国网山东电科院创新研发国家电网公司首套自动化建模工具

Core Insights - The article discusses the development of the first automated modeling tool by the State Grid Shandong Electric Power Research Institute, aimed at enhancing the integration of big data and digital technology in the energy sector [1][2]. Group 1: Tool Development and Features - The tool, named "AutoML-based Power Big Data Automated Modeling Tool," automates the entire process from data preprocessing to model construction and evaluation, significantly improving the efficiency and accuracy of power data analysis [1]. - It integrates advanced technologies such as artificial intelligence, big data, and cloud computing, allowing for efficient access to various data sources including historical load, distributed photovoltaic output, and weather forecasts [1]. - The tool has successfully automated modeling for over 10 key scenarios, including distributed photovoltaic output prediction and transmission line inspection defect analysis [1]. Group 2: Performance and Impact - Traditional power data analysis relied on manual modeling, which faced challenges such as long cycles, high professional thresholds, and insufficient model generalization capabilities [2]. - Since its establishment in June 2024, the AutoML tool has demonstrated outstanding performance, achieving over three times the efficiency of traditional manual modeling and a 5 percentage point increase in prediction accuracy during the peak summer load period [2]. - The company plans to continuously optimize the model, expand external feature indicators, and explore more feature transformation, combination, and selection methods to enhance the tool's application in grid planning and energy conservation [2].