Recently, State Grid Henan Electric Power Company has started using an icing monitoring and alarm platform for new energy generator sets, successfully completing three rounds of icing monitoring and alarm during rain, snow, and freezing weather. This platform helped dispatchers to quickly correct new energy power generation forecasting information, and warned 11 new energy stations to carry out de-icing operations on 63 units in a timely manner. Through this platform, the company has realized remote automatic monitoring of icing on new energy generator sets, providing important support for scientifically and efficiently predicting new energy power generation and ensuring the safe and stable operation of the power grid.

Currently, the new energy generator set icing monitoring and alarm platform has implemented icing monitoring, prediction, and early warning functions across four dimensions: stand-alone units, stations, cities, and provinces. The platform can issue early warnings for new energy generator sets in the province up to three days in advance, support professionals in real-time monitoring of unit icing and outage capacity based on different statistical metrics, dynamically correct the prediction results of new energy power generation, and construct a "data-driven-monitoring linkage-prediction closed loop" regulatory strategy for new energy power generation under adverse weather conditions. This improves the quality and efficiency of power grid operation scheduling and anti-icing and de-icing efforts at the stations.

Based on this platform, State Grid Henan Electric Power Company classifies the new energy generator sets in the province into four categories of icing severity - "Easy," "Normal," "Difficult," and "Very Difficult" - based on environmental differences such as altitude, plains, and mountainous areas. It then developed an intelligent icing judgment model for new energy generator sets, based on deep learning algorithms, tailored to each category. This model, utilizing real-time data collected from each unit such as power generation and wind speed, automatically monitors and comprehensively assesses the icing status of each unit, while dynamically calculating the outage capacity and power loss. Meanwhile, by analyzing historical unit icing and meteorological data, the platform proposes critical thresholds for meteorological elements such as temperature, humidity, wind speed, and rain and snow, which are closely related to the occurrence of icing, and constructs an icing prediction model. This model can dynamically forecast the proportion of new energy generator sets that will experience icing in various cities in Henan Province in the next three days.