基于关联数据挖掘的继电保护定值风险评估方法研究

继电保护是确保电网安全运行的关键，为降低继电保护运行风险，现提出一种基于关联数据挖掘的继电保护定值风险评估方法。通过分析常见连锁故障类型及影响因素，采用数据挖掘求得电力系统故障关联度，计算挖掘数据间关联信息熵。利用关联信息熵系数挖掘电力系统内历史相关数据，深度分析每日数据波动特征，明确系统负荷孤立、电源孤立和电网解列情况发生的风险概率，以及功率损失及风险程度，建立继电保护定值风险评估模型，完成继电保护系统的综合风险评估。实验结果表明：所提方法能够有效评估出继电保护系统的风险概率和风险程度，算法收敛速度快，且风险评估的相对均方根误差、平均绝对误差低于2%,均方根误差低于20 kW,说明该方法有利于电力系统继电保护定值风险控制。

Research on Relay Protection Setting Risk Assessment Method Based on Association Data Mining

Relay protection is the key to ensure the safe operation of power grid. In order to increase the reliability of power system relay protection and reduce the cascading failures of large power system caused by relay protection, a relay protection setting risk assessment method based on association data mining is studied. By analyzing the types and influencing factors of common cascading faults, the correlation degree of power system faults is obtained by data mining, and the correlation information entropy of mining data is calculated. Use the correlation information entropy coefficient to mine the historical relevant data in the power system, deeply analyze the daily data fluctuation characteristics, define the risk probability of system load isolation, power source isolation and power grid splitting, as well as the power loss and risk degree, and complete the comprehensive risk assessment of the relay protection system by virtue of the constant value risk assessment in each case. The experimental results show that the proposed method can effectively evaluate the risk probability and risk degree of the relay protection system, the algorithm converges quickly, and the relative root mean square error, the average absolute error ratio of the risk assessment is less than 2%, and the root mean square error is less than 20kW, which is conducive to the risk control of relay protection settings in the power system.