Search

Weak signal features of early bearing faults are interfered by environmental noise, which seriously affects the accuracy of diagnosis results. Moreover, a large amount of data calculation and manual parameter adjustment during model training will affect the timeliness and intelligence of diagnosis results. Aiming at the above problems, an intelligent diagnosis method for motor bearing fault based on music theory knowledge novel generative adversarial networks (MTKGAN) is proposed for the first time. First, the game between generation and discrimination models is used to generate fault samples. The Earth-Mover distance is used to measure the distance between the real and generated distribution. The method generates and enhances weak signal features, and the interference of environmental noise on the signal is effectively solved to improve the accuracy of fault diagnosis. Second, inspired by music theory knowledge, the fault feature affine invariance migration method based on adaptive chord transformation strategy is proposed. The problems of Big Data training and manual parameter adjustment are effectively solved to improve the timeliness of fault diagnosis. Finally, the advantages of MTKGAN in early fault diagnosis of motor bearings are verified by comparing the public dataset and motor bearing fault experiment platform with the existing advanced methods.

High dynamic response has always been one of the goals of the surface-mounted permanent magnet synchronous motor (SPMSM) servo system. To this end, this article proposes an antidisturbance enhanced finite-time composite control strategy for the SPMSM servo system. First, a finite-time controller is designed to enhance the rapidity of the SPMSM speed loop. Second, an antidisturbance enhanced finite-time disturbance observer is proposed, which optimizes the disturbance compensation channel and accelerates the disturbance compensation rate. Meanwhile, this method decouples the parameters of the controller and disturbance observer, making it easier to tune the parameters. Finally, combining the controller with the disturbance observer, the proposed composite control scheme has better dynamic performance than the traditional finite-time composite controller. The experimental evaluation results verify the superiority of the proposed method.

When the interior permanent magnet synchronous motor drive system is affected by both dc and ac disturbances, the estimation error of ac disturbance cannot be converged to the steady-state value by the traditional linear extended state observer (T-LESO). At the same time, the tracking and antidisturbance ability of the traditional linear active disturbance rejection controller(T-LADRC) to dc disturbance is insufficient due to the influence of observer bandwidth frequency. To solve the above-mentioned problems, an improved LADRC combining a quasi-proportional resonance (QPR) controller and disturbance differential compensation LESO is proposed in this article. In this controller, the linear state error feedback in the T-LADRC is replaced by QPR, so that the ac disturbance to the system can be effectively suppressed. Based on T-LESO, the derivative of the disturbance is introduced, whose estimation range of LESO for dc disturbance is broadened without increasing the observer bandwidth, thereby avoiding system oscillation caused by excessive observer bandwidth. Finally, the validity and availability of the proposed method are proved through theoretical analysis and experiment.

In the interior permanent magnet synchronous motor (IPMSM) drive system, the conventional linear active disturbance rejection controller (C-LADRC) is affected by the bandwidth, which will lead to insufficient tracking and disturbance compensation capability. To solve the above problem, a doubly-fed differential compensation LADRC (DFDC-LADRC) combined with input feedforward differential compensation and output feedback differential compensation signals is proposed in this article. The DFDC-LADRC takes the total disturbance error as the basis for adjusting the total disturbance of linear extended state observer (LESO), while an output feedback differential compensation signal is introduced so that the total disturbance received by the system can be effectively estimated and compensated by LESO. In DFDC-LADRC, the actual value of the output signal is directly taken as the feedback signal of the linear state error feedback (LSEF), and the input differential compensation signal is given simultaneously, so as to reduce the observation burden of LESO and enhance its tracking capability. Finally, the correctness and effectiveness of DFDC-LADRC are verified by theoretical analysis and experiments.

Traditional sliding mode observers (SMOs) are no longer able to meet the high standard requirements for interturn short-circuit (ITSC) fault diagnosis in permanent magnet synchronous motors (PMSM). The first reason is the system jitter problem caused by traditional SMO control functions. The second reason is the amplitude attenuation and phase delay problems caused by low-pass filters. A diversified diagnosis strategy based on the novel SMO for PMSM ITSC fault detection, localization, and severity estimation is proposed to address the above issues. First, the system jitter is minimized by introducing a nonlinear saturation function. Second, a diagnostic method based on second-order generalized integrator (SOGI) SMO is designed. This method can realize the undistorted extraction of back electromotive force signal. Importantly, SOGI can phase shift the filtered signal. The coupled signal used for fault severity estimation is easier to obtain. Finally, the effectiveness of the method is verified under experimental working conditions of system steady-state, dynamic, and load disturbances. The experiment shows that the proposed method can complete current calculation within 0.1 s and fault diagnosis within 1 s.

