Proceedings of IncoME-VI and TEPEN 2021

This paper analyzes the error model of inductosyn install in an angular-measuring instrument, and then build up a test error system using the ultra-high precision Ring Laser Gyroscope to get the error of inductosyn. Compare the traditional Fourier Function calibrations and the linear interpolation calibrations methods; this paper presents a bit memory based method to calibrate the error of inductosyn in electrical resolution. As a result, after compensation, the inductosyn error is ±0.8 arcsecond, RMS of the error is less than 0.3 arcsecond.

Predictive maintenance is fundamental for modern industries, in order to improve the physical assets availability, decision making and rationalize costs. That requires deployment of sensor networks, data storage and development of data treatment methods that can satisfy the quality required in the forecasting models. The present paper describes a case study where data collected in an industrial pulp paper press was pre-processed and used to predict future behavior, aiming to anticipate potential failures, optimize predictive maintenance and physical assets availability. The data were processed and analyzed, outliers identified and treated. Time series models were used to predict short-term future behavior. The results show that it is possible to predict future values up to ten days in advance with good accuracy.

Vibration signals are often used in the fault diagnosis of rotating machinery. However, due to the influence of complex environment, environmental noise is often doped, and the diagnostic accuracy is reduced. The traditional deep self-encoder is used in the noise reduction process of rotating machinery fault diagnosis. The pooling model is poor and easy to lead to over-fitting problems, and deep learning training needs a large number of labeled data. Therefore, this paper proposes a reinforcement learning fault diagnosis method based on less label data. The random pooling is used to replace the pooling layer of the original convolutional self-encoder, and the exponential linear unit (ELU) is used to replace the original activation function to enhance the convolutional self-encoder. A large number of unlabeled samples are used for training, and then the deep reinforcement learning is used for network fine tuning. The experimental results of the sensor data collected by the fault diagnosis test bench show that the method used has a good improvement in denoising ability and feature extraction ability, and the recognition accuracy and stability are better than traditional convolutional autoencoder and traditional machine learning methods.

In order to improve the processing precision and surface quality of line-drawing and boring in ship manufacturing process, a virtual prototype of micro-boring machine was built on the basis of existing portable boring machine, aiming at the inconvenient boring operation and low processing precision of small and medium-sized ships at present. Then it carries on the optimization design to its transmission mechanism, uses ADAMS to carry on the kinematic simulation analysis to the virtual prototype; Finally, reasonable simulation data are obtained to ensure the performance, accuracy and efficiency of boring, and provide basis for the subsequent optimization design.

Underwater vehicles (UV) with a deeper operation ability are the important research field in the marine industry. In order to obtain a better and safer operation performance, the strength, stability and vibration characteristics of the pressure shell of UV should be analyzed. In this paper, a finite element model of the pressure shell of buoyancy regulator is developed. The influences of the stiffener thickness, width, position, and shell thickness on the strength, deformation, and load factor of the pressure shell are studied. In addition, a lighter and safer shell structure is obtained by using the response surface optimization method. The simulation results show that the above factors have great influence on the shell characteristics, such as strength, stability and modal parameters. Moreover, a lighter pressure shell used in the UV can be helpful for providing a better possibility to carry more equipment.

The vibration experiment of the sugarcane harvester is of significant value, and it is mainly done in the sugarcane field. This method has low efficiency, poor security and reliability. So, a sugarcane field excitation device is designed in this paper based on the sugarcane field excitation signal already collected. The dynamic characteristics of the sugarcane field excitation device are studied by using theoretical analysis, simulation analysis and experimental research methods. The multi-body dynamic model is studied by using rigid-flexible coupling simulation technology. Based on the simulation results, the sugarcane field excitation device is manufactured. The output frequency of the sugarcane field excitation device is calibrated by the speed calibration method. Finally, based on the experimental optimization results, the function of the sugarcane field excitation device is verified.

Aimed at improving the sugarcane ratoon cutting quality of sugarcane harvesters for hilly areas, a sugarcane harvester experiment platform was developed. During sugarcane cutting experiments, flaws may appear in sugarcane ratoons. The flaw number, thickness and length were measured. An experiment index, y was introduced as the comprehensive cutting quality evaluating value of the sugarcane ratoon cutting quality by the improved entropy method with these three parameters. A regressive mathematical model was set up by orthogonal experiments and to study effects of the vibration frequency, the vibration amplitude, the cutter rotating velocity, the sugarcane feeding velocity and the cutter installing angle on the sugarcane ratoon cutting quality. It is shown in experiments that there is a strong linear relationship among y, vibration amplitude and frequency and the amplitude as well as frequency had great effects on y while the moving velocity, the cutter rotating velocity and the cutter installing angle had relatively less significant effects on y. According to the fact whether the effect on y is significant, the significance order is as follows, the vibration amplitude, frequency, the sugarcane feeding velocity, the cutter rotating velocity and the cutter installing angle. Interaction between the vibration amplitude and frequency and that between the amplitude and the cutter rotating velocity also have effects on y. In details, the greater the vibration amplitude and frequency are, the greater y will be, which means the worse the sugarcane ratoon cutting quality will be. The greater the vibration amplitude and the cutter rotating velocity are, the greater y will be. This research was done to study the effect mechanism of the sugarcane ratoon cutting quality and lay the foundation of design and development of sugarcane harvesters with a high sugarcane ratoon cutting quality for hilly areas.

Taking sands as the research object, the influences of changes in particle sizes on the simulation results were analyzed in the simulation test based on DEM (Discrete Element Method) using tests and comparative simulation methods, so as to study the influences of particle changes on simulation results in the simulation process using DEM. According to the simulation using particles of varied sizes, the angle of repose is changed marginally with minor errors in the test of the angle of repose. Also, particle size has insignificant influence on the simulation results in the test of the angle of repose. That is to say, the simulation test of the angle of repose is less sensitive to the particle size. However, particle size exerts a remarkable influence on the simulation result in the vane shear test. Specifically, when the particle size is changed, then the maximum torque of the vane will be significantly fluctuated, indicating that the vane shear test is highly sensitive to the particle size. When the magnification factor m ≤ 4, the variation trend of torque will be consistent with reality. In that case, the particle magnification should be less than 4 times.

A wind turbine condition monitoring method based on cointegration analysis is proposed. The co-integration residual obtained by the co-integration process of the SCADA data of the wind turbine is used for monitoring the operation state of the wind turbine. Take the experimental data of a 1.5 MW doubly-fed wind turbine from Jinjie Company in Baotou City, Inner Mongolia, under different environmental and operating conditions, and conduct experiments on the proposed method. The method was tested with known failure cases. The results show that this method can effectively monitor the running status of wind turbines.

Smartphone glass covers defects detected by human, which is inefficiency, high costs, low detection accuracy and labour intensive, while the automatic detection methods based on traditional machine vision is poor flexibility, low yield and poor generalisation capability. Therefore, this paper introduces YOLO (You Only Look Once) v3 to smartphone glass cover defects for the first time. The YOLOv3 algorithm was improved for the actual characteristics and specific requirements of defect detection. First of all, the channel attention mechanism SENet (Squeeze and Excitation Networks) was added to the feature extraction network to detect inconspicuous defect features. Moreover, a 104 × 104 scale detection layer was added to the YOLOv3 detection network to solve the problem of multi-scale defects. Finally, the scaling factor coefficient of the BN (Batch Normalization) layer in the convolutional network is used as the important factor for model pruning to improve the defect detection speed. The improved YOLOv3 algorithm is applied to smartphone glass cover defect detection, and a high accuracy and high detection speed method for smartphone glass cover defects is proposed. 15,914 production site images covering four types of defects, including chipped edges, pits point, soiling and scratches, were obtained from smartphone glass cover manufacturers, 14,321 were annotated as the training set and 1593 were used as the test set to compare and analyse the proposed method and the original YOLOv3 algorithm in this paper. These experiments showed that the mAP (mean average precision) of the detection was 81.0% and the detection speed was 43.1 sheets/s. Compared to the original YOLOv3 algorithm, the mAP of the detection increased by 3% and the detection speed increased by 6.7 frames/s, which meets the need for high precision and efficient detection of defects in the industrial production of smartphone glass covers.

A company mainly produces varistor (e-var). Through observation and analysis, the e-var production line of company a has problems such as low production efficiency, high rate of defective products, more in-process products, nonstandard operation of production line and long waiting time. This paper explores and improves the above problems by the methods of value flow chart analysis, 5W1H method, ECRS principle, standard operation sequence, rapid model change, setting up control experiment and causal chart. The e-var production line is integrated and standard work is formulated. The die changing can be realized quickly in the pressing process, and the quality problems of the products are comprehensively managed. After improvement, company a has improved production efficiency and the rate of defective products decreased significantly, reducing production cost.

