The Central Electricity Authority (CEA) has released extensive guidelines to forecast the energy demand in the medium and long terms from discoms, states and union territories. The guidelines aim to facilitate the realistic assessment of future electricity demand from discoms/states by the CEA in its Electric Power Survey of India, aligned with the methodology followed at the central level. It will also serve as a guiding document for such an assessment with respect to new and emerging market segments such as electric vehicles, solar rooftop and green hydrogen.
Key highlights of the guidelines
Term, periodicity, scenarios: The guidelines state that the forecast should be prepared for the medium term (more than one year and up to five years) as well as the long term. The long term forecast should cover the next 10 years, at least. The detailed power demand forecasting exercise should be undertaken every five years. However, the forecast should be reviewed on a yearly basis and updated, if required. According to the guidelines, the base year for the forecast should ideally be taken as the third year (T-3) preceding the year during which the forecast exercise is being carried out. This is to be done to test the performance of the forecasting model by comparing the forecast results obtained for the T-2 and T-1 years with the actual available data (termed as “Out of Sample Validation”). For example, if the forecasting exercise is being done in 2022-23, then the base year for the forecast should be 2019-20 and the performance of the forecasting model should be tested by comparing the forecast results obtained for the years 2020-21 and 2021-22 with the actual available data.
In terms of spatial granularity, the forecasts should be prepared at the discom/state level at the least. In addition, forecasting at more granular levels such as the zonal level, circle level, district level, substation level and feeder/transformer level should also be carried out if adequate granular-level data is available. Such granular forecasts would be more useful in power infrastructure planning. The forecast should be worked out year-wise, at the least. In addition, month-wise/day-wise/hour-wise/time-block-wise forecasts should also be done if adequate granular-level data is available. The forecast should be carried out for at least three scenarios – optimistic, business as usual and pessimistic.
Impact of emerging aspects: The impact of emerging aspects should be quantified in sync with the targets set by the government. In case of non-availability of a target, suitable assumptions should be made, which should be spelled out clearly. As far as possible, the impact of the emerging effects should be apportioned to the corresponding pre-defined consumption categories only – for example, electric vehicle penetration should impact domestic and commercial consumptions, green hydrogen production should impact industrial consumption, solar pump penetration should impact irrigational consumption, etc. In the absence of a suitable category, a new category may be created if the impact is substantial. Otherwise, it should be put in the “Others” category.
Electrical energy requirements of discoms and states: The total electrical energy requirement of a discom should be worked out by adding its distribution losses and the intra-state transmission losses attributed to that particular discom to its total category-wise electrical energy consumption. Meanwhile, the electrical energy requirement of a state should be derived by adding the T&D losses of all the discoms of the state to the sum of their electrical energy consumption. The inter-state transmission losses should be added to the electrical energy requirement of the state at its periphery to determine the energy requirement of a state incident upon the ex-bus of the generators.
Peak demand: Peak demand forecast should be derived from the energy requirement by applying the appropriate load factor. The appropriate load factors in upcoming years should be estimated based on past trends. However, any expected change in the specific consumer mix should also be accounted for. For example, if there is an increase in industrial consumption share, an increase in load factor may be expected.
Checks and balances: This segment highlights the threshold or the range for the load factors and diversity factors of discoms/states. The T&D losses of a state should be equal to the sum of the T&D losses of all its discoms. Unaccounted consumptions, such as those of small discoms, franchisees, temporary connection categories and special categories (ex-Centre-state category in Jammu & Kashmir) should be avoided. The accounting of energy should be cross checked in order to avoid double counting across utilities in cases such as the creation of new states/discoms and the merging of tariff slabs. Data consistency should be cross checked across both the demand and supply side. The periphery’s energy requirements should be the same as the total net generation of the state, and the net imports from outside the state.
Overall, the comprehensive CEA guidelines for long-term and medium-term demand forecasting will provide guidance to power utilities at the state level for carrying out demand forecast holistically. These guidelines will assist in making fairly accurate demand projections, thereby ensuring the availability of reliable energy supply and helping avoid under- or over-utilisation of assets.