Study On Parameter Identification And Control Optimization Of Governing System For Pumped Storage Unit

With the deepening reform of energy structure in China,pumped storage plant,considered as one of the energy storage means for renewable resources and the effective tool for peak load and frequency regulation,has achieved rapid development and large-scale construction.However,some negative traits existed,i.e.,the complicated layout of the conduit tubes,the reversible operation mode,the continual condition transformations and the intrinsic 'S' characteristic and 'hump' effect,result in the characteristics of time variant,non-minimum phase and strong nonlinearity and hydraulic-mechanical-electrical coupling in the governing system of pumped storage unit.These characteristics not only bring about great difficulty to the running of the machine and the stability of the power plant,but also cause increasingly prominent security issue,which leads to a series of scientific frontier problems and engineering technical problems to be solved for safe,stable and efficient operations of the pumped storage unit.Further,researches on the key techniques,such as advanced control strategy and optimal operation,are eager to be carried out due to the complex nonlinearity property in control and optimization of the governing system caused by the penetration of large-scale intermittent renewable energy and violent fluctuations in power grids.The precise modeling of the governing system of pumped storage unit is the fundamental of conducting the aforementioned scientific problems.Most of the conventional modeling schemes simplify the pump turbine as a first-order linear model which is troublesome to describe the operation features of the unit and strong nonlinearity in the governing system.Thus,the study and development in the key technical fields of parameter identification,intelligent control strategy,optimal operation and so forth has been restrained.To this end,by taking the safety,efficiency and stability in the operation of pumped storage unit as the pointcut,the system identification,modern control theory and intelligent optimization as the basis along with the object modeling,simulation calculation and engineering practice validation as the leading clue,system identification and control optimization of the pumped storage unit has been well studied in this thesis on account of the critical scientific problems and technical bottlenecks in precise modeling and control optimization.Moreover,the research findings have been practically demonstrated in Hongping pumped storage power plant in Jiangxi Province,China.The main research achievement and novelty in this thesis are given in the following:(1)The linear model of the governing system of pumped storage unit satisfying different research and application demand and the nonlinear and computational models of the pump turbine based on the Logarithmic projection transformation of complete characteristic curves has been established by synthetically analyzing the structure and mathematical description of every element in the governing system of pumped storage unit and combining the operation characteristics and the numeric mapping method of the complete characteristic curves.Subsequently,the bifurcation phenomenon of the unit rotation speed in the control parameter domain has been discovered and nonlinear system dynamics-based time-domain diagram,phase trajectory and Pioncare map have been obtained according to the nonlinear model of the governing system being built.The influence of the 'S' characteristic of pump turbine on the oscillations of the unit rotation speed has been revealed and the nonlinear dynamic behavior and its law of dynamic evolution has been investigated,which provides the research on advanced control strategy with theoretical foundations.(2)For the sake of implementing parameter identification of the real-time operating condition of the governing system of pumped storage unit,a bacterial foraging chemotaxis-gravitational search algorithm(BCGSA)that is suitable to solve high-dimensional parameter identification is proposed by taking the nonlinear model of governing system of pumped storage unit as the study object.The precise identification of the nonlinear system is realized on the establishment of BCGSA-based adaptive integrated parameter identification model with the introduction of a parameter identification weighting factor.Then,by taking the governing system linear system as the study object,the parameter identification method of governing system white-box linear model mapping-based BP neural network is proposed to overcome the defects of reliance on large amount of sampling data in the conventional neural network system identification and its weakness in generalization capability of identification results,thus to improve the research paradigm of precise modeling of the governing system.(3)Focused on the rotation speed fluctuation and difficult synchronizing problems when the unit is working at low head and entering 'S' characteristic zone,a novel fractional order PID controller inspired by the idea of fractional order calculus has been founded to effectively alleviate the adverse impact on the system stability under no-load condition.Furthermore,a set of practical fuzzy logic rules is deduced and the adaptive fast fuzzy fractional order PID(AFFFOPID)controller is developed by the in-depth study on theories of fuzzy control and the synthetically trait analysis of fuzzy fractional order PI and fuzzy fractional order PD controller.Simulation experiments have indicated that AFFFOPID could further improve control performances of the governing system of pumped storage unit in view of the frequency and the guide vane opening transient process.(4)As for the instability issue of pumped storage unit in condition transformations,the influence of different guide vane closure rules on the hydraulic characteristics in every hydraulic element has been discussed.By introducing the safe threshold boundary of the hydraulic transient calculation and multiple constraints on velocity of the governor oil,et al.,the multi-objective optimization model for guide vane closure rule is elected based on the proposed computational model of "single tube-double unit" governing system of the pumped storage unit considering the balance of rotation speed rising ratio and water-hammer pressure fluctuations.Moreover,an improved multi-objective gravitational algorithm(IMOGSA)based on mechanisms including population reconstruction,multi-objective chemotaxis operations and elite archive set guidance,and so forth,has been presented to implement the efficient solution to the muti-objective optimization problem of guide vane closure rules and swiftly formulate the evenly distributed and widespread non-dominant scheme set under load rejection and pump outage conditions,thus to provide theoretical clue and scientific guidance on the safety assurance of the stable operation of the pumped storage unit.(5)Based on the theoretical achievements about the study on system identification and control optimization of the governing system of the pump storage unit in this thesis,a service-oriented application system for control optimization and performance evaluation of the governing system of pumped storage unit has been designed and developed.The application system,which has been successfully applied in Hongping pumped storage power plant in Jiangxi Province,implements the functions of parameter identification,control optimization and performance evaluation of the governing system of the pumped storage unit and provides the power plant operators with scientific auxiliary decisions and suggestions,thus to accelerate the science promotion and development of pumped storage industry in China.