Research On Controlled Fault-current Interruption Based On RLS Algorithm

High voltage intelligent circuit breaker is becoming hot following the development of smart grid technology. As an important technology for the development of smart grid, phase control has attracted more and more attention. The essence of phase control is to control the circuit breaker to close or open at the optimal phase, it has been used in the switching operation of capacitor, reactor and transformer. But it has not been used in the opening of short-circuit fault, so controlled fault-current interruption (CFI) is promising. It can improve the power quality, reduce the arcing time, extend the life and improve the breaking capability of circuit breakers.In order to realize CFI, the fault current zero-point must be predicted fast and accurately. Research and simulation are put on taylor expansion algorithm and complete response algorithm based on Recursive Least Square (RLS) for parameters estimation. The simulation result show that both the two algorithm can get a predicted zero-point with less than 0.3ms error in the environment of 40dB noise. The time convergence of predicted zero-point error of complete response algorithm is faster than taylor expansion algorithm. Because of the existing problem of F test method, R test method has been applied in the detection of short-circuit fault, it can save the software and hardware resources.Based on FPGA, Verilog HDL language has been used to design the software scheme of CFI. It has been verified in Modelsim. The FPGA minimum system, AD sampling module and peripheral circuit have been designed as hardware circuit.A test about CFI was put by using the testing device for relay protection as the generation of fault current. The result show that the system can detect the current fault and recognize the type of fault correctly, fulfill basic requirements of controlled fault interruption. The time convergence of predicted zero-point error of complete response algorithm is faster than taylor expansion algorithm. Taylor expansion algorithm is suitable for small time constant, and complete response algorithm is suitable for big time constant. It matches the simulation result. Both the two algorithm can get a predicted zero-point error with less than 0.45ms in the experiment. The design of the CFI system can fulfill the CFI requirement.