REGULATORY BASICS
Electricity DemanD Assessment
Demand assessment is the backbone of a reliable power system since it ensures adequate resources to maintain grid stability. Assessing demand in traditional power systems require operators to plan on a medium to long-term basis considering the commissioning of thermal and hydro power plants that consumed a fair amount of gestation period. Usually, the periodicity for such long-term demand assessments is a minimum of 10 years, estimating the peak demand/load on a monthly or yearly basis while projecting the annual demand growth rates using econometric methods[i]. Some countries like India also estimate demand in a similar way while also simultaneously assessing sectoral consumption demand.
Demand assessment is crucial for the following reasons:
Traditional demand assessment techniques have become relatively less synchronous with the evolution in energy-mix. Renewable integration in power systems with growing changes in demand patterns have highlighted two main drawbacks arising out of traditional demand assessment methods: the inability of current metrics to capture the impact of supply variability and the normalisation of the duck curve[ii] , especially in the case of solar generation; and the lack of spatial and temporal granularity in capturing hourly demand changes. Changes in demand assessment primarily requires a pedagogical review of the classification of various demand assessment methods in short and long terms; these changes must also account for the stochastic factors that affect demand in shorter time frames. Eventually, these demand assessment models become capable of predicting low-probability high-impact events that can be averted.
As far as resource adequacy is concerned, demand assessment is primal since it provides insights into consumption and demand estimates that helps understand the load and its allied risk metrics that calculate the capacity to meet the peak or overall demand. Moreover, it is the very first step in any analysis aimed at ensuring sufficient resources to meet electricity needs. This process involves predicting both the peak power requirements (measured in Megawatts or MWs) and the total energy consumption (measured in Megawatt-hours or MWhs) for various timeframes. These forecasts are crucial for planning in three key horizons: short-term, medium-term, and long-term[iii] .
Accurate demand forecasts are crucial to ensure sufficient resources are available to meet future needs. When demand is expected to increase, additional generation resources or energy purchases become necessary. Two key demand types are considered in resource planning:
- Annual Demand: This refers to the total energy consumption (measured in Megawatt-hours or MWh) of a specific area over an entire year. Due to the influence of weather and seasonal patterns, annual demand is typically broken down by season. Understanding annual demand and its seasonal variations helps in strategically building a resource portfolio.
- Peak Hour Demand: This refers to the highest level of energy consumption (measured in Megawatts or MW) within a specific area during a single hour of the day or year. Identifying trends in peak hour demand helps planners anticipate changes that need to be factored into resource planning strategies.
Electricity demand assessment methods have been constantly evolving. Many countries like India were earlier exercising methods that used to extrapolate on past demand trends and on the consumption elasticity to that of its Gross Domestic Product (GDP). However, with the advent of alternative supply-side and end-use technologies, impact of microeconomic and macroeconomic factors, demography and lifestyle changes, modelling techniques had to evolve to capture the effects of price, income, population, technology and other economic, demographic, policy, and technological variables[iv] .
Conventional demand assessment methodologies that used historical data, growth rates, assumptions, and scenarios to assess demand often outstripped the actual demand, leading to supply overhangs. Moreover, these methods often adopted assumptions based on roadmap figures than actual data and trends; the methodologies did not account for baseline correction that ought to factor in unserved demand to measure anomalies from baseline over years. Another drawback of such assessments was that they had forecast periodicity- a considerable time gap (5-10 years) that affects assumptions, baselines, and other such dynamic factors4.
Demand assessment methods will have to account for non-linear relationships that may exist between dependent and independent variables; a combination of techniques that not only expresses latent demand but one that also includes sensitive parameters of weather, per capita income among other tail risks and seasonal factors that affect the baseline demand.
Demand assessment is the first and most crucial step of any resource adequacy and planning analysis.
It involves the forecasting of peak (MWs) and energy (MUs) requirement for multiple horizons (short/medium/long-term) and considers various input parameters such as historical consumption, consumer categories, weather data, econometric data, policies, and drivers, etc.
Factors that Affect Demand Assessment
A multitude of economic, environmental, demographic, technological and regulatory factors affect the assessment of electricity demand while much of these directly affect the demand patterns due to their nature of uncertainty[v]. The main factors that affect demand assessment are:
Time Horizon of Demand Assessment
Demand assessment methods require constant evolution to capture the effects of technological advancements, price, income, demography, economic, and policy variables and their impact on the overall/peak demand. Additionally, conventional models will have to eventually account for non-linear relationships between dependent and/or independent variables, a combination of techniques to express latent demand and include other tail risks and seasonal factors that tend to affect the baseline demand. Demand assessment models will have to grow out of forecast periodicity, i.e., a considerable time gap between assessment exercises that affects assumptions, baselines, and other dynamic factors.
