Research On Condition Monitoring Of Control Loop Based On Digital Twin

In today’s world,informatization and industrialization are integrated and developed deeply.With the implementation of "Industry 4.0","Made in China 2025",and "Industrial Internet",Cyber Physical System(CPS)which interacts and integrates physical space and information space has become a research hotspot.As the main technology of CPS,the digital twin(DT)has developed rapidly.The principle is to make full use of data,such as physical objects,sensor collection,and operating history to establish a twin model of a physical entity.The twin model runs in parallel with the physical entity,and optimizes the operation of the physical entity through the calculation and analysis of the physical object and the virtual model.At present,digital twin is mainly used in equipment monitoring,as for control loops,poor performance control effect will shorten the service life of the equipment and reduce the quality of the product.Therefore,how to apply digital twin to control the state monitoring of loop status has become an urgent need in the industry.In this paper,aiming at the"virtual model" and "service system",the on-line modeling method of thermal objects and the evaluation method of loop performance are studied deeply.(1)According to the five-dimensional structure of the digital twin,and the operating characteristics of the control loop,a set of control loop full life cycle simulation system is proposed.Combined with the state monitoring system of the control loop,the overall architecture of the control loop digital twin system is proposed.(2)For the "virtual model" part,in order to solve the nonlinear,time-varying,large delay and large inertia characteristics of the thermal object,a modeling method based on Least Squares Support Vector Machine(LS-SVM)is proposed.The strategy is updated by training samples,more specifically,the training samples are updated in real time by calculating the Euclidean distance between the new sample and the training sample,so as to realize the online update of the model.(3)As for the "service system" part,in order to solve the loop evaluation problem of noise obeying non-Gaussian distribution,this paper proposes a control system performance evaluation method based on the minimum entropy benchmark.The residual error is estimated by the ARMA algorithm,therefore the lower limit of the minimum entropy of the loop output is obtained.Compared with the entropy value of the loop output,the minimum entropy performance index is obtained.Then combined with other evaluation criteria,a comprehensive evaluation method of control loop performance is formed.(4)According to the overall architecture of the proposed control loop digital twin system,the work flow of the loop full life cycle simulation system is simulated for the high pressure heater’s water level control loop in power plant.At the same time,a control loop performance evaluation platform was built,and the comprehensive performance evaluation of the control loop is realized by combining various control indexes,which verifies the feasibility of the control loop digital twin system proposed in this paper.