Modal analysis is critical to better understand structural dynamic vibration characteristics by extracting system’s natural frequencies, damping ratios and mode shapes. Modal analysis has been widely used to structure optimization in the design stage, damage detection and structural health monitoring or condition monitoring. According to whether need artificial exaction, the modal analysis techniques can be categorized as experimental modal analysis and operational modal analysis. Conventional experimental modal analysis has to measure the excitation and corresponding response in the meantime, while operational modal analysis measure system’s response only during normal operating condition. Therefore, operational modal analysis also called output-only modal analysis methods, which have developed dramatically in recent decades because it is promising as means to achieve structural online monitoring, which is highly desirable for critical mechanical system, important buildings and bridges, etc. This paper made a brief review of the development of popular operational modal analysis techniques and their applications in condition monitoring.

Underwater gliders are mainly used to monitor the marine environment. In order to reduce the interference of self-vibrations, the rudder drive system of gliders is analyzed. Firstly, the meshing frequencies of the gears are calculated. Then, the vibrations of the actuator of rudder drive system are tested; and a finite element model of the rudder drive system of underwater glider is developed. Finally, the gear meshing frequencies and the vibration frequencies of the actuator are compared with the results from the finite element model analysis. The results show that the gear meshing frequencies and the vibration peak frequencies of the actuator are different from the natural frequencies of the rudder drive system, the system has no resonance and the structure design is reasonable one.

Kaplan turbines rely on an operating mechanism inside the runner hub to control blade angles. Contact surfaces of the moving parts on these mechanisms are constantly subjected to frictional and contact forces, inflicting wear which can lead to malfunctioning and performance reduction. In this paper, a novel method for wear assessment in individual blade joints of the runner operating mechanism is presented. The technique consists in monitoring blade angles separately during turbine operation through inductive proximity probes mounted on the discharge ring. These angles are contrasted with the operating mechanism positioning data at several instants and the performance of each joint is evaluated. This technique has been implemented on a 92 MW Kaplan turbine. In October 2018, excessive clearances in three blade joints were detected and an inspection was recommended during the next programmed maintenance. The runner hub was later disassembled and all joints inspected, which confirmed those joints had been worn down and were replaced. This result shows that the proposed method can effectively assess clearances on blade joints during operation, providing an early detection method to anticipate mechanism malfunction and incorporate in the Condition-Based Maintenance plan for production optimization.

In the wind power generation system, the bearing plays a very important role. Whether it can run stably directly determines the quality of the electricity produced and has a great influence on the efficiency of power generation. Due to the harsh working environment, the bearing has become one of the most vulnerable components in the entire wind turbine system. Therefore, bearings of wind turbines need to be maintained regularly. However, it needs to be shut down every time for maintenance, which will incur high maintenance cost. So, the fault diagnosis of the bearing is particularly important. A fault diagnosis method is proposed based on deep learning in this paper. This method is based on the residual module to construct a new ResNet model and embeds the attention mechanism in it to select information that is more critical to the current task goal from a lot of information. In addition, a long short-term memory is added to the network to extract the long-term dependence of the vibration signal and ensure that the information on the time series will not be lost as the training progresses. The experimental results show that the method proposed in this paper is very effective for the fault classification of fan bearings.

This case makes full use of the knowledge of IE in various aspects, such as work research, ergonomics, and other methods to put forward four different improvement proposals, including two work table design, one fixture design and one operator man–machine operation design, solving the problem of one company H product has a large inventory of WIP and so on. After the proposal was put forward, the Flexsim simulation technology and other methods were used to evaluate the scheme which is proposed in this case and verify the feasibility of the scheme. The final improvement proposal resulted in a total reduction about 2709 s in appearance inspection engineering cycles, at the same time the new improved workstation made it easier for employees to work, to reduce physical injury to employees due to long working hours; Four employees were reduced in the electrode magnetic coil project, and the annual wage expenditure was saved by about 198,144 yuan.

A viable data driven approach for determining dynamical systems describing engineering processes would be a valuable tool in condition monitoring. The application of the SINDy algorithm for dynamical system discovery is investigated in the context of a reciprocating compressor. A feasibility study was carried out in which an attempt was made to recover a model of the compressor from synthetic data obtained from that model. A simplified model of the compressor with two degrees of freedom was developed from an existing model. Following the SINDy approach a parsimonious model was constructed from a large library of functions using sparse regression. This model has the same structure as and similar coefficients to the original model thus demonstrating the potential of this approach.

Bearings are the most critical components in modern industrial rotating machinery. If a bearing is damaged, it can lead to serious consequences such as an interruption to a production line and financial losses. It is important to monitor the bearing operation condition and to predict the remaining useful life (RUL) of bearings so that a scheduled maintenance can be planned ahead. In order to improve the accuracy of a bearing RUL prediction, a new data-driven RUL prediction technique based on Long Short-Term Memory (LSTM) network and Transformer network is proposed. Firstly, a total of 8 degradation characteristics in both time and frequency domains are extracted from the bearing data to be used as the input features. After the data preprocessing steps such as normalization and sliding window interception, the degradation characteristic dataset is obtained. Then, the proposed LSTM-Transformer technique is applied to the characteristic dataset for training and prediction. The prediction result shows that the proposed technique can effectively overcomes the information loss of LSTM network caused by the increase distance between the input and output sequences to produce a more accurate RUL prediction. The RUL prediction obtained using the proposed technique is compared with those using existing techniques such as GRU, LSTM and CNN networks for an evaluation of the effectiveness and efficiency of the proposed technique. It is confirmed that the proposed technique can yield a more accurate bearing RUL prediction than the existing techniques.

Under the condition of variable speed, the traditional signal processing method cannot accurately determine the fault location of motor bearing. This paper proposes and studies the order analysis method without tachometer. Through the short-time Fourier transform and speed tracking of the collected fault signal, the speed signal of the motor fault bearing is obtained indirectly. The extraction process of the speed signal is completed and the order analysis is carried out to obtain the bearing fault diagnosis results. The results show that this method can effectively judge the fault location of motor bearing.

Axis parallelism error cause edge contact and stress concentration, which lead to uneven load distribution and uneven wear of tooth surface, and seriously reduce service life. In this paper, a tooth contact analysis (TCA) model and a gear wear analysis model considering the axis parallelism error were established based on the basic equation of contact problem and Archard’s wear equation, and the change of contact characteristics and wear depth were analyzed. The results show that when the axis parallelism error exists, the tooth load is uneven, and the load distribution becomes more uneven with the increase of the error, and the uneven contact is improved with the increase of load. The wear depth decreases in the tooth width direction, and the wear depth decreases firstly and then increases along the line of action, and the wear depth of the pitch point is zero. The wear depths reach their maximum value at the root of the pinion where the stress concentration occurs, and the maximum wear depth increases non-linearly.

It is difficult to obtain enough data to train a robust diagnosis model for different rolling bearing faults, and the existing intelligent bearing fault diagnosis algorithms have insufficient generalization ability. Therefore, a rolling bearing fault detector based on the time–frequency graph and convolution neural network (TFG-CNN) is introduced to improve the generalization performance of the fault diagnosis algorithm as much as possible under the condition of considering the diagnosis accuracy and sample size. The specific implementation method is to use Fast Fourier transform (FFT) to transform the vibration data of rolling bearing into a two-dimensional network graph, and then use CNN to classify them. Finally, the performance of the proposed method is analyzed by using the rolling bearing fault datasets of Case Western Reserve University, and analysis results show that the proposed method can simultaneously diagnose the fault location and severity of rolling bearing, and has good cross-domain diagnosis ability and anti-noise performance.

Degradation of gas path performance is the focus of gas turbine condition monitoring. In order to realize gas turbine maintenance cycle performance monitoring, diagnosis and prediction, combining mechanism knowledge and data information, a gas turbine gas path performance digital twin modeling method is proposed, and a health monitoring framework of the gas turbine gas path performance digital twin is constructed. Taking the split-shaft gas turbine as the research object, the gas turbine performance calculation model is established without relying on component characteristic information, the parameter matching method based on differential evolution is studied, and the gas turbine performance digital twin model is developed. The simulation test was carried out the results show that the gas turbine gas path digital twin model realizes the quantitative characterization of the performance degradation of gas path components during the maintenance period, and provides a basis for the gas turbine gas path performance fault diagnosis.