Demand assessment can be classified into short-term demand estimation and long-term demand forecasting methods based on temporal granularities.
- Demand estimation deals with understanding current or short-term demand patterns. It uses historical data and additional factors like weather, holidays, and present events to get an idea of the amount electricity that is being used at a specific time.
- Demand forecasting, on the other hand, with a broader scope, predicts the medium to long-term future demand for electricity. It does use historical data, but also considers some of the anticipated changes in the economy, population growth and demand-side regulatory changes. This is crucial for planning production levels, inventory management and future infrastructural investments.
Data Requirements for Demand Assessment
Accurate demand forecasting plays a crucial role in utility system planning and operation. This process necessitates a comprehensive dataset encompassing various aspects of electricity and alternative fuel consumption within the service area. Key data categories for effective forecasting include:
- Sales Records: Historical data on electricity sales across multiple years provides a foundation for understanding consumption patterns. This data should be disaggregated by geographical area and customer class (residential, commercial, industrial) to capture variations in demand across different sectors and locations. Additionally, the number of customers in each class and area aids in demand projections.
- Demand Records: Time-series data on power demand measured in megawatts (MW) across various timeframes (days, weeks, months, years) is essential. Analyzing this data helps establish the relationship between electricity sales and the required generation capacity. Disaggregated data, further dividing demand by customer class or sector, offers more granular insights. The shape of the load curve, reflecting peak load variations over time (load profile), informs decisions regarding the most suitable types of generation capacity to meet demand fluctuations.
- Economic and Demographic Data: Historical information on economic performance alongside population or household count data provides context for demand forecasts. This data helps understand the impact of economic growth and population changes on energy consumption.
- Weather Data: Real-time data points on temperature, humidity, wind and rainfall patterns, season and weather event predictability, among others can help in assessing the demand for different regions.
- Economic and Demographic Projections: Utility companies can leverage their own economic and demographic forecasts for their service territory. Alternatively, projections might be obtained from government planning ministries or specialized institutions.
- Energy End-Use Data: Ideal data for end-use analysis includes:
- The number or proportion of households using specific electric appliances.
- The number or proportion of commercial, institutional, or industrial consumers employing different types of electric equipment.
- Electricity consumption per customer per specific end-use (e.g., lighting, heating).
- This data, often referred to as penetration/saturation (e.g., percentage of electrified households) and energy intensity (e.g., kWh/household/year), should ideally be available for each customer class and major end-use category. However, obtaining historical data with this level of granularity can be challenging.
- Data Acquisition Strategies:
- National census documents may offer partial data on appliance ownership or use, primarily for the residential sector.
- In developing countries, government agencies or NGOs may have conducted energy end-use studies or participated in data collection initiatives funded by international aid programs. Despite these efforts, data completeness may not always meet forecasting requirements.
- Conducting new end-use surveys often becomes necessary to obtain the level of detail required for robust end-use forecasts.
[i] Mehra, Meeta, and Aarti G Bharadwaj. 2000. “Demand Forecasting for Electricity.” New Delhi. https://regulationbodyofknowledge.org/wp-content/uploads/2013/03/Mehra_Demand_Forecasting_for.pdf
[ii] Synergy. n.d. “Everything You Need to Know about the Duck Curve.” Synergy. Accessed May 27, 2024. https://www.synergy.net.au/Blog/2021/10/Everything-you-need-to-know-about-the-Duck-Curve.
[iii] Forum of Regulators. 2023. “Report on Resource Adequacy Framework.” https://forumofregulators.gov.in/Data/study/Report%20on%20Resource%20Adequacy%20Framework.pdf
[iv] WEC India. n.d. “Alternate Methodology for Electricity Demand Assessment and Forecasting: Executive Summary.” https://www.wecindia.in/downloads/weci_publications/Executive_Summary_Demand_side_forecasting.pdf
[v] Momani, Mohammad Awad. 2013. “Factors Affecting Electricity Demand in Jordan.” Energy and Power Engineering 05 (01): 50–58. https://doi.org/10.4236/epe.2013.51007.