Research On The Method Of Monitoring The Motion State Of CRDM Based On Coil Current

Control rod drive mechanism(CRDM)is the core component inside the reactor,so the status of CRDM will affect the safety of the reactor and effective monitoring of CRDM is extremely necessary.In the normal state,CRDM realizes the stepping movement of the control rod through the alternating action of three sets of coil currents.When an abnormality in the coil current is detected,it means that the CRDMmovement fails,and in more serious cases,it may cause the control rod to accidentally drop.In view of the characteristics of CRDM coil current changes that can effectively reflect the motion status of CRDM,this paper first determines the current hardware acquisition system,and then designs a wavelet current motion point fault detection algorithm based on feature amplification,and on this basis,further use support vector machine(SVM)to explore the possibility of coil current fault classification,and finally the CRDM motion state monitoring software based on LabVIEW is designed.The work done is as follows:First,the mechanical structure of CRDM and the role of each moving part are introduced,the CRDM movement process under the interaction of the coil currents is specifically explained,and the timing and amplitude characteristics of the CRDM coil current are analyzed.According to the requirements of CRDM coil current acquisition,the current acquisition system structure is designed,the principle and advantages of non-contact current collection are analyzed,and the CRDM coil current collection method is finally determined.Secondly,it analyzes in detail the cause of the failure of current no-motion point,and designs an algorithm to reliably detect the current motion point of wavelet transform based on feature amplification.The CRDM circuit is used to construct a curve model of the current rising section with no motion point,and the model is used to amplify some features of the current motion point,and then wavelet decomposition is performed on the current motion point of the amplified feature,so as to realize the detection of the coil current motion point.And finally the algorithm was verified.Then,a CRDM coil current fault classification algorithm based on wavelet energy entropy and SVM is designed to realize CRDM coil current fault classification.Based on the principle of wavelet energy value algorithm,the window theory and entropy value theory are introduced,and the wavelet energy entropy algorithm is constructed,which can be used to extract features.Use Genetic Algorithm(GA)to optimize the parameters of SVM,and use Particle Swarm Optimization(PSO)as a comparison,and input the obtained feature vector into the optimized SVM.The comparison results prove the superiority of the wavelet energy entropy feature as extraction algorithm compared to the wavelet energy value and the efficiency of GA as an SVM parameter optimization algorithm compared with PSO.Finally,The software for realizing CRDM motion state monitoring is designed.A design plan based on the four basic modules is drawn up,and the five algorithms of CRDM motion state recognition,current motion point detection,single-step current overrun analysis,current fault classification and waveform review are analyzed in detail.Using the UDP network transmission protocol to simulate the scene of real-time data collection and transmission,the software’s real-time display function,single-step current overrun analysis function,current motion point detection function,current fault classification function and current waveform review function are tested.It shows the feasibility of the CRDM motion state monitoring software based on LabVIEW.