From 10 Days to 20 Minutes: AI Speeds Up Coal Inventory Calculations at Samruk-Energo

Artificial intelligence is becoming one of the key tools in the digital transformation of Samruk-Energo. AI technologies are already being applied to enhance occupational safety, equipment diagnostics, digital inspection of energy infrastructure, and electricity generation forecasting.

One of the most sought-after solutions is intelligent video analytics. At energy facilities, systems analyze video streams in real-time and automatically identify violations of occupational safety requirements, lack of personal protective equipment, unsafe behavior by personnel, and the presence of people in hazardous areas. This enables prompt responses to potential risks and enhances workplace safety.

A significant focus is predictive equipment diagnostics. At Ekibastuz GRES-1 (State Regional Electric Station), a project is underway to detect defects using machine learning. This solution analyzes historical and current data, identifies deviations, and helps detect signs of possible malfunctions at an early stage. The approach aims to increase equipment reliability, reduce the risk of unplanned shutdowns, and minimize downtime.

AI is also utilized for digital inspection of energy infrastructure. Unmanned aerial vehicles equipped with computer vision and AI technologies are used to survey facilities. They provide visual and thermal imaging data, which is then analyzed automatically to identify potential defects. This reduces the need for staff involvement in potentially dangerous high-altitude work.

The practical impact of these technologies is markedly reduced time for certain operations. For instance, the calculation of coal reserves is reduced from 10 days to just 20 minutes while maintaining high accuracy. Inspections of transmission lines, poles, and transformers are accelerated 5 to 10 times faster. During chimney inspections, the use of drones reduces inspection time to two hours, eliminating the need for prolonged high-altitude work. Analysis of aerial survey data of hydraulic structures, in turn, enables timely detection of potential damages.

A separate initiative is focused on the development of renewable energy. Samruk-Energo is implementing AI models to forecast electricity generation 24 to 72 hours ahead. When building forecasts, meteorological data, as well as information from AMR (Automated Meter Reading) and SCADA systems, are considered.

Forecasting plays a crucial role for wind farms, where the volume of generation directly depends on weather conditions, particularly wind speed. The project's goal is to improve the accuracy of generation forecasts to 80%, enhance operational planning, and reduce deviations of actual generation from predicted values.

The integration of AI spans various processes within Samruk-Energo, from ensuring safety and monitoring equipment conditions to inspecting infrastructure and managing generation. Using such solutions increases the efficiency of production processes, the reliability of energy assets, and the quality of operational management.

Photo: gov.kz