Abstract:In order to effectively extract the characteristics of electric shock faults and realize the separation ofelectric shock current from residual current，this paper proposes a new method of electric shock current extractionbased on the combination of Variational Mode Decomposition（VMD）and Long Short Term Memory（LSTM）. Theparameters of VMD［K，α］were optimized by fruit fly optimization algorithm to obtain the optimal combination of pa?rameters［6，280］. Based on the mutation characteristics of the optimal modal component of the residual current de?composed by VMD，the growth rates η1 and η2 of the sum of the current amplitudes in adjacent periods are defined asthe characteristic quantities for judging electric shock accident. The electric shock current signal is reconstructedfrom the six layers modal component signal，and the electric shock current detection model based on LSTM networkis constructed. The results of 240 groups of electric shock signals show that，when at least one of η1 and η2 is greater than 1%，electric shock occurs，otherwise no electric shock accident occurs. Compared with the VMD-BP andVMD-RBF detection models，the average correlation coefficients of the VMD-LSTM detection model are increased by6.2% and 2.3%，respectively，and the average root mean square error is reduced by 36.8% and 27.1%，respectively.The method proposed in this paper has higher detection accuracy. The research results provide a certain reference forthe development of residual current protection devices based on the electric shock current of the biological body.

Abstract:Aiming at the problem that the parameters of fault diagnosis model are difficult to be optimized of wind turbine pitch system, a fault diagnosis method of wind turbine pitch system based on multi-class optimal margin distribution machine optimized by the state transition algorithm (mcODM-STA) is proposed. In this method, the wind turbine power output is selected as the main state parameter, and Pearson correlation coefficient is used to analyze the historical operation data of wind turbine in wind power data acquisition and monitoring control system, and the features with low correlation of power output state parameters are eliminated. The remaining features are analyzed twice to reduce the sample features. The data set is divided into training set and test set. The training set is used to train the proposed fault diagnosis model, and the test set is used for testing. The operation data of a domestic wind farm is used for experimental verification. Experimental results show that the proposed method has higher fault diagnosis accuracy and Kappa coefficient than other parameter optimization methods.

Abstract:Due to the shortcomings of traditional fault diagnosis system, such as too complicated hardware system and fault recognition algorithm, a distributed intelligent fault diagnosis system based on particle filter was proposed and studied. Real-time collection of distributed multi-variable parameters was realized by adopting ZigBee wireless sensor network, on-line processes variable data based on particle filter, and precise estimate about real system states were obtained based on simple pattern recognizing algorithm in order to realize the intelligent forecast and diagnose for system fault. The distributed fault diagnosis system includes ZigBee network, particle filter algorithm, system states model and malfunction mode recognition. Particle filter can filter data collected by sensor, suppress and eliminate the interference or significant error that affects the estimate of system states based on sequential importance sampling and Monte-Carlo method. Finding a system state model that has the minimum sum of residuals with an estimate curve about system states from a particle filter is the process of the malfunction mode recognition. Realization of the distributed intelligent fault diagnosis system and the result of the experiment show that the system can realize the remote monitor, accurate state estimation and fault diagnose, and it has the advantage of low cost, high reliability and easy to realize. The work can expand the application range of distributed sensor network and improve the diagnosis level of the fault diagnosis system.

Abstract:The Kriging model was introduced to detect and diagnose the faults in the chillers of building air-conditioning systems. This model was built and validated by using the normal data from ASHRAE RP-1043. The methods of T-statistic and exponentially-weighted moving average (EWMA) were compared by the sensitivity of performance indexes. The results show that the EWMA can achieve better performance sensitivity. Combined with the EWMA, Kriging model and the rules of fault diagnosis, the chiller faults like condenser fouling, refrigerant overcharge, refrigerant leakage, non-condenser gas, reduced evaporator water flow rate, and reduced condenser water flow rate were diagnosed using the measured data from ASHRAE RP-1043. The diagnosis results show that the chiller faults at different levels can be accurately and efficiently detected and diagnosed by using the Kriging model.

Abstract:To address the computing instantaneous frequency of the product function (PF) in local mean decomposition (LMD), a new instantaneous frequency of a signal computing method was introduced. This method is piece-wise wave based. Firstly, a signal was separated to a number of full waves. Then, the instantaneous phase of each full wave was defined by a set of monotonic increasing arcsine functions. Therefore, the instantaneous frequency of a signal was obtained. Theoretically, the instantanoues frequency obtained in this method was positive, stable and could guarantee the characteristic information of signal integrity. This method was applied to compute the instantaneous frequency of simulated signals and actual gear fault vibration signals, and the results were compared with those obtained in other methods. It has been shown that this method is quite suitable for extracting the instantaneous frequency of a signal.

Abstract:According to the principle of fault detection and diagnosis, the technology of data fusion with neural network was used to deal with a lot of data obtained from fault detection and diagnosis. A new method based on Dempster-Shafer theory of evidence to sol

Abstract:In this paper,the used fault detection method based on the SequentialProbability Ratio Test(SPRT)is simply discussed,and the influence of the auxiliary signal on the average fault test time and the false alarm probability of the SPRT method isalso did.A s