基于混合優(yōu)化算法的電網(wǎng)故障診斷

 

謝瑞，張興旺

（南昌工程學(xué)院 江西省精密驅(qū)動與控制重點實驗室，江西 南昌 333000）

 

    摘 要：電網(wǎng)故障過程中保護和斷路器動作及告警信息存在不確定性，會使原有電網(wǎng)故障解析模型診斷出現(xiàn)錯誤。在現(xiàn)有解析模型基礎(chǔ)上，通過電網(wǎng)結(jié)構(gòu)、保護配置及斷路器的動作規(guī)則進行解析，考慮各級保護之間的互相影響，針對可疑母線和線路分別建立目標函數(shù)，構(gòu)建新的解析模型。采用混合優(yōu)化算法對目標函數(shù)進行求解，將模擬植物生長算法（PGSA）與粒子群算法（PSO）結(jié)合，初始生長點的選取對于PGSA能否收斂于全局最優(yōu)解起著決定作用，先通過PSO的高魯棒性初選優(yōu)秀的初始生長點，再基于PGSA的高效搜索能力得到最終的全局最優(yōu)解。算例結(jié)果表明，改進的解析模型更加合理，混合優(yōu)化算法搜索速度與收斂精度大幅度提高。

    關(guān)鍵詞：故障診斷；優(yōu)化模型；模擬植物生長算法；粒子群算法；告警信息

    中圖分類號：TM711     文獻標識碼：A     文章編號：1007-3175(2020)11-0031-05

    Abstract: There are uncertainties in the protection and circuit breaker action and alarm information in the process of power grid failure, which will cause errors in the diagnosis of the original grid fault analysis model. On the basis of the existing analytical model, analyze the power grid structure, protection configuration and the action rules of the circuit breaker, consider the mutual influence between all levels of protection, establish objective functions for suspicious buses and lines, and build a new analytical model. A hybrid optimization algorithm is used to solve the objective function, and the simulated plant growth algorithm (PGSA) is combined with the particle swarm algorithm (PSO). The selection of the initial growth point determines whether the PGSA can converge to the global optimal solution. First pass the PSO The high robustness of PGSA initially selects excellent initial growth points, and then obtains the final global optimal solution based on the efficient search ability of PGSA. The results of calculation examples show that the improved analytical model is more reasonable, and the search speed and convergence accuracy of the hybrid optimization algorithm are greatly improved.

    Key words: fault diagnosis; optimization model; simulation plant growth algorithm; particle swarm algorithm; warning information