Aiming at the problem that there are all kinds of noise interference in the planetary gearbox of wind turbine in the general experimental scene, the vibration data obtained is less and the fault characteristics are not obvious. A MKFCNN-LSTM migration learning algorithm based on multi-kernel fusion convolution neural network (MKFCNN) and long and short time memory neural network (LSTM) is proposed to realize the fault diagnosis of wind turbine planetary gearbox. Firstly, the MKFCNN is constructed to extract the multi-scale spatial features of the sample signal, and then it is connected in series with LSTM to extract the corresponding time information of the sample signal. In view of the associated fault feature information between the rolling bearing data set and the planetary gearbox data set, the rolling bearing vibration signal of the Western Reserve University is input into the MKFCNN-LSTM as the source domain sample data, and iterative training is used to update the network weight and offset value. The pre-trained MKFCNN-LSTM is obtained, and then fine-tuned by inputting the vibration data of the planetary gearbox with small samples in the target domain, the weight and offset values are transferred from the source domain to the target domain, and finally the accuracy of fault recognition based on the number of small samples in the target domain is improved. The experimental results show that the proposed method can apply the original fault diagnosis knowledge to the vibration data set of the planetary gearbox in the laboratory. Compared with stack autoencoders (SAE), support vector machine (SVM) algorithm, the accuracy of fault identification and classification is improved to a certain extent.

The economic sustainability of any industry is directly linked to the management and efficiency of its physical assets. The maintenance of these assets is one of the key elements for the success of a company since it represents a relevant part of its Capital and Operational Expenses (CAPEX and OPEX). Due to the importance of maintenance, a lot of research has been done to improve the methodologies aiming to maximize physical assets’ availability at the most rational costs. The introduction of Artificial Intelligence in the world of maintenance increased the quality of prediction on equipment failures, namely when associated to continuous equipment monitoring. This paper presents a case study where a neural network is proposed to predict the future values of various sensors installed on a paper pulp press. Data from the following variables is processed: electric current; pressure; temperature; torque; and speed.

Wind turbine blade is an important component of wind turbine. Wind turbine blade crack damage will cause hidden danger to the operation of wind turbine. The current wind turbine blade defect detection mainly relies on manual inspection, and the image detection technology can improve the inspection efficiency and reduce the unit maintenance cost. In view of the existing wind turbine blade crack defect detection algorithm with low recognition rate and low accuracy, a YOLOv4-based wind turbine blade crack detection method is proposed. First establish the wind turbine blade crack image dataset, then the anchor box parameters in YOLOV4 are optimized by K-means++ algorithm to make the anchor box parameters match the crack defect size; BiFPN is used instead of PANet to achieve better feature fusion, and finally the Focal Loss function is introduced to balance the number of small size defect samples in the data. The comparison tests show that the AP of the improved YOLOv4 algorithm reaches 93.49, which is better than the original YOLOv4 and the other three comparison algorithms, and has better efficiency and practicability.

Metal debris detection technology in online oil monitoring has drawn significant industrial attention recently since it can provide information about wear, lubrication and friction conditions of friction pairs. In the paper, the principles of several types of metal debris sensors based on photoelectric or imaging, X-ray luminescence spectrum, ultrasonic detection and electric impedance measurement are reviewed. Especially, the inductive debris sensors which gain advantages of simple structure, complete flow measuring and the ability to distinguish ferromagnetic and non-ferromagnetic metal particles, have received extensive attention. The developing progress, detection principle, typical sensor structure, and industrial applications are presented in detail. We also provided the prototype of debris sensor we developed and the performance and application in bearing health status evaluation and remain life prediction. Finally, the main problems confronted and the development trend for debris sensor using in industrial applications are also proposed.

Deep learning methods such as the one using Convolutional Neural Network (CNN) have made remarkable achievements in computer vision and natural language processing. Compared with the conventional neural network structures, CNN features low complexity, fewer parameters, and higher degree of nonlinearity. As the sizes of sensor signal input are often different from those of image input, using CNN to monitor the equipment status is a new issue compared with image recognition. To examine the impacts of various one-dimensional CNN structures on the regression of performance parameters, this paper conducts a preliminary study on the application of CNN in equipment status recognition, and utilizes published simulation datasets of ship propulsion equipment to train and test one-dimensional CNN models with different structures. The results show that the size of convolution kernels hinges on the attributes of input features when one-dimensional CNN is used for data regression prediction. In the case of independent and direct feature input, the training effect can be effectively improved by using 1 × 1 convolution kernels and the Network In Network (NIN) structure.

Because the extra-wide slab continuous caster has larger sectional width, heat emission condition is different between the length and width. Therefore, it is inevitable that the water distribution is regional separation along the casting slab width and length. It brings a lot of difficulties to determine process parameters and control continuous casting process, and causes quality defect. Based on 3250 mm × 150 mm extra-wide slab casting and production process, the paper established the finite element model, calculates the temperature distribution and shell thickness of the slab with different casting speed and superheat, to study influence of process control parameters of wide slab quality.

Gearboxes are the most significant components of an electromechanical system. They are often exposed to various abnormal working condition which causes damage to the gear. These damages can be of any form such as tooth breakage, tooth wear. Furthermore, this breakage will cause changes in gear tooth mesh stiffness, thereby reducing the gear dynamics. Recently, several attempts have been performed for the detection of localized gear tooth faults using electric signatures from induction motor with promising results. However, the interaction between the motor-gear dynamics to detect and diagnose this fault has not been fully investigated. Therefore, this study proposed a 6° of freedom (DOF) electrical motor model integrated with an 18° of freedom (DOF) gear dynamic model to fulfil the detection of gear faults using motor current signature. In this model, a comparison of electric torque and constant torque as the input for the gear model has been investigated to study the influence of electric motor torque on gear dynamic. Furthermore, this study considered different severities of tooth breakage to demonstrate the performance of motor current in gear tooth breakage detection and its location. The numerical results show an increase in amplitudes at the frequency of $$f_s \,f_{r1}$$ f s f r 1 and $$f_s \,f_{r2}$$ f s f r 2 as the severity of the fault increases at different stage of the gearbox which can reflect the presence of tooth breakage. This proposed numerical analysis does have a good agreement with the experimental validation.

Multiple faults diagnosis is a critical problem in mechanical fault diagnosis. Fault behavior analysis aims to find out the response characteristics of operating parameters under different faults and is the most primitive task for multiple faults diagnosis. This paper proposes a bond graph-based approach to analyse the mechanical fault behavior for diagnosis. The analytical model of an engineering system is firstly established via bond graph. The temporal causal graph is derived from the bond graph model to depict the analytic relationships system variables. The operating parameters response characteristics under different faults are then derived by representing the faults with abnormal change of system variables. The approach is illustrated via an engine lubrication system. The presented approach avoids time-consuming formula transformation which is necessary in mathematical model-based approach, and therefore provides an efficient for fault behavior analysis.

Enlarged clearance in connecting rod big end bearing is known to be a typical fault in internal Combustion (IC) engines such as diesel engines. Enlarged clearance causes inefficient operation that leads to reduced performance of the engine. In this work, the influence of enlarged clearances in big end bearing on the vibration of the system is studied. The kinematics and dynamics of the big end bearing were explored using numerical simulation to investigate the effect of the clearance on the associated response of acceleration signals obtained by multiply the simulated bearing forces with frequency response function (FRF). Different clearance and engine speed conditions were simulated for analysis of various degree of bearing clearance fault. Results obtained for these cases are evaluated for vibration analysis which shows that simulation model developed is able to simulate the vibration signals issuing from the dynamics of engine system with various clearance phenomenon.

Centrifugal pumps are commonly used in pipelines with moderate head and discharge requirements for hydraulic transportation of liquids and solids over long distances. The performance characteristics of the pump and erosion wear are the most significant design and selection parameters. Mechanical wear can cause considerable damage to the impeller in the pump, reducing the pump’s lifetime and efficiency. This paper investigates the impeller wear of a centrifugal pump using an acoustic mechanism at different flow rates. As the peripheral velocity or circling radius at the impeller surface increases, a small amount of wear in the impeller inlet region rapidly develops. The uniform corrosive wear area is defined as the portion of the impeller where the impact velocity is less than the critical value, and it expands as the impeller velocity increases. In fact, wear mechanisms differ from one region to the next; when the tangential portion of the impact velocity is high, the impeller wears out faster. Because of the high capability of noise reduction when modulating the signals, modulated signal bispectrum (MSB) analysis is used for extracting the incipient fault signature. The experimental results show that the diagnostic features produced by modulated signal bispectrum allow for early detection of the initial impeller wear fault. Furthermore, the MSB study of acoustic signals demonstrates the efficacy of airborne sound-based monitoring. It offers a strong attestation of complete distinction between safe and defective conditions at various flow rates.

