摘要
为解决"海量数据"和"有限知识"之间的矛盾,提高继电保护通信电路风险评估准确性和算法效率,提升事前预警能力,论文提出一种融合关联规则分析和神经网络的风险评估方法 Apriori-BPNN。该方法首先基于多源异构的数据构建初始指标体系,在此基础上采用改进的Apriori算法确定多种因素与目标的关联度,实现指标筛选;采用BP神经网络算法确定各因素的权重,并加权求得最终的综合风险评估指标。仿真结果表明,Apriori-BPNN既避免了传统层次分析法的不足,又简化了神经网络的结构,提高了继电保护通信电路风险评估准确度和算法效率,使运维人员能及时发现存在的隐患和风险,指导继电保护通信电路的主动预警与智能检修,提高继电保护业务运行的安全性和稳定性。
With combination of both association rules analysis and neural network method,an algorithm named " AprioriBPNN" for assessing the risk is proposed in order to solve the contradiction between the " massive data" and "limited knowledge" and improve accuracy and efficiency of risk assessment on the relay protection communication circuit.After an intimal index system is set up based on the heterogeneous multi data,the improved Apriori algorithm is used to determine the correlation of various factors and objectives to achieve the index screening.The BP neural network algorithm is used to determine the weight of each factor to obtain the final comprehensive risk evaluation index.Simulation results indicate that the AprioriBPNN avoids disadvantages of the traditional analytic hierarchy process,simplifies the structure of neural network and improves accuracy and efficiency of the risk assessment algorithm of the relay protection communication circuit.The operation and maintenance staff can timely detect the hidden dangers and risks so as to guide relaying protection communication circuit of active warning and intelligent maintenance,finally improves the security and stability of the relay protection operation.
出处
《计算机与数字工程》
2017年第4期700-705,739,共7页
Computer & Digital Engineering
基金
江苏省电力公司运用大数据技术的通信网络全景监测与智能调配技术研究与应用项目(编号:500327478)资助