To address the problem of gear clearance causing the limit cycle oscillation at the positioning end, further deteriorating the positioning accuracy of the servo system, and even causing tooth breakage in severe cases, a backstepping control (BC) method based on shaft torque observer (STO) is proposed in this article for a fully closed-loop gear transmission servo system. First, a gear transmission servo control system model is established, which includes a permanent magnet synchronous motor model, a backlash model, and a gear transmission system model. Second, the description function method is utilized to analyze the limit cycle oscillation mechanism of nonlinear systems. Next, speed loop and current loop BC subsystems are proposed. To avoid the problem of limit cycle oscillation and decreased positioning accuracy caused by backlash nonlinearity in the fully closed-loop gear transmission servo system, a STO is proposed and real-time observation and online compensation of shaft torque are carried out, which effectively eliminates the limit cycle oscillation at positioning end. Ultimately, a 2.0-kW gear transmission servo system is used as a case study, the validity of the proposed BC-STO is tested via experimental results, the limit cycle oscillation is effectively eliminated, and the positioning accuracy is significantly improved.

Synchronous reluctance motors (SynRM) have a highly nonlinear flux-current mathematical relationship, making it difficult to obtain accurate inductance information. The Kalman-filter-based position estimation method is designed by using a single active back electromotive force (AEMF) model. When inductance parameter mismatch occurs, the estimation performance of the AEMF observer designed by using the single-model resonant Kalman filter (SM-RKF) is more sensitive to inductance parameter deviation as the load increases. To inhibit the effects of inductance parameter mismatch on the SynRM sensorless control system, the position estimation method based on a generalized multiple-model resonant Kalman filter (GMM-RKF) is proposed. First, a new AEMF model of the SynRM is derived at the maximum inductance-oriented coordinate frame. Second, the model set of GMM-RKF is built by RKF1 and RKF2 observers, which are derived from the two AEMF models. The two observers exhibit different parameter sensitivity characteristics at different current vector angle ranges. In comparison with SM-RKF, GMM-RKF has superior robustness against inductance mismatch since it combines the areas in two RKFs where the position estimation is robust to inductance mismatch. The effectiveness of the proposed sensorless control scheme is verified through an experimental prototype.

In the interior permanent magnet synchronous motor drive system, the conventional linear active disturbance rejection controller (LADRC) has coupling problems with tracking and antidisturbance under different operating conditions and various complex disturbance conditions due to the influence of bandwidth. To solve abovementioned problem, a series connection leading correction LADRC (SCLC-LADRC) is proposed in this article. As the core of SCLC-LADRC, the SCLC linear extended state observer (SCLC-LESO) is composed of the total disturbance transfer function of the conventional linear extended state observer and a series connection leading correction network. As a result, the estimation range of the LESO for total disturbance is expanded, and the problem of system oscillation caused by excessive observer bandwidth is also solved. Finally, the validity and availability of the proposed method are proved through theoretical analysis and experiments.

The small capacitor drivers use film capacitors instead of electrolytic capacitors in the dc-link, which will improve the life and reliability of the drivers. However, the dc-link voltage of the small capacitor drivers fluctuates seriously, and the ac component in it causes |(ω+6kωg)/ω|th and |(ω−6kωg)/ω|th harmonics in the extended back electromotive force. These harmonics will increase the estimated position error in the sensorless driver. A harmonic suppression method based on adaptive synchronous rotating frame transformation (ASRFT) is presented in this article to improve the control accuracy of sensorless drivers with small capacitors. First, the motor current harmonics caused by harmonics in the dc-link voltage are analyzed, and the relationship between these harmonics and the estimated sensorless position error was derived. Second, ASRFT and fixed synchronous rotation frame transformations can be used to extract current harmonics in corresponding harmonic coordinate systems. Then, the voltage compensation component generated by the harmonic compensation link will be injected into the driver to suppress the current harmonics extracted in the corresponding coordinate system. Finally, the controller parameters of the proposed method are designed. Using a 2 kW permanent magnet synchronous motor sensorless driver with a small capacitor, the proposed method is demonstrated to be effective.

Under certain working conditions, the permanent magnet synchronous motor (PMSM) traction system of urban rail transit shows negative impedance characteristics, which is easy to cause voltage resonance of DC-link and lead to instability of the system. In order to ensure the stability of the system, a DC-link oscillation suppression method based on q-axis voltage compensation is proposed. Firstly, the PMSM traction system and oscillation mechanism of DC-link are analyzed. Secondly, a stable control method based on q-axis feedforward decoupling voltage compensation is adopted to realize the decoupling control of PMSM and the suppression of DC-link oscillation. Finally, the method is verified on a 180 kW PMSM traction system, simulation and experiment verify the effectiveness of the DC-link oscillation suppression method.