It is important rotating machinery for gas turbines in aviation, shipbuilding, and other industries. Given the high failure rate of the gas turbine rotor system, fault diagnosis of the rotor system is completely vital. Aiming at the fault diagnosis of the gas turbine rotor, we adopt a method based on Multipoint Optimal Minimum Entropy Deconvolution Adjusted (MOMEDA)—Variational Nonlinear Chirp Mode Decomposition (VNCMD) in this paper. For the gas turbine rotor test rig data, the original data is first analyzed for effective value, the fault signal is extracted, the fault signal is filtered by MOMEDA, the processed filtered signal is subjected to VNCMD decomposition, and the signal is reconstructed according to the magnitude of spectral kurtosis, and passed Envelope analysis to extract fault characteristics. This paper analyzes the data of the gas turbine rotor test bench, and the results show that the proposed method has achieved excellent results in the fault diagnosis of the gas turbine rotor.

In view of the problems that the signal features are difficult to extract when a fault occurs in a rolling bearing, and the time domain image of the original vibration signal cannot obviously show the feature differences of different faults, and the direct deep feature learning and recognition will have a large impact on the system performance, etc., a bearing fault classification and recognition method based on symmetry dot pattern-convolutional neural network (SDP-CNN) is proposed. First, the SDP method is used to analyze the vibration signals of different faults, and the signal SDP images obtained can clearly show the feature differences of different faults; then, the SDP images are input into the CNN network for feature learning and state recognition; finally, Validation was performed using the Case Western Reserve University (CWRU) bearing dataset. The results show that the recognition accuracy of this method is 97.5%, which further verifies that the deep learning algorithm can adaptively extract the features of the SDP image and effectively identify bearing faults.

Due to the complex structure of the spindle and many influencing factors, the failure analysis of the spindle has always been a focus. Use Soildwoks to model the Dalian machine tool VDL600 machine tool spindle, analyze the modalities of the spindle and main components through the ANSYS Workbench platform, and analyze the possible resonance frequencies. Taking the data measured by Lion's gyration accuracy tester as a reference, the MEMS acceleration sensor is used to obtain the vibration signal of the spindle, and the modulation signal bispectrum (MSB) analysis method is used to obtain the characteristic frequency of the vibration signal. By comparing and analyzing the characteristic frequency of the vibration signal and the modal analysis result, it can be proved that the MSB method can extract the resonance frequency of the spindle well.

Aiming at the effective identification of failure modes of rolling bearings, a support vector machine (SVM) and Levenberg–Marquardt (LM algorithm) fault diagnosis method for rolling bearings is proposed. First, use wavelet packet decomposition to obtain sub-bands, reconstruct the decomposition coefficients, and expand the decomposed sub-band signals to the original signal length; then, use SVM to classify the fault state; finally, input the feature vector into LMNN (LM algorithm Neural network) to realize failure mode recognition. The method is verified by the rolling bearing fault diagnosis experiment. The results show that the SVM-LMNN based on wavelet packet decomposition has a rolling bearing fault diagnosis accuracy rate of up to 99.456%. The method proposed in the study is compared with the instantaneous energy method of the VMD component of the kurtosis criterion and the enveloping spectrum solution diagnosis method, and the higher accuracy is obviously obtained, which proves the feasibility and effectiveness of the proposed method.

Nowadays, neural network become popular in modeling. However, the model training needs a lot of data, long training time and high hardware conditions. It is inefficient for ordinary computing devices to be used in training models. In this paper, VGG16 model was modified to fit ten labels and used as feature extractor. The default image size of model was 224 × 224 pixels. Then the images were reduced into low resolution as 112 × 112, 75 × 75, 56 × 56, 45 × 45, 32 × 32 pixels, which were 1/2, 1/3, 1/4, 1/5 of default side length and the minimum size. Next these images were sent to model for training. The training results illustrated that the images of 112 × 112, 75 × 75, 56 × 56 groups can still be adequate for modified VGG16 to classified and achieve high accuracy and meanwhile significantly reduce the training time. However, when the size dropped to 45 × 45, 32 × 32, overfitting appears and the training accuracy significantly dropped. Thus, it is recommended that set a target accuracy first and begin training from a small size. If the accuracy was not high enough, enlarge the size and train again.

Internal combustion (IC) engines are widely employed in power systems such as marine ships, small power stations and vehicles. However, due to its complex working conditions and sophisticated degradation mechanisms, IC engines commonly suffer various types of malfunctioning and faults, which affects their performance in power delivery. Therefore, it is important to monitor the condition of IC engines and detect faults occurred in time. In this paper, two unsupervised data-driven models using machine learning techniques are employed and investigated for the purpose of online condition monitoring and fault isolation of IC engines. A misfire and a lubrication system filter blocking faults are experimentally studied on a purposely built marine engine test rig. The performance of the two models and their contribution maps are discussed, which provides guidance for using such unsupervised models for the condition monitoring and fault detection of IC engines.

Piping systems are widely utilized in industry and home. Leakage of piping systems induced by prolonged corrosion, severe weather, or man-made damage will lead to serious consequences like explosion disasters, severe damage to industrial equipment, unforeseeable waste of resources and even threaten human life. WSNs significantly attract attentions in Industry 4.0 in recent years due to their advantages of wide distribution, remote controllability, convenient portability, easy programming, and economy. Meanwhile, as a non-intrusive measurement technique, vibration manifests a great potential for leakage detection of piping systems. In this paper, a vibration-based wireless sensing system is developed to remotely monitor the condition of piping systems in real time. According to the analytical results of vibration signals at two different positions on the piping system, the effective statistical features are extracted at the wireless sensor node to detect the leakage and its severity of the piping system. Furthermore, it can reduce the amount of data transmitted to reduce the power consumption then prolong the service life of the designed wireless sensing system. The diagnostic result can be conveniently observed on the mobile device in real time.

Recently, intelligent fault diagnosis technology based on deep learning has been extensively researched and applied in large industrial equipment system for ensuring safe and stable production. However, these deep models only effective when enough data for each observed failure category are available in the training durations. Otherwise, the performance of these models will notably decrease. As the critical component in large machinery, the gearbox often changes the speed and load along with the production demand in the practical application, which caused few data samples to be collected at certain conditions. This phenomenon introduces the few-shot fault diagnosis, and its goal is to identify the fault types with extremely limited data samples. To address this problem, a Meta-learning guided Few-shot Fault Diagnosis method, named MFFD, is proposed for gearbox fault diagnosis under limited data conditions. The results verify the effectiveness of our MFFD method at one-shot and five-shot fault diagnosis tasks under different speed and load conditions.

As motor faults could lead to unwanted loss in industry, it is important to find out the motor faults in time. Currently, with the popularity and mature application of deep learning, researchers in the field of electrical machine health assessment have begun to focus on deep learning methods. It is hoped that motor fault detection can be achieved with the help of deep learning methods. This paper presents to adopt deep learning methods represented by LSTM neural network for motor fault diagnosis and evaluates on our own experimental platform. Considering two typical motor faults with two different degrees of severity, the results show that the proposed LSTM approach has a high accuracy (98.81%) on motor fault classification. The results also confirm that: (1) adequate effort of preprocessing, including sample length selection in the time domain and frequency band selection in the frequency domain, can significantly improve accuracy and computational efficiency; (2) different faults can be separated through the information in frequency band of 100–1000 Hz, which has not been fully modelled analytically before.

Health condition monitoring (HCM) of an aircraft engine is crucial to enhance its reliability running. In this paper, a novel method is proposed using the complex model to monitor its degradation trend. With the help of the data collected by multi-sensors, a complex dynamic model is built using the data-driven approach, which aims to achieve the purpose of HCM of aircraft engine. First, the Gath-Geva fuzzy clustering method is utilized for health condition division. Second, the network model based on correlation analysis is conducted. Finally, the dynamic improved logistic model is developed to describe the changes of sensors data of aircraft engine degradation trend. To verify the effectiveness of the proposed method, simulated aircraft gas turbofan engine data is utilized for validation. The results demonstrate that our method is effective to track its degradation process of aircraft gas turbofan engine.

Gearbox is a key component in rotating machinery and prone to chipping fault due to poor working environment. Hence, it is necessary to carry out research on chipping fault. Time-varying mesh stiffness is a periodic function caused by the change in the number of contact tooth pairs and the contact positions of the gear teeth. Time-varying mesh stiffness of is one of the main sources of vibration of a gear transmission system. Time-varying meshing stiffness provides the important information about the health status of the gear system. When a chip happens in one gear, mesh stiffness will decrease and consequently the vibration properties of the gear system will change. The vibration change can be characterized through gearbox dynamic modelling approach. In order to comprehensively understand the vibration properties of a gear set, it is essential to evaluate the time-varying mesh stiffness effectively. In this paper, the potential energy method is applied to analytically evaluate the time-varying mesh stiffness of a gear set. A modified cantilever beam model is used to represent the external gear tooth and analytical equations are derived without any modification of the gear tooth involute curve. A chip model is developed and the mesh stiffness reduction is quantified when chipping fault occurs in the pinion or the gear.

In the rail vehicle structure design, the key of the strength design and fatigue life estimation is to analyse the stress state of the structure under the dynamic load. As the stress cannot be measured directly, the displacement response model of the body structure is established by the displacement mode analysis method, and the strain response of the car body is obtained by the relationship between displacement and strain, thus the stress state is obtained. Since displacement to strain is a differential process, the variation of displacement will be magnified and the error will be generated. Strain mode theory and its property are derived from displacement mode theory. The results show that it is more sensitive for strain mode than displacement mode through the simulation analysis of the car body equivalent vertical model calculation. Strain mode difference curve can determine the structural damage location. The vehicle FE model verifies this result. The strain and stress versus time history of car body can be obtained by the mode superposition method, which provides basis for fatigue life prediction and load spectrum research.

For chip testing, in the process of the actual chip manufacturing since most of the chip size is very small, so the artificial extremely difficult to discern defect signals, such as lack of pin, pin bending, surface defects such as scratches, lack of the shape signal, thus easy to cause the yield is not ideal, therefore in the process of actual production introduction of machine vision. Chip defect detection system based on machine vision is a kind of machine vision, chip bearing platform, automatic rotating disc, etc., on the basis of combining computer terminal to control the whole test system, in view of the chip pins, surface, shape features such as visual algorithm analysis, finally through the man–machine interface technology of motion control system and chip testing results show that the Finally, the system is made and the best detection state is debugged. It can improve the yield of products and improve the production efficiency in the actual manufacturing process.

As the amount of data generated by monitoring the condition of rolling bearings is increasing, matrix factorization-based collaborative filtering can effectively dig out valuable fault information from it. However, in practice, the amount of data generated by the normal state of the bearing is much larger than the amount of data of the bearing fault. As the total amount of data increases, this imbalance will become more and more and more severe, bearing fault information is often overwhelmed in it. In response to this problem, this paper starts from the perspective of mathematical statistics, a method of mean conjugate prior is proposed for the bearing normal condition data of bearing score matrix, from which the prior distribution of the probability distribution parameters of the bearing fault data is obtained. Then combined with the Bayesian method, we get the posterior distribution. According to the distribution, the random number is used to construct the Bayesian probabilistic scoring matrix (BPSM). Relying on BPSM, the collaborative filtering recommendation algorithm is used to identify different types of faults in rolling bearings. Under unbalanced data, comparing with the identification under a conventional joint score matrix (CJSM), the model built based on BPSM has a better identification effect on bearing fault state.

Bending deformation is easy to occur when turning a long slender workpiece due to its low stiffness, which seriously affects the machining dimensional accuracy. Currently, the dimension of the part is generally measured off-line after the completion of the operation. The purpose of this paper is to explore an on-line monitoring method for the dimensional error of slender shaft in turning processes. First the deformation of the whole workpiece in the process of machining is analyzed. The deformation correlation at the measuring point and that at the cutting point is deduced. Then an on-line monitoring approach to radial dimension is proposed using a single fixed displacement sensor and Wavelet Transform. Finally, the reliability of the monitoring method is verified by machining experiments. The experimental results show that the presented on-line monitoring model enables to predict the dimensional error of the machined workpiece effectively.

The Naxi people in Lijiang, China, have created a pictograph that represents the Dongba culture. The ancient books written with this script are one of the three world heritages in Lijiang, and are known as “the only living ancient script in the world” (Likun in Dongba ancient books and documents of Naxi nationality 2021 [1]). Dongba hieroglyphs are written by different Dongba elders. Different writing habits lead to the phenomenon of variant characters in multiple versions of the same character, as well as the complex structure, different forms, complex background and image noise of Dongba text, this paper puts forward two parts to realize feature extraction and image recognition, topological characteristics and characteristics of grid, Input of neural network training, and combining the neural network using multilevel identification model, template matching and through experiment verification, this algorithm is 9.92% higher than that of the recognition rate of template matching algorithm, and the algorithm of recognition is got improved significantly, the results show that the method to achieve accurate and efficient implementation of Dongba hieroglyphics identification purposes.

As a clean energy, the development of wind power has attracted wide attention. In view of the characteristics of non-linear and non-stationary mixed signals in the vibration state of wind turbines, the separation of noise is the key problem of information feature extraction. In this study, sensors are utilized to collect blind source signals and mixed matrix information in order to retrieve source signals and extract features from information. This paper integrates EMD (Empirical Model Decomposition) with ICA (Independent Component Analysis) with the aim of extracting feature signals from the wind turbine generator system (WTGS). By analyzing signals with obvious fault characteristics, this approach considerably increases the accuracy in extracting feature signals from the WTGS transmission system.

Fault diagnosis in reciprocating compressor (RC) requires time-consuming feature-extraction processes due to the complexity of the compressor operation and fluid–solid interaction. This causes the useful information to be corrupted and difficulty in accurately diagnosing the faults with traditional methods. The aerodynamic phenomenon has a large impact on acoustics signal compared to the vibration. Thus, this paper presents analytical modelling of compressor sound highlighting the important sound sources and their generation. The additional contribution of this paper is the application of a state-of-the-art signal processing technique: Modulation Signal Bispectrum (MSB) which overcomes the challenges by showing good noise suppression capability and characterising the modulating components present in the signal, thereby resulting in stable modulation components for accurate diagnostics. The result reveals that the fault diagnosis based on airborne acoustics using MSB method outperformed the vibration-based method.

Rolling bearings are widely used in modern machinery and equipment, and the tough working environment is easy to cause their failure. To solve the problem of extracting fault signals of rolling bearings in a strong noise environment, a method based on Normalized Least Mean Square(NLMS) adaptive filtering and Ensemble Empirical Mode Decomposition(EEMD) noise reduction method is proposed. Firstly, NLMS is used to filter the signal, which is used for primary noise reduction. Then the signal is decomposed into a series of Intrinsic Mode Functions(IMFs) by EEMD, and the kurtosis value, root mean square value and sample entropy value of each IMF are calculated respectively. The appropriate one is selected according to the comprehensive index. Finally, the signal is reconstructed and the Hilbert transform is performed on the reconstructed signal to obtain the envelope spectrum, and the fault characteristic frequency is extracted. Simulation and experimental results show that the method can effectively reduce noise and successfully extract fault features.

As the main parts of rotating machinery, rolling bearing is prone to failure due to its harsh working environment. Aiming at the problem that the early fault features of a rolling bearing are easily submerged by noise and difficult to extract, a fault diagnosis method based on weighted variational mode decomposition (WVMD) and cyclic spectrum slice energy (CSSE) is proposed. Firstly, the signal is decomposed into intrinsic mode functions (IMFs) by VMD and the sparsity is used to measure the amount of information contained in each IMF, and all IMFs are weighted and reconstructed to suppress the noise interference components in the signal. Secondly, the advantage of the CSSE which can accurately mediate the fault information is used to analyze the reconstructed signal, and then the fault characteristic frequency of the reconstructed signal is extracted. Finally, the bearing simulation signal and outer ring fault signal are used to verify that the proposed diagnosis method can effectively extract the early fault features of rolling bearing.

Planetary gear systems often work under severe working conditions, causing gear failures to occur frequently. When the gear fails, the dynamic characteristics of the system will be altered with the excitation of nonlinear parameters. To study the dynamic response of the planetary gear system with gear fault, a nonlinear dynamic model for both health system and faulty system containing the backlash, time-varying mesh stiffness and manufacturing error is put forward. Then, the backlash and rotation frequency are taken as the excitation parameters to study the nonlinear characteristics of the faulty and healthy system through the global bifurcation diagram. At the same time, the local characteristics of the two systems are analyzed via the Poincaré maps and phase diagrams. The analysis results show that the faulty system has a more complex movement as the excitation parameters variation. This research can provide a reference for the dynamic design of planetary gear systems.

Rolling bearing is a kind of easily damaged mechanical equipment. The quality of rolling bearing is related to the normal operation of the equipment. Because the resonance demodulation method is susceptible to noise interference, and the band-pass filter parameters are largely dependent on personal experience selection. This paper proposes an analysis method based on the combination of Ensemble Empirical Mode Decomposition (EEMD) and the selection criterion of kurtosis-cross-correlation coefficient. Firstly, the vibration signal is decomposed by EEMD to get intrinsic mode functions (IMFs); Secondly, since the decomposed IMF components will produce mode aliasing, two criteria of cross-correlation coefficient and kurtosis are introduced to extract effective IMF components for signal reconstruction; Finally, the reconstructed signal is subjected to Hilbert transform and envelope analysis. Compared with the resonance demodulation analysis method, the EEMD decomposition method is selected to replace the band-pass filter to reduce the noise of the signal, which enhances signal to noise ratio and makes the fault characteristics more obvious. The experimental signal analysis results of rolling bearing faults show that a refinement of methodology presented in this article can effectively extract the fault characteristics of rolling bearing, and has more advantages than traditional envelope analysis methods.

