VIJAYAWADA : Andhra Pradesh energy department, which has successfully developed ‘Day Ahead Power Demand’, forecasting using artificial intelligence and machine learning to optimise power purchase cost for the benefit of customers and won the SKOCH Silver Award in the Power and Energy sector for the year 2023, is now coming up with two more innovative tools for optimal resource mobilisation and reduce the financial burden on the department.
The two new projects—Intra-State Open Access Settlement and Unit Commitment Model—have been developed using various computer programming languages like C and Python. While the Intra-State Open Access Settlement programme is currently being implemented, the Unit Commitment Model is being fine tuned.
Addressing the department staff, Special Chief Secretary (Energy) K Vijayanand said that Andhra Pradesh is the first State in India to develop ‘the most accurate’ day-ahead forecast model using artificial intelligence and machine learning. APTRANSCO, which has reduced the transmission losses to under 3 per cent, which was at 2.69 per cent, and is striving for further reduction of transmission losses.
In 2020, the State Load Dispatch Centre of APTRANSCO under the energy department had developed the model with an objective for better power management, minimising power purchase cost. “The forecast is made for the next 24 hours, which are divided into 96 time blocks of 15 minutes each. To ensure occurrence, 25 years data of various parameters like climate, holidays, day developments, unforeseeable variables among others are taken into consideration. Using AI, an accurate price forecast for 96 time blocks has been designed. This helps the officials to make informed decisions while purchasing power from the market,” a senior official from APTRANSCO explained.
After its successful implementation, the department has come up with another innovative tool namely Intra-State Open Access Settlement. Stating that wind and solar power is being purchased from different consumers and pricing is not uniform, he said that this new tool reduces complexity and time, further reducing the expenditure incurred.
Highlighting the role of the Unit Commitment Model, the senior official said that the new tool is for forecasting the demand for the next day. “To meet the demand, power has to be sourced out from different projects like thermal, hydro electric, solar and wind projects within the State besides nuclear power projects and power markets. Different projects generate power differently and the demand for the day varies. The new tool, which is in development stage, will help in deciding what to do in case of deficit production and what to do in case of surplus,” he explained.
Forecasting demand for next 24 hours
‘Day Ahead Power Demand’ forecasts energy demand for next 24 hours, which are divided into 96 time blocks. Using AI, an accurate price forecast for 96 time blocks has been designed, which helps the officials to make informed decisions while purchasing power