Aiming at the problem of motor control performance degradation caused by the cross-coupling and low switching frequency condition in the high-power direct-drive permanent magnet synchronous motor (PMSM) control system, a robust current control of PMSM based on internal model control (IMC) and hybrid pulse width modulation (PWM) is proposed and applied to the control of a MW-level direct-drive PMSM in electric locomotive traction system. Firstly, combined with the characteristics of the low switching frequency of the electric locomotive inverter, a hybrid PWM strategy is adopted. Secondly, the IMC of PMSM is designed, and the IMC is adopted to solve the problems of more control parameters and cross-coupling of direct-drive PMSM in the d-qaxis under low switching frequency, which improves the robustness of the whole locomotive system. Finally, the experiments have been carried out on a 1.2 MW direct-drive PMSM for traction system in electric locomotive, the correctness and effectiveness of IMC and hybrid PWM strategy are verified.

When estimating the active back electromotive force (AEMF) of synchronous reluctance motor (SynRM) at medium and high speed, the traditional Kalman filter (KF) will produce large phase delay due to the low-pass filter characteristics. A novel speed sensorless control scheme of SynRM based on resonant Kalman filter (RKF) is proposed in this article. First, the AEMF observer based on KF is designed and the bandwidth performance is analyzed. To solve the phase delay problem caused by KF, a generalized integral resonance perturbation estimator (GIRPE) is introduced to the framework of KF and the proposed RKF is formed. The estimation error of AEMF is compensated through GIRPE in the current error feedback loop of KF observer. In this way, the high-precision position estimation at high speed range is obtained. Finally, an improved dynamic position compensator is proposed to enhance the position estimation performance under dynamic case. The effectiveness of proposed scheme is validated at a 1.5-kW SynRM drive, and the computational burden of the proposed method is assessed quantitatively.

• Various factors are corrected with durability and reliability of photovoltaic backsheet. • Detection methods of insulation deterioration are summarized innovatively. • Emerging novel materials and structures are summarized in photovoltaic cell. The insulation degradation in polymeric backsheets has been identified as a main cause of catastrophic accidents induced by short circuit or ground faults in photovoltaic module. To ensure quality, the photovoltaic industry is therefore faced with urgent demand in discovering degradation mechanisms. Moreover, the development of environmental-friendly backsheets and the establishment of backsheets recycling specifications are vital to fulfilling the requirements of a future reliable photovoltaic system with improved productivity. In this review, we innovatively summarize the detection methods of insulation deterioration from the viewpoints of spectroscopic, thermal and mechanical approaches. The corresponding ambient conditions in measurement and their accelerating effect on the degradation of photovoltaic backsheets are discussed. Subsequently, emerging novel materials and structures for enhancing insulation properties, anti-aging performance and optical-electrical energy conversion efficiency of photovoltaic cell are summarized. It offers a comprehensive strategy to design materials with optimal structures in photovoltaic module for a future zero-carbon energy system. [Display omitted] [ABSTRACT FROM AUTHOR]

Purpose: This paper aims to propose a novel multiple transient modeling scheme for the 12-pulse phase-shifting reactor (PSR) rectifier to enhance the efficiency of full-cycle design evaluation. Design/methodology/approach: The detailed time-domain method is adopted to model the rectifier at the behavioral layer. The diode bridges/transformer model at the architecture layer is established by using the switch function and Park transformation. The frequency domain model at the functional layer is derived with the time-varying Fourier decomposition and frequency-shifting. At the component layer, the magneto-thermal characteristics of the rectifier are analyzed with field-circuit and magnetic-thermal coupling methods. A computer-aided design program integrating multiple modeling is also developed for industrial product design. Findings: The function layer modeling is preferred in the initial design stage, making up for the lack of modeling accuracy at the architectural layer and the lack of modeling rapidity at the behavioral layer. The component modeling is irreplaceable for the detailed evaluation in the latter design stage. The multiple modeling scheme based on the four-layer modeling helps the designers achieve high-quality products with a short development cycle. Originality/value: The singular transient modeling cannot cover the needs of different stages in the full-cycle design evaluation. This paper fills this gap with a novel multiple modeling scheme. Meanwhile, the proposed multiple modeling scheme and developed computer-aided design program provide a great convenience for full cycle design evaluation of the 12-pulse PSR rectifier. [ABSTRACT FROM AUTHOR]