The vari-speed drive, which transmits motion and torque through planetary gear and centrifugal rotor, plays an important role in mechanical transmission system. It is mainly used in ships, rolling mills, automobiles and other fields. The vari-speed drive can improve the dynamic performance and economy of the vehicle through the continuous change of transmission ratio. In this paper, an innovative planetary centrifugal vari-speed drive is presented. The structure and working principle of the device are introduced. The multi-body dynamics modeling of the vari-speed drive is established to obtain the working characteristics and vibration characteristics of the mechanism under different working conditions. A compression spring is installed between the centrifugal rotors, and the influence of spring on the system performance is analyzed. This paper can provide a theoretical basis for the design of planetary stepless transmission.

The track is the main failure part of vari-speed drive. The failure damage of the track disc will generate huge vibrations. The fatigue fracture became a problem during the normal working process of these machines. However, the steel roller has a higher deformation resistance capacity. The large contact stress of the steel roller will generate larger amplitude of cyclic stress and decrease the life of vari-speed drive. In order to improve the reliability of the vari-speed drive and increase the fatigue life, the rubber roller is used to replace the steel roller. A 3D model of the vari-speed drive is established by using Solidworks. The contact analysis of track is calculated by using Adams. The dynamic stress is analyzed by using the rain-flow counting method, which can determine the amplitudes and mean values of counted cycles. According to the assumption of a linear Palmgren–Miner hypothesis of damage accumulation and typical fatigue characteristics of the material, the fatigue failure life of the roller is calculated. The results show that the contact force of the steel roller is more than that of the rubber roller. The contact stress of the steel roller is much larger than the rubber roller. The fatigue life of the steel roller is less than that of the rubber roller. This paper presents a new method for solving the fatigue failure of the vari-speed drive.

Accurate state of health (SOH) estimation of the lithium-ion battery plays an important role in ensuring the reliability and safety of the battery management system (BMS). The data-driven method based on the selection of degradation features can be effectively applied to SOH estimation. In practice, lithium batteries often work in complex discharge conditions, but they are charged under constant current (CC) conditions. Therefore, the suitable degradation features of the battery are extracted in this work for accurate SOH estimation. First, the degradation features are summarized and extracted from the CC charging data. Second, the Pearson correlation coefficient is utilized to quantify the relationship between the extracted degradation features and the battery SOH, thus determining the most influential degradation feature. Finally, the long short term memory (LSTM) is used for model training and SOH estimation based on the selected feature. The results show that LSTM model can give reliable and accurate SOH estimation with $$R^2$$ R 2 of 1 and lower mean absolute error (MAE) and maximum error (MAX).

The finite element model of a dual-rotor system (coaxial structure) was established by Ansys entity model. The inner rotor is supported by three bearings while the outer rotor is supported by two bearings. The outer rotor connects the inner rotor by an intermediary bearing. The critical speed Characteristic of the dual-rotor system was calculated by selecting the inner rotor and outer rotor separately as the main excitation source. The harmonic response analysis of the dual-rotor system with mass unbalance were analyzed in order to study the influence of mass unbalance of the rotor system. The results indicate that the outer rotor with mass unbalance is more likely to cause resonance than the inner rotor with mass unbalance, but the resonance amplitude is relatively small. It is possible to reduce the harmonic resonance frequency in the dual-rotor system by improving the design of the outer rotor. Harmonic resonance amplitude gradually increases with the increase of excitation frequency.

The assembly precision of high pressure compressor (HPC) multi-stage rotors is critical to the healthy and high-performance operation of aeroengine. The traditional trial assembly method usually requires multiple disassembly and assembly of rotors to ensure assembly precision, which is cumbersome and may cause damage to parts. It is important to predict and optimize the stacking precision of multi-stage rotors to improve the assembly quality and reduce aeroengine operational failure. This paper proposes a novel method for stacking optimization of aeroengine multi-stage rotors based on 3D deviation prediction model. First, the three-dimensional deviation propagation and accumulation process in the stacking process of a four-stage rotor is deduced using the coordinate transformation method, and the 3D deviation prediction model is established to derive the concentricity and perpendicularity of rotors under different bolt hole phase combinations; Second, the Gaussian distribution is used to simulate the radial and axial runout data of the upper end surface of each stage of rotor, and the center and unit normal vector of the upper end surface of rotors are obtained by the least squares method, then the predicted values of concentricity and perpendicularity of rotors under different bolt hole phase combinations are calculated through the 3D deviation prediction model; Third, a dual-objective integer optimization function is established, and the optimal installation phase and the corresponding predicted values of concentricity and perpendicularity of rotors are obtained. The proposed method can predict the concentricity and perpendicularity of rotors and optimize the stacking phase accurately, so as to achieve the target stacking precision requirements through a single assembly, which can provide theoretical guidance for the actual stacking process of aeroengine multi-stage rotors.

Accurate identification of fault types is an important part of sensor fault diagnosis. Therefore, a sensor faults identification method based on Adaptive Particle Swarm Optimization Support Vector Machine (APSO-SVM) is proposed in this paper. Firstly, the appropriate Time-domain parameters are extracted from the fault data to realize feature extraction and dimension reduction. Then the Particle Swarm Optimization (PSO) algorithm is improved by adjusting the particle velocity with weight and introducing mutated particles, so as to improve the optimization ability of the algorithm and to optimize the parameters of Support Vector Machine (SVM). Finally, the optimized model is used to identify the sensor faults, and compared with other advanced algorithms, the results show that the proposed method can identify the sensor faults more accurately.

Due to the complex internal structure of the sensor, the corresponding fault causes are also diverse. Once a fault occurs, the cause of the fault is difficult to determine. This paper proposes a sensor fault correlation analysis method combining fuzzy Petri net (FPN) and Apriori algorithm. First, obtain the typical fault type waveform of the sensor according to the method of fault simulation, calculate its fault waveform characteristics, find out the residual between it and the normal waveform characteristics, and normalize the residual; then, use the modeling method of FPN to establish the correlation analysis model between fault types, fault characteristic indicators and fault modes; finally, the establishment of model weights and transition threshold parameters is achieved through the Apriori algorithm based on association rules. The maintainer can analyze the fault correlation of the sensor through the abnormal waveform of the sensor to preliminarily judge the fault cause, to achieve the purpose of improving the efficiency of maintenance.

At present, commercial software of the multi-body dynamics is widely used in the research of vehicle ride comfort simulation and optimization. This paper reviews some literatures on vehicle ride comfort optimization based on ADAMS, and focuses on the research based on suspension models. The suspension rigid-flexible coupling models and the simulation research about optimizing suspension model parameters to achieve multi-objective optimization are the main areas of concern. Finally, the paper is summarized and the future trend of ADAMS applied to the simulation and optimization of vehicle ride comfort is prospected.

Vehicle handling stability is one of the most important performance of automobile. At present, ADAMS software is widely used in the study of vehicle handling stability. Taking suspension as research objects, this paper reviews the progress of ADAMS software in the study of vehicle handling stability, introduces what evaluation indexes scholars have selected and what simulation tests have been carried out to study the vehicle handling stability. The results of simulation experiments of handling stability under different working conditions are summarized, and the contribution of these methods to improving handling stability is analyzed, which provides a useful reference for scholars in related research. Finally, the paper is summarized, and the future trend of ADAMS software applied to the simulation of vehicle handling and stability is prospected.

Nine Degrees of Freedom Motion Simulator (9-DOF-MS) is the key equipment for calibration of Camera-type Rendezvous & Docking Sensor (CRDS) in spacecraft space rendezvous & docking Guidance Navigation and Control (GNC) sub-system, and it must be high with relative position precision. For meeting this demand, the components of errors impacting this system’s integral indexes are analyzed systemically in this paper. At first, the relationship and interactions among the components of system errors are analyzed. Then the error model is built. By decomposing and redistributing the systematic precision index, 9-DOF-MS designed fulfils the precision requirements.

In recent years, the necessity of implementing sensor-based process condition monitoring (CM) in additive manufacturing (AM) has attracted the attention of many foreign governments and academic institutions. To verify the feasibility of the condition monitoring on additive manufacturing, an experiment is carried out. The experiment focuses on the abnormal status of the 3D printer to explore the relationship between the printing signals and the printing quality. There are two methods are used to process the experimental data. One method is the Short-Time Fourier Transform (STFT), which is used frequently used in this report. Its processing result indicates that the signal changes in both the time domain and the frequency domain. The other method is the Mean function. By the comparison of the two methods, the mean function turns out to be better than STFT at presenting the differences in detail and proving the consistency of the signals and the 3D printer features. This experiment lays a firm foundation and points out the directions for future research, such as mathematical simulation, etc.

