CN-121997243-A - Multisource measurement data fusion method for operation and maintenance management of distribution network equipment

Abstract

The invention relates to a multisource measurement data fusion method for operation and maintenance management of distribution network equipment, which comprises the steps of firstly constructing a layered data acquisition architecture through operation and maintenance requirements of the distribution network equipment, and establishing a unified data transmission protocol and a standardized data management standard, so as to complete construction of a multisource measurement data acquisition system, then acquiring equipment operation electric quantity data, environment perception data, mechanical state quantity data and historical operation and maintenance record data by using the constructed multisource measurement data acquisition system, preprocessing the data, extracting key features capable of reflecting equipment states from the data, finally carrying out weighted preliminary fusion on the same type of data by adopting a layered fusion mode, and then carrying out depth fusion of cross-latitude data, so that spatial association and time sequence features of the data are mined, and the problems of non-uniform multi-source data format and information redundancy conflict are effectively solved.

Inventors

• GU GUOZHI
• ZHOU XU
• SHI WENLONG
• WANG LEI
• HU ZHIWEI
• Jia Aohui
• BI LIANG
• SHEN CHAOQUN
• JIANG ZIYU
• SUN YILONG
• WANG WEIMIN
• XU MING
• YANG ZHONGLI
• SU TAO

Assignees

• 国网河南省电力公司商丘供电公司

Dates

Publication Date

20260508

Application Date

20251202

Claims (6)

1. 1. A multisource measurement data fusion method for operation and maintenance management of distribution network equipment is characterized by comprising the following steps: step 1, constructing a layered data acquisition architecture according to operation and maintenance requirements of distribution network equipment, and establishing a unified data transmission protocol and a standardized data management specification, thereby completing construction of a multi-source measurement data acquisition system; Step 2, determining the type of a measurement data source of the distribution network equipment, wherein the measurement data source comprises equipment operation electric quantity data, environment sensing data, mechanical state quantity data and historical operation and maintenance record data; Step 3, according to the type of the data source measured by the distribution network equipment, the data is acquired in real time by utilizing sensors, intelligent meters and fault wave recording devices which are deployed in the transformer, the switch cabinet, the cable line and the periphery in a layered data acquisition architecture; step 4, preprocessing the accessed data, namely sequentially performing data cleaning, data standardization alignment, data normalization and data complement operation; Step 5, extracting the characteristics of the multisource measurement data after the pretreatment in the step 4, namely extracting key characteristics capable of reflecting the state of equipment from various data, such as extracting current effective value trend, harmonic distortion rate and voltage sag/temporary rise times from electric quantity data; Step 6, establishing a layered data fusion model, wherein a first layer adopts a weighted average algorithm to perform preliminary fusion on the measurement data of the same type and dynamically distributes weights according to the reliability of the data; and 7, verifying the validity of the fusion result, and based on the verification result, finely adjusting parameters of the layered data fusion model to ensure the accuracy of the fusion result.
2. 2. The method for multi-source measurement data fusion for operation and maintenance management of distribution network equipment according to claim 1, wherein the method is characterized in that in the step 4, abnormal values and null values are removed through data cleaning, the data is normalized and aligned, namely, the timestamp and interpolation processing are unified, the data is normalized, namely, the incoordination caused by different dimensions and sizes between data characteristics is eliminated, and the missing values are reasonably estimated through known information through data complementation, so that a complete and consistent data set is constructed.
3. 3. The method of claim 1, wherein the deep learning network in the step 6 is a combination model of an improved convolutional neural network and a long-short-term memory network, the data space features are extracted through the convolutional neural network, and the long-short-term memory network is used for capturing the data time sequence features.
4. 4. The method of claim 1, wherein the step 6 uses an improved gating cell network to enhance the extraction weight of key data features by introducing attention mechanisms.
5. 5. The method of claim 1, wherein the validity verification in the step 7 is performed by using a mean square error and a mean absolute error as evaluation indexes, and when the evaluation indexes exceed a preset threshold, re-optimizing the weighting weights and network parameters of the deep learning model.
6. 6. The method for multi-source measurement data fusion for operation and maintenance management of distribution network equipment according to claim 1, wherein the electrical quantity data in the step 2 comprises current, voltage, active power, reactive power, power factor and three-phase imbalance, the environment sensing data comprises environment temperature, environment humidity and meteorological information, the mechanical state quantity data comprises oil temperature, winding temperature, opening and closing positions and energy storage states, and the historical operation and maintenance record data comprises operation and maintenance record, historical fault information and an emergency treatment scheme corresponding to the historical fault information.

Description

Multisource measurement data fusion method for operation and maintenance management of distribution network equipment Technical Field The invention belongs to the technical field of operation and maintenance management of distribution network equipment, and particularly relates to a multisource measurement data fusion method for operation and maintenance management of distribution network equipment. Background At present, with the continuous improvement of the automation level of a smart grid and a power distribution network, various and huge measuring devices are deployed in the power distribution network, and the devices generate massive, heterogeneous and multi-time-scale operation data, so that a data base is provided for the fine operation and maintenance management of distribution network equipment; however, measurement data from different sources are usually collected and managed by an independent system, the formats are not uniform, the precision difference is large, and the data are lack of effective association and fusion, so that a data island is formed, the comprehensive and unified cognition on the equipment state cannot be formed, and the huge amount of multi-source measurement data cannot be subjected to deep fusion and effective utilization, so that the intelligent level improvement of the operation and maintenance of the distribution network is restricted, and meanwhile, the effective prediction and discrimination on the health state and the operation trend of the equipment hidden behind the distribution network cannot be performed, therefore, in order to solve the problems, the development of the multi-source measurement data fusion method for the operation and maintenance management of the distribution network equipment, which is strong in adaptability and high in fusion precision, is necessary. Disclosure of Invention The invention aims to overcome the defects of the prior art and provide a multisource measurement data fusion method for operation and maintenance management of distribution network equipment, which has strong adaptability and high fusion precision. The invention aims to realize the multi-source measurement data fusion method for the operation and maintenance management of the distribution network equipment, which comprises the following steps: step 1, constructing a layered data acquisition architecture according to operation and maintenance requirements of distribution network equipment, and establishing a unified data transmission protocol and a standardized data management specification, thereby completing construction of a multi-source measurement data acquisition system; Step 2, determining the type of a measurement data source of the distribution network equipment, wherein the measurement data source comprises equipment operation electric quantity data, environment sensing data, mechanical state quantity data and historical operation and maintenance record data; Step 3, according to the type of the data source measured by the distribution network equipment, the data is acquired in real time by utilizing sensors, intelligent meters and fault wave recording devices which are deployed in the transformer, the switch cabinet, the cable line and the periphery in a layered data acquisition architecture; step 4, preprocessing the accessed data, namely sequentially performing data cleaning, data standardization alignment, data normalization and data complement operation; Step 5, extracting the characteristics of the multisource measurement data after the pretreatment in the step 4, namely extracting key characteristics capable of reflecting the state of equipment from various data, such as extracting current effective value trend, harmonic distortion rate and voltage sag/temporary rise times from electric quantity data; Step 6, establishing a layered data fusion model, wherein a first layer adopts a weighted average algorithm to perform preliminary fusion on the measurement data of the same type and dynamically distributes weights according to the reliability of the data; and 7, verifying the validity of the fusion result, and based on the verification result, finely adjusting parameters of the layered data fusion model to ensure the accuracy of the fusion result. Further, the data in the step 4 is cleaned, namely abnormal values and null values are removed, the data is normalized, namely the time stamp and interpolation processing are unified, the data normalization, namely the incoordination caused by different sizes and dimensions among data features is eliminated, and the data complementation, namely the known information is utilized to reasonably estimate the missing values, so that a complete and consistent data set is constructed. Furthermore, the deep learning network in the step 6 is an improved combination model of a convolutional neural network and a long-term memory network, the data space characteristics are extracted through the convolutional neural network, and