In the position sensorless control of permanent magnet synchronous motor (PMSM), the traditional sliding mode observer (SMO) uses the switching function, which presents serious chattering. Otherwise, the traditional SMO uses a low-pass filter to filter the harmonic components in the estimated back-EMF. The low-pass filter usually causes phase delay in the estimated rotor position. These problems restrict the further development of SMO. In this paper, an improved sliding mode observer (ISMO) with suppressed chattering and no phase delay is studied. In the ISMO, a sine saturation function is designed as the control function, and a synchronous reference frame filter (SRFF) is designed to eliminate the phase delay of ISMO. Different from other filters, the SRFF can not only filter out the harmonic components in the estimated back-EMF, but also effectively makes the phase delay almost zero. In addition, in order to improve the poor anti-disturbance performance of position sensorless control, a load disturbance observer (LDO) is introduced on the basis of the ISMO. The experimental results prove the effectiveness of the ISMO, and the addition of LDO improves the anti-disturbance ability of the system.

The ultra-local model-based deadbeat predictive current control (ULM-DPCC) is attractive to a permanent magnet synchronous motors (PMSM) drive for its rapid current response and robustness to the parameter changes. However, the accuracy suffers from the model noise and measurement noise in the application of ULM-DPCC to a PMSM drive. To improve the precision of current control under the uncertain Gaussian noise, an augmented fading Kalman filter-based ULM-DPCC (ULM-DPCC-AFKF) for the PMSM drive is proposed. First, the Kalman filter (KF) is employed to estimate the unknown part in ULM-DPCC. Second, to ameliorate the estimation performance under uncertain noise conditions, an augmented fading Kalman filter (AFKF) is proposed to adapt to changes in noise, and an appropriate reference voltage vector is obtained by combining ULM-DPCC and AFKF. In AFKF, the augmented fading factors are derived to make the AFKF observer work optimally, and high-precision current control for PMSM is achieved. Last, the availability of the proposed ULM-DPCC-AFKF strategy is proved via experiment and simulation under the different Gaussian noise.

There are two contradictions between the occurrence of multiple failure types in motor bearings and traditional remaining useful life (RUL) prediction methods: 1) the contradiction between time series data and prediction accuracy among neurons and 2) the contradiction between the large amount of data calculation and prediction speed. Aiming at the above problems, inspired by the music theory knowledge, the method for RUL prediction of motor bearings based on a novel whale optimization algorithm (WOA) is proposed. First, based on the conventional WOA, nonlinear time-varying adaptive weights and convergence factors are proposed. The method solves the problem of decoupling the optimization of time series data due to the overlapping of multiple fault features and improves the accuracy of RUL prediction. In addition, the optimization process for multiple types of faults is complex and time-consuming, so a feature mapping method based on the principle of chord variation is proposed for the first time. The method renders single-fault features into multiple-fault features by mimicking the chord transformations of stringed instruments playing. Finally, the validity of the method is verified on the PRONOTIA public dataset and self-built bearing aging experimental platform. The method improves the prediction speed by 35% compared with the existing advanced methods under a variety of experimental conditions.

The fault diagnosis method based on generative adversarial networks (GANs) has been successfully applied to the early fault detection of motor bearings, and it has effectively solved the problems of small samples, unlabeled sample features, and data imbalance in early faults. The existing methods, however, ignore the weak features of bearing vibration signals when solving the problem of missing early fault samples, which seriously affects the accuracy of diagnosis results. In addition, a large amount of human parameter adjustment work is required during the network model training process. This makes the whole diagnosis process lack objectivity and intelligence. To address these issues, this article proposes an intelligent diagnosis algorithm for early faults of motor bearings based on Wasserstein GAN meta learning (WGANML). First, the method performs feature extraction on the collected bearing vibration signals. Then, the generation of missing samples and the enhancement of weak features are completed by using the game between the generative and discriminative models. Second, to get rid of manual intervention, meta learning (ML) is applied to Wasserstein distance GANs (WGANs) parameter variation training for the first time. Finally, the feasibility and effectiveness of the proposed WGANML algorithm are proved by the open dataset of Case Western Reserve University (CWRU) and the data of the motor bearing fault experimental platform. In addition, compared with the existing advanced methods, the superiority of the proposed method in motor bearing early fault diagnosis is further verified.