Acoustic sensing is a non-destructive technology that plays an essential role in condition monitoring. For high-quality data collection, condition monitoring relies on various sensing methods that further complicate the wiring of the system. Moreover, weak signals such as evanescent waves carrying valuable information are usually hard to capture. With the emerging field of metamaterials, such issues could be optimized and solved. This paper presents a metamaterial that is designed by two kinds of typical unicells, purely geometrically, with the aim to enhance the acoustic signal without any external power source. As a result of transmission through the designed metamaterials, the acoustic pressure level at a particular range of frequency is efficiently enhanced. Furthermore, the frequency shift of the enhancement is achieved by altering specific structural parameters, which demonstrates its tunable characteristics. This study intends to provide ideas for the design of acoustic metamaterials for applications such as remote sound measurement, energy harvesting, fault diagnosis, etc.

Drone technologies are widely used for various purposes in many fields. However, the onboard imaging platform is severely compromised by low-frequency vibration during flight, which cannot be suppressed by a general vibration isolation method, leading to poor image quality and failure to fly. In this paper, a novel cylindrical mechanical metastructures (CMMS) vibration isolator was proposed to overcome the drawback of general vibration isolators based on the vibration of the drone imaging platform. By using additive manufacturing, a prototype of the CMMS was fabricated, and experiments were carried out to verify the mechanical properties. In research results, the CMMS isolator has been observed to suppress the full-band frequency vibration of the drone, enabling the imaging platform to operate in a stable environment.

Piezoelectric bimorph cantilever is a typical collecting structure for vibration energy, however it can not adapt to the low frequency and random of vibration excitation in natural environment. In this paper, the physical model of linear vibration system of piezoelectric bimorph cantilever is analyzed, where the piezoelectric energy harvester with multi-cantilever is designed for it’s disadvantages. By increasing the cantilever beam with different natural frequencies, the energy harvester has the characteristics of broadband. By measuring the dynamic strain under sweep excitation in the simulation of Comsol, Compare the output voltage and working bandwidth between the piezoelectric bimorph cantilever and the piezoelectric energy harvester with multi-cantilever, verify the broadband characteristics of the latter. This paper also designs the rectifier circuit, to convert alternating current from the energy harvester into direct current.

Pipes, like blood vessels, play an important role in industry and people's life. Once a pipeline leak occurs, it will bring huge economic losses and even cause serious accidents. Usually, pipeline leak monitoring is carried out by manual inspection or the installation of many sensors, which have great limitations. In this paper, an intelligent mobile robot is proposed for more effectively monitoring the large-scale pipeline systems. Equipped with microphone the robot can be set with the inspection paths, realize the monitoring of leakage anomaly by power spectrum and correlation analysis of sound signal according to the collected real-time data, and pinpoint the location of leakage by calculating the sound pressure. The experimental results show that this method is convenient and effective in a typical industrial environment.

Additive manufacturing has the characteristics of gradual accumulation of materials in the manufacturing process. It is often superimposed layer by layer in the process of material physical shape change, which may be accompanied by hot melting, liquid material solidification, particle sintering and other processes. Due to the influence of physical environment, machine state, manufacturing principle and other factors in the whole process, performance defects of parts may occur. The traditional monitoring methods such as vision, optics and CT tomography have limitations, or can only observe the defects on the outer surface, or it is difficult to find the defects in time in the processing process, or the micro defect identification accuracy is not enough. A series of research on acoustic emission detection technology, due to the high sensitivity to high-frequency signals, can observe various phenomena of the machine itself in the processing procedure, and monitor the spatial micro faults of the whole part in the process of parts made of additive materials.

Industrial robots are widely used in modern factories. Robot faults and abnormal working state will lead to the shutdown of the production line inevitably. Robot condition monitoring can improve production capacity. However, due to the changes of robot in dynamic working state, this is a challenge. This paper presents a methodology of condition monitoring for industrial robots using vibration signals. The main purpose of this paper is to identify the occurrence of the fault and its different degrees. Experiments was performed on a 6-dof industrial robot (IR). Firstly, the Frequency Response Function of the IR was obtained by using the Experimental Modal Analysis method. And the characteristic frequency in each axis was found. Then, based on the Short-time Fourier Transform analysis method, the vibration data under normal conditions and different degrees of abnormal working conditions were analysed. In some characteristic frequency bands, the amplitude will increase with the increase of the binding force at the joint. Finally, this trend was further verified by the calculation of RMS value. The results show that the proposed frequency domain and model analysis method can monitor the operating condition of industrial robots.

Vehicle reliability, battery life, and increased costs due to increased system complexity will hinder the marketization of hybrid electric vehicles. Improving vehicle reliability is the basis for improving product safety and performance. The dissertation studies the failure modes and failure laws of hybrid electric vehicles, and uses hybrid electric buses of electric vehicle demonstration operation companies as test objects to conduct road assessment tests to verify the matching and optimization of the entire vehicle and improve its performance and reliability. Develop a fault monitoring and acquisition analysis system for electric vehicle battery systems and power switching systems, and establish a mathematical model of the failure mode to summarize the rules for the maintenance and use of electric vehicles and the operation of electric vehicles.

This paper introduces a new two-dimensional condition-based maintenance model for complex and repairable machining systems like computer numerical control machining tools. A joint distribution of condition and reliability indicators is constructed by using copula. The maintenance threshold is set on the cumulative hazard rate conditioning on intensity of work. A numeric example with assumed settings is provide to demonstrate the relationship between the maintenance threshold and expected cost rate. This is the first model jointly considering condition and reliability indicators in maintenance area; and being benefited by the features of copula, this model can be easily extended to model dependences among multiple indicators in practice.

Since the discovery of the piezoelectric effect, the application of piezoelectric technology has been increasing. The rotary valve has precise and controllable characteristics for the contact dispensing process, and has become an indispensable part of the dispensing industry. For a long time, the working mode of piezoelectric ceramic valve controller is relatively single. The traditional point mode, line mode and single channel can no longer adapt to the development of industrial automation. With the increasing complexity of dispensers, customers have put forward more and more customized requirements. In order to better adapt to the development of dispensers, reduce the difficulty of control, and shorten the development cycle, a more flexible and convenient controller system is needed.

This article describes a method for measuring the positioning accuracy of dispenser and the compensation method. This method uses customized high-precision calibration boards, CCD imaging technology and vision algorithms to measure the XY two-dimensional positioning accuracy of the machine, and compensate the measured error results to the motion system to improve the positioning accuracy of the machine. Since the deviation is calculated by visually grabbing the Mark points on the calibration board, it is lower in cost, easier to operate, and more efficient than the traditional laser interferometer measurement method. Furthermore, the traditional laser interferometer can only perform single-axis compensation, while the calibration plate method can compensate XY at the same time, which truly improves the positioning accuracy.

This paper mainly describes the structure design of a high-speed and high-precision operation platform in the automatic dispenser industry. Firstly, through the analysis of the development of domestic automatic dispenser and the mechanism of automatic dispenser, the importance of X/Y/Z high-speed and high-precision operation platform is summarized. Referring to the relevant information, this paper analyzes and compares the existing automatic dispenser mechanisms in the market, describes their advantages and disadvantages, and makes its own improvement on the basis of these mechanisms. This design mainly includes the design of x-axis module, Y-axis module and z-axis module, which mainly involves the material comparison and parameter design of casting base, selection of linear motor, linear guide rail, servo motor, precision grinding screw rod, grating ruler and photoelectric sensor, design parameter check, material comparison of x-axis beam and tolerance analysis of important parts. Furthermore, this design chooses to use SolidWorks three-dimensional modeling software to complete the X\Y\Z high-speed and high-precision operating platform structure design, 2D assembly drawing and drawing of its main parts.

Deep learning is widely used in the field of rotating machinery fault diagnosis. However, manually designing the neural network structure and adjusting the hyperparameters for specific fault diagnosis task are complex and requires a lot of expert knowledge. Aiming at these problems, Differentiable Architecture Searched Network with Tree-Structured Parzen Estimators (DASNT) is proposed for fault diagnosis. Differentiable Architecture Search (DARTS) is utilized to automatically search network structure for specific fault diagnosis task. Tree-Structured Parzen Estimators (TPE) is utilized to optimize the hyperparameters of the network searched by DARTS, which can further improve the fault diagnosis accuracy. The results of comparison experiments indicate that the network architecture and hyperparameters optimized by DASNT can achieve superior fault diagnosis performance.

