微型电脑应用

为了有效减少调度失误以及保证操作票拟定的合理性，避免因人为疏忽、操作不当等原因导致的电力系统事故和故障发生，保证电力系统管理的规范化和标准化，研究基于改进极限学习机与交互逻辑的倒闸操作指令网络化复核模型。该模型以电气设备现场接线数据、典型接线防误逻辑数据、现场防误装置数据为依据，通过改进极限学习机识别倒闸动作；依据倒闸防误交互逻辑规则，在评估、复核、计算分析该动作识别结果后，复核倒闸操作指令，并完成调度中心与厂站或者运维中心之间的网络化交互。测试结果显示，所研究模型能够完成不同类别倒闸动作的可靠识别，F₁指标结果均在90%以上，识别结果的可信度均在0.933以上，可清晰呈现相关倒闸操作的具体执行情况以及执行结果，可生成倒闸操作指令安全复核申请文件以及复核审核文件，全面完成倒闸操作指令复核。

In order to effectively reduce scheduling errors and ensure the rationality of operation ticket formulation, avoid power system accidents and faults caused by human negligence, improper operation, and ensure the standardization and normalization of power system management, a networked review model of switching operation instructions based on improved extreme learning machine and interactive logic is studied. This model is based on on-site wiring data of electrical equipment, typical wiring error prevention logic data and on-site error prevention device data, and identifies switching actions through improved extreme learning machine. Based on the logic rules of anti-error interaction for switching, this paper evaluates, reviews, calculates and analyzes the identification results of the action, reviews the switching operation instructions, and completes the networked interaction between the dispatch center and the plant or operation and maintenance center. The test results show that the model can reliably identify different types of switching actions, F₁ indicator results are above 90%, and the reliability of the recognition results are above 0.933. It can clearly present the specific execution status and results of related switching operations, generate safety review application documents and re-review documents for switching operation instructions, and comprehensively complete the review of switching operation instructions.