To improve the robustness of the conventional finite-time convergent speed controller (FTCSC) against unpredictable disturbances, a finite-time generalized proportional integral observer (FT-GPIO) is proposed based on the state observer to achieve finite-time convergence of the closed-loop control system. First, an FTCSC is devised to improve the speed convergence performance of the induction machine (IM) control system. However, the system will inevitably be affected by multiple sources disturbances, such as friction, load disturbance, parameter mismatch, and torque current tracking error, resulting in performance degradation. To improve the disturbance rejection ability of the IM control system, an FT-GPIO based on multiple sources disturbance estimation technology is designed. Finally, a 1.1-kW IM is used as a case study, and the experimental results are provided to verify the effectiveness of the proposed robust composite finite-time convergent speed controller.

Model predictive direct speed control (MPDSC), with a simple structure and without parameter tuning required, can use only one controller to constrain the speed and current of the motor at the same time. First, a direct voltage vector selection-based three-vector model predictive speed control, simplifying the optimization process, is proposed in this article, which reduces the pulsation of current and speed without increasing the computational burden compared with MPDSC. At the same time, because of the reasonable distribution of voltage vectors in the proposed three-vector method, the quality of the control and modulation has been significantly improved. Second, a multidisturbance Luenberger observer is designed to observe mechanical and electrical disturbances and to make further compensation, which reduced the influence of model mismatch, load torque, and unmodeled dynamics simultaneously and avoided the use of multiple observers, with fewer parameters and easier tuning. The experiment is presented to validate the effectiveness of the proposed method on the 7.5-kW experiment platform.

The high-frequency (HF) voltage injection method can accurately estimate the rotor position of an interior permanent magnet synchronous motor (IPMSM) at low speed and zero speed. However, the fixed injection frequency for the traditional HF-voltage injection method will produce loud audible noise. The random HF-voltage injection method has a spread spectrum, which can reduce the audible noise. However, random HF voltages are controlled by constant probability rules, and it is inevitable that continuous multiple random numbers are larger or smaller than the mathematical expectation in the digital implementation, which will lead to spectrum concentration of the HF current. Thus, the random HF-voltage injection method cannot reduce the audible noise effectively. A novel random HF square-wave voltage injection method for an IPMSM based on the Markov chain (MC) is proposed to reduce the audible noise. The transition of the injected HF square-wave voltages obeys the MC instead of constant probability rules, which not only can shape the power spectrum of the HF current but also enable the explicit control of the HF current time-domain performance to reduce the audible noise. In order to reduce the position error caused by the phase mismatch between the HF current and the demodulation signal, an enhanced signal processing method is presented. Then, the power spectra density of the proposed method is analyzed to verify the effectiveness. Finally, the correctness and the effectiveness of the proposed method are verified by experiment on a 2.0-kW IPMSM drive platform.

The conventional internal model control (IMC) has been used widely due to its advantages of less computational burden and simple implementation. Since the internal model controller has a fixed filter, disturbances created by mismatched models, parameter variations and other unstructured dynamic uncertainties in induction motors (IMs) cannot be eliminated by a fixed IMC. To solve these problems, the control strategy of an induction motor using internal model control with an extended state observer (IMC-ESO) was proposed. IM parameter variations and other unstructured dynamic uncertainties are considered in IM drives. Based on this model, an ESO is defined as a hypothetical equivocal function. Then the estimated disturbance is applied as a feed-forward compensation to accurately control the current loop. Since the designed ESO works concurrently with IMC, the fast dynamic response of the IMC is maintained. The feasibility and validity of the proposed method are validated by experimental results.

To improve the conventional predictive current control (PCC) performance of an induction machine (IM) system, a robust PCC method based on an immune-optimized disturbance observer (DOB) is proposed in this paper. First, a predictive current controller is established and then the performance of the PCC is analyzed when machine parameters mismatch. Second, the model PCC for an IM based on an immune-optimized DOB is proposed in this paper, and then the performance of disturbance rejection is analyzed. However, the fixed constant time of a traditional DOB will largely limit the performance of a closed-loop system. Therefore, the immune algorithm is introduced in this paper and it is used to optimize the fixed parameters of DOB. Finally, the experimental results show that both smooth dynamic responses at transient-state and high-accuracy performance at steady state can be achieved, and the effectiveness of the proposed algorithm is verified.

In order to suppress the problems of chattering, harmonic and three phase electromotive force unbalance of sliding mode observer in the sensorless control of permanent magnet linear synchronous motor, an improved sliding mode observer is studieded in this paper. These negative disadvantages can be addressed by adaptive complex coefficient filters with band pass and positive and negative sequence frequency selection characteristics. Therefore, the low pass filter of traditional sliding mode observer is replaced by adaptive complex coefficient filter in this paper.