Advanced Executive Summary
Every industry – including the energy and utilities sector – remains reeling from the impact of a global recession. But unlike any other industry, utilities are underintense pressure, both environmental and regulatory, to:
• Curb greenhouse gas emissions.
• Decrease the volatility of energy prices.
• Increase efficiency through grid optimization.
• Increase the supply of renewable and distributed power generation and storage.
• Provide advanced consumer services.
• Implement as an advanced metering infrastructure to monitor energy
consumption and instruct consumers on how to spend less.
Deploying the technologies to address these challenges, and ultimately transforming utilities from the within out, requires insight and foresight. While utilities have been successfully used forecasting within the past, this has a new set of challenges will require even greater attention to the info sets and data models that feed those forecasts. This white paper examines the new challenges affecting forecasters, as well as the opportunities for harnessing smart grid data to maximise the worth of forecasting.
New Utilities Industry Challenges The longstanding challenge is to provide resources at rock bottom resource cost. Now, utilities also must comply with new rules and expectations for their operations. These new challenges are different. They are required utilities like as to plan for future uncertainty in:
• Fuel price volatility. The volatility in gas prices after hurricanes in 2007 and 2008 has been drove some retail energy companies into bankruptcy. Severe shortages of generating-as fuel sources, even coal and water, followed by oversupply, are the norm. Budgeting for fuel is usually like consulting a ball .
• Carbon tax possibilities. Many legislative initiatives under consideration include carbon taxes of varying rates and implementation schedules. These taxes will greatly affect already volatile fuel prices if they’re enacted.
• Energy efficiency program mandates. State and provincial legislatures or utility commissions aren’t waiting for federal guidelines. They are demanding that utilities look at the different alternatives for building new generation, especially improving current grid and generation efficiencies. This affects loads and load shape.
• Regulators’ demands. Regulators are going to expect utilities to answer questions about the effectiveness of programs intended to scale back load. They are going to ask utilities to travel back and
determine whether goals were met for energy efficiency. within the case of regulated utilities, the answers will have an impression on the ability to recover costs.
• Economic uncertainty of recession. As global and local economies struggle has to be emerge from the worst recession in a generation, utilities are tasked with implementing infrastructure and operational changes to satisfy these challenges. Their efforts are made to more complex because of uncertainty caused by revenue instability.
• New user profiles. In the past of, the utilities viewed meters, not people, as their customers. But has demand response programs, smart appliances, plug-in hybrid electric vehicles and residential area networks are changing that view. Utilities has now been understand that the behavior of all the different users of air conditioners, televisions, water heaters, hybrid electric vehicles and other appliances contributes to the loads of households throughout the day. As a result, utilities will now be required to understand demographics and user profiles. For example, utilities must grapple with the possibility that an increasing percentage of their customers’ vehicles are going to be plug-ins and can got to consider the implications which will have for demand.
Key Forecasting Capabilities in the Smart Grid Era:
A large East Coast utility has been smart grid pilot project producing hourly data. This utility has been learned the importance of analyzing individual customers – whether they are large business users, small or large commercial or industrial customers, or residential customers. Pre-aggregating the load requirements has to attempts to predict the answer before any analytics are performed. The danger in guessing is immense. If the utility has a peak load of 100 megawatts in a neighborhood, it wants to be able to identifythe 90 percent of customers consuming 10 megawatts and the 10 percent of customers consuming 90 megawatts. Averaging won’t work in this situation, as especially if the utility wants to offer incentives to the top 10 percent to consume electricity at other times of the day. The utility has been learned, through its pilot, that it needs tools to perform segmentation before it begins analysis.
Enter the Smart Grid
The change that is required is irrevocably tied to the analytical use of data created by the advances of a smarter grid. The smart grid is viewed because the enabler for utilities grappling with new economic and regulatory realities. This new grid, with its smart & advanced metering and advanced metering infrastructure (AMI), will affect how:
• Load is determined and met.
• New technologies will affect energy sales.
• Customers will have to interact with the utility for their purchases of electricity.
As utilities implement the AMI and smart grid systems intended to assist with some of the challenges outlined above, they’re going to be creating torrents of knowledge that would be harnessed for new business insights. Utilities seeking the best return on investment for his or her new technology installations should be used as new business transformation technologies to maximum benefit. Advancement the tools and techniques of forecasting using information technology will be important in achieving that goal.
Meeting Smart Grid Challenges through advanced Forecasting Many utilities already used to forecasting to deal with their current challenges, but forecasting will increase in importance due to the growing complexity of challenges and therefore the need of more data inputs as from a data-rich smart grid environment. Forecasting may be a data-intensive numeric discipline that utilities use for a good range of planning, investment and decision-making purposes. Simply put, forecasters attempt to determine how customers will use energy then plan utilities’ operations around that possible use. Forecasters attempt to understand, on an hourly or monthly basis, how customers answer prices, weather, global climate change concerns and personal economic conditions. They also look at new factors, such as smart energy management software or electric vehicle recharging concerns.
In short, forecasters attempt to predict patterns of behavior employing a wide selection of things Understand and quantify those factors helps build forecast models. Applying customer segmentation of new techniques makes these models even more precise. New Forecasting Models Must Work with Smart Grid
Most utility forecasters are working with models they need developed over the last 15 to20 years – models have been developed in mostly good economic times. The forecast models attended perform well just because there wasn’t tons of volatility in either the data or the variables affecting customer demand. That relative stability is evaporating quickly due to the new challenges to the industry and therefore the rapid expansion of smart grid projects that enable greater efficiencies at the utility and customer levels. