设备资产运维精益管理系统(power production management system,PMS)SF6气体量数据不全且误差较大,无法为电网企业核算碳储量以及实现待建变电站碳规划提供基础数据。针对上述情况,研究了计及母线和断路器的变电站碳储量核算方法,并结...设备资产运维精益管理系统(power production management system,PMS)SF6气体量数据不全且误差较大,无法为电网企业核算碳储量以及实现待建变电站碳规划提供基础数据。针对上述情况,研究了计及母线和断路器的变电站碳储量核算方法,并结合宁夏电网现场实测数据,通过MIC法筛选神经网络输入参数,构建了6输入参数的GA-BP、PSO-BP、HPO-BP神经网络模型,结果表明HPO-BP神经网络模型的评估指标及预估结果相对误差(6.28%)均优于其余2种神经网络模型,可以准确核算断路器SF6气体量。针对参数不确定情况,根据PCCs法分析不同参数之间的线性关系,构建了3输入参数的HPO-BP神经网络模型,预估结果相对误差为9.72%。通过遍历输出方式,在参数不确定情况下输出多组断路器SF6气体量预估数据,利用求和累积方法获取变电站总SF6气体量,并量化为变电站碳储量,从而为电网企业实现“双碳”目标提供数据支撑。展开更多
In order to realize on-line quantitative detection on SF6 and effective control of nlnning state of SF6 high voltage power supply system, a concentrated SF6 quantitative ultrasonic on-line deteetion system has been de...In order to realize on-line quantitative detection on SF6 and effective control of nlnning state of SF6 high voltage power supply system, a concentrated SF6 quantitative ultrasonic on-line deteetion system has been developed based on the actual demand of electric power system consumers. There are four major characteristics in this system. Firstly, the gas of maximum 64 detection points is transferred through the specific air path to the detection devices to he detected and analyzed, thereby the electrical lines and the complicated installation of the collectors can be avoided; secondly, the differential technique is used to shield the influence of environmental factors, which effectively improves the accuracy of the acoustic detection; thirdly, the SF6 coneentration is determined by the speed and phase in the ultrasonic wave trans- mission process, therefore there is no secondary pollution for the purely physical means; finally, the ma- ture embedded technique is applied in this system to improve its intelligence and stability.展开更多
文摘设备资产运维精益管理系统(power production management system,PMS)SF6气体量数据不全且误差较大,无法为电网企业核算碳储量以及实现待建变电站碳规划提供基础数据。针对上述情况,研究了计及母线和断路器的变电站碳储量核算方法,并结合宁夏电网现场实测数据,通过MIC法筛选神经网络输入参数,构建了6输入参数的GA-BP、PSO-BP、HPO-BP神经网络模型,结果表明HPO-BP神经网络模型的评估指标及预估结果相对误差(6.28%)均优于其余2种神经网络模型,可以准确核算断路器SF6气体量。针对参数不确定情况,根据PCCs法分析不同参数之间的线性关系,构建了3输入参数的HPO-BP神经网络模型,预估结果相对误差为9.72%。通过遍历输出方式,在参数不确定情况下输出多组断路器SF6气体量预估数据,利用求和累积方法获取变电站总SF6气体量,并量化为变电站碳储量,从而为电网企业实现“双碳”目标提供数据支撑。
基金Supported by the National Natural Science Foundation of China (No. 10574038)
文摘In order to realize on-line quantitative detection on SF6 and effective control of nlnning state of SF6 high voltage power supply system, a concentrated SF6 quantitative ultrasonic on-line deteetion system has been developed based on the actual demand of electric power system consumers. There are four major characteristics in this system. Firstly, the gas of maximum 64 detection points is transferred through the specific air path to the detection devices to he detected and analyzed, thereby the electrical lines and the complicated installation of the collectors can be avoided; secondly, the differential technique is used to shield the influence of environmental factors, which effectively improves the accuracy of the acoustic detection; thirdly, the SF6 coneentration is determined by the speed and phase in the ultrasonic wave trans- mission process, therefore there is no secondary pollution for the purely physical means; finally, the ma- ture embedded technique is applied in this system to improve its intelligence and stability.
