Bi-LSTM Algorithm Based Identification of Load Model Parameter for Fault Analysis

Because the increasingly promoted high proportion of power electronics based load devices and distributed power sources, the impact on short circuit currents caused by clustered loads when fault occurred cannot be ignored. It is of great value to establish a correct and accurate model for the purpose of simulation in fault cases, for the exiting load model is founded and only suitable for small disturbance cases. So an AI based parameter identification algorithm is proposed. Firstly, according to the sensitivity analysis, the parameters which have a significant impact on simulation result of short circuit currents are identified, both of the comprehensive load model which considering distribution network indirectly (CLM) and of the synthesis load model which considering distribution network directly(SLM). Secondly, the AI based algorithm namely bidirectional long term and short term memory networks (Bi LSTM) is adopted to parameters identification. Finally, the feasibility of the aforementioned method is verified by comparison with some other identify algorithms, and the correctness of identified parameters are also confirmed by simulation cases.