With the rapid development of power electronics technology and the proposal of intelligent ships, electric propulsion systems on ships are becoming more and more widespread. As the power source for ship navigation, timely and accurate diagnosis and prediction of faults of electric propulsion system play a vital role in the operation safety of ships. This paper summarises the common faults of electric propulsion systems, reviews the latest developments and applications of fault diagnosis techniques based on fault signal analysis in electric propulsion system fault diagnosis, and discusses the advantages and disadvantages of typical methods in the light of the latest literature and current research problems. The paper concludes by proposing future trends in fault diagnosis and prediction for ship electric propulsion systems.

As one of the core components of the gas turbine compressor, the blades are prone to cracks and even breakages when working in harsh environments such as high pressure, high speed, and alternating heavy loads. The generation and propagation of cracks will change the natural vibration characteristics of blades. Therefore, it is necessary to research the influence of different types of crack faults on the natural vibration characteristics of blades, which can provide a theoretical reference for the blade crack fault diagnosis. Firstly, the finite element model of the healthy gas turbine compressor blade was established, and the modal parameters were analyzed; secondly, in order to verify the accuracy of the finite element model, the blade modal experiment platform was built to carry out the modal experiment of the healthy blade based on the moving hammer method, and the influence of different sensor installation methods on the modal test results was analyzed; finally, the modal parameters of the blade with different types of crack faults were analyzed based on the finite element model, and the mapping relationship between the crack faults and the natural vibration characteristics of the blade was established. The results show that crack type variation would affect the natural vibration characteristics of the blade, which will lead to modal coupling and modal shape switching characteristics. As a result, the same order vibration mode of blade has different mode shapes at different crack positions and shapes. The results of this paper may serve as a theoretical basis for the diagnosis of compressor blade crack faults based on vibration characteristics.

As the key component of the steam turbine, the steam turbine blade needs to work in a complex and rigorous operating environment, which easily leads to blade cracks or even fractures and other faults. Excessive vibration is one of the main causes of blade failure, which may affect the safe and stable operation of the equipment. Therefore, it is significant to detect and analyze blade vibration characteristics. Blade tip-timing (BTT) technology has the advantages of non-contact and simple installation, which is widely used in online blade vibration monitoring of turbomachinery. In this paper, the research of using BTT technology to measure the vibration characteristic parameters of the last stage moving blade of a steam turbine with integral shroud and snubber is carried out. Firstly, a finite element model of the last stage blade is built, the stress distribution and mode shape of the blade are obtained through simulation analysis. Secondly, the blade vibration measuring experiment is accomplished on a dynamic balancing test-bed, and the synchronous vibration parameters such as resonance speed are calculated correctly under lowing speed working condition based on BTT technology. Furthermore, the strain gauge method is used simultaneously to verify the accuracy of the measurement results. The resonance frequency and engine order of the blade are measured successfully. The analysis results show that parameters such as resonance speed identified by BTT method are consistent with that measured by strain gauge method. The research results prove the effectiveness of BTT technology in measuring the vibration parameters of the last stage blade with integral shroud and snubber, which can provide reference for design rationality verification and vibration characteristics detection of the steam turbine blade.

The paper presents a novel magnetic coupled piezoelectric energy harvester for supplying an on-rotor sensing (ORS) IoT device. It operates based on a rotating piezoelectric beam and a fixed permanent magnet placed remotely. When the free end of the beam rotates passing through the magnet fixed on stators, an impulsive magnetic force will excite the beam to vibrate and produce electricity. A lumped electromechanical model is calibrated by fixed beam tests and subsequently used to competently evaluate the basic configuration and performance of the harvester. Simulations has verified that the harvester can performs outstandingly not only in the resonance frequency band of the beam but also the frequency range lower than half of the resonance frequency, thanks to the impulsive excitations produced by the when the beam tip passing the fixed magnet. Simulation studies also shows that this harvester can operates for both horizontal and vertical rotor systems.

This paper addresses the problem of reconstructing depth and silhouette images of wind turbine from its photos of multiple views using deep learning approaches, which aims for wind turbine blade fault diagnosis. Some previous multi-view based methods have extracted each photo’s silhouette and combined them into separate channels as the input of convolution; others use LSTM to combine a series of views for reconstruction. These approaches inevitably need a fixed number of views and the output result is divergent if the order of the input views is changed. So, we refer to a network, SiDeNet (Wiles and Zisserman, Learning to predict 3d surfaces of sculptures from single and multiple views. Int J Comp Vision, 2018), which has a flexible number of input views and will not be affected by the input order. It integrates both viewpoint and image information from each view to learn a latent 3D shape representation and use it to predict the depth of wind turbine at input views. Also, this representation could generalize to the silhouette of unseen views. We make the following contributions to SiDeNet: improving the resolution of predicted images by deepening network structure, adopting 6D camera pose to increase the degrees of freedom of viewpoint to capture a wider range of views, optimizing the loss function of silhouette by applying weights on edge points, and implementing silhouette refinement with point-wise optimizing. Additionally, we conduct a set of prediction experiments and prove the network’s generalization ability to unseen views. Evaluating predicted results on a realistic wind turbine dataset confirms the high performance of the network on both given views and unseen views.

To solve the problem of maintenance lag caused by long-term ocean voyage, a health assessment method is proposed based on the testability model for diesel engine. Firstly, the testability model is applied to generate the “fault-test” correlation matrix and accurately describe the interaction of each module and the fault signal propagation in the system structure; Then, the current health state corresponding to the bottom fault mode can be quickly deduced by using the model to infer the test information entropy; Finally, the health status of the whole diesel engine is evaluated by mapping to the health status of the upper structure through the support vector machine algorithm. The method can be used to determine the maintenance requirements in advance and improve the accuracy of fault prediction.

Cutting tool plays an important role in modern manufacturing industry, however, tool wear is unavoidable during machining which could reduce the efficiency. Aiming at studying an appropriate and efficient tool condition monitoring method to improve the accuracy of finished parts, the roughness of the turned surface, a novel On-Rotor Sensing (ORS) is installed on the rotating workpiece to obtain vibration signals. To get an in-depth understand of the vibration data, a multi-degree-of-freedom (MDOF) system consisted of spindle, chuck and workpiece is established and its multi-mode natural frequency is obtained by finite element model (FEM) method. It is found that the dynamic response of the spindle rotor determines machining accuracy in the turning process and shows that the first several modes in the frequency range within 2000 Hz are the main responses of the system, which can be effectively captured by the ORS. Especially, the spring stiffness is calibrated based on the FEM results and the accuracy of the dynamic modal responses of this model are verified when the mass of the workpiece decreases during the turning process. According to the results, two frequency bands are advocated for ORS based online monitoring of tool wear conditions.

The high requirements on the integrity and operational reliability of the equipment in the process of carrying out tasks is required by the ships and other major equipment. To ensure the long-term safe and healthy operation of their power transmission devices is one of the key links to achieve this requirement. In response to this urgent need, the real-time monitoring and health assessment method of equipment power transmission based on wireless sensor network is researched, and the related software and hardware prototype systems is constructed, after that, the system testing and use verification in actual tasks of ships are carried out. The verification results show that the prototype system can complete the online collection and analysis of vibration and temperature data of the key parts of the ship's power transmission device, and the real-time monitoring and health assessment of the power transmission device condition is effective. As a result of that, the maintenance workload is reduced effectively while work efficiency is improving.

In the fault diagnosis method of data-driven method, it is difficult to obtain fault data and high cost of experiment due to the particularity of special equipment and health condition for a long time at the beginning of operation. Based on the analysis of slow-changing parameters such as temperature and pressure collected under normal operation, this paper establishes signal prediction models under different conditions and puts forward a historical view. The dynamic threshold method of measuring data eliminates the false alarm and improves the ability of early fault detection at the initial stage of equipment operation, and provides a new idea for fault detection under the condition of only normal samples. It provides scientific and accurate support for fault early warning theory and method of special equipment and realizes the direction of special equipment from regular maintenance and preventive maintenance to condition-based maintenance change.

Due to the high bearing capacity and good alignment advantagements, splines are widely used in various mechanical equipment. However, alternating stress can be easily observed in meshing of spline coupling, which leads to fatigue failure of spline pairs such as cracks. The underlying mechanism and suppression method of impact force of spline pair are seldom studied. In order to settle this problem, this paper proposed a new two-mass nonlinear impact dynamic model of spline pair, and studies the generation mechanism, influence law and suppression method of dynamic impact force of spline pair. Based on Hertzian contact theory, this paper establishes the calculation model of dynamic impact force of spline pair firstly, both the coupling relationship between pressure angle, backlash and eccentricity and impact force are taken into consideration. Then, the calculation and suppression method of impact force of spline pair are investigated and discussed. The results show that the impact force gradually increases with the growth of backlash and eccentricity, and local extremum appears. This paper presents a new model to study and suppress the dynamic impact force of spline teeth, it is found that the root mean square (RMS) value of impact force reaches to the smallest when pressure angle is set as 25.