真空断路器状态评估系统研究

发布时间：2018-03-14 10:10

本文选题：真空断路器　切入点：在线监测　出处：《大连理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：真空断路器作为控制设备,担负保障电网安全运行的重要责任,而真空断路器的在线监测,作为一种状态检修形式,可以根据电力设备的在线监测量,预测真空断路器的工作状态,进行有目的性的检修,从而可以节省资源,节约人力物力。首先,本文在前人研究的状态量基础上结合真空断路器的特性,选取最具代表性意义的真空度、分／合闸线圈电流、触头温度和剩余电寿命四个状态量作为真空断路器在线监测的指标,并对四个指标的在线监测的意义和具体检测方法进行研究。在确定了评估模型特征量的前提下,本文设计了真空断路器在线监测及状态评估硬件采集电路,可对真空断路器在线监测指标实施采集、处理和分析；并利用以FPGA为主控芯片的采集电路对实验室真空断路器样机的四个在线监测指标进行在线采集。实验结果表明,硬件电路满足实验条件,得到了指标数据。完成硬件采集电路对四个特征量的在线采集,本文采用基于混合核函数的最小二乘支持向量机算法(LS-SVM)和基于模糊理论的最优组合权重法实现真空断路器状态评估模型,其中基于混合核函数的LS-SVM算法分别采用粒子群优化算法(PSO)和遗传算法(GA)进行模型参数寻优,通过仿真分析,带入200组训练集和测试集得出利用PSO寻优的基于混合核函数的最小二乘支持向量机算法分类准确率可达98.5%,基于混合核函数GA-LSSVM分类准确率只有90%,因此采用基于混合核函数PSO-LSSVM算法可靠性更高,但PSO-LSSVM算法模型适用于解决具有大样本,高维数问题,随着特征量的不断完善可以对模型进行拓展。最后,本文采用基于最优组合权重的模糊算法,选用高斯函数作为隶属度函数对真空断路器运行状态进行评估,仿真结果同样达到要求,实现可靠分类。而基于最优组合权重的模糊算法模型适用于需要引入专家意见来指导评估建模的情况。采用两种模型对真空断路器状态评估改善了原有研究中建模方法单一的缺点,提高了评估模型的适应性,为智能化真空断路器状态评估和检修提供参考,实际运行过程中,可根据具体情况选取不同算法。
[Abstract]:As a kind of control equipment, vacuum circuit breaker is responsible for ensuring the safe operation of power network. The on-line monitoring of vacuum circuit breaker, as a form of condition maintenance, can be based on the on-line monitoring quantity of power equipment. In order to predict the working state of vacuum circuit breaker and carry out purposeful overhaul, it can save resources and manpower and material resources. Firstly, this paper combines the characteristics of vacuum circuit breaker based on the state quantity studied by predecessors. The vacuum degree of the most representative sense, the current of the branch / closing coil, the contact temperature and the residual electric life are selected as the online monitoring indexes of the vacuum circuit breaker. The significance of on-line monitoring of the four indexes and the specific detection methods are studied. On the premise of determining the characteristic quantity of the evaluation model, the on-line monitoring and state evaluation hardware acquisition circuit of vacuum circuit breaker is designed in this paper. The on-line monitoring index of vacuum circuit breaker can be collected, processed and analyzed, and the four on-line monitoring indexes of vacuum circuit breaker prototype in laboratory can be collected by using the acquisition circuit with FPGA as the main control chip. The experimental results show that, The hardware circuit meets the experimental condition and gets the index data. In this paper, the state evaluation model of vacuum circuit breaker is realized by using the least squares support vector machine (LS-SVM) algorithm based on mixed kernel function and the optimal combined weight method based on fuzzy theory. Particle swarm optimization (PSO) algorithm and genetic algorithm (GA) are used to optimize the model parameters of LS-SVM algorithm based on hybrid kernel function. It is concluded that the classification accuracy of least squares support vector machine based on mixed kernel function can reach 98.5, and the accuracy of GA-LSSVM classification based on mixed kernel function is only 90. Therefore, the hybrid kernel function is used to optimize the classification accuracy of least squares support vector machine based on hybrid kernel function. The kernel function PSO-LSSVM algorithm is more reliable. But the PSO-LSSVM algorithm model is suitable for solving the problem with large sample and high dimension. The model can be expanded with the improvement of the feature quantity. Finally, the fuzzy algorithm based on the optimal combination weight is adopted in this paper. Gao Si function is chosen as the membership function to evaluate the operating state of vacuum circuit breaker. The simulation results also meet the requirements. The fuzzy algorithm model based on optimal combination weight is suitable for the case where expert opinion is introduced to guide the evaluation and modeling. Two models are used to improve the state evaluation of vacuum circuit breaker. The disadvantage of single mode method, It improves the adaptability of the evaluation model and provides a reference for the condition evaluation and maintenance of intelligent vacuum circuit breakers. In actual operation different algorithms can be selected according to the specific conditions.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TM561

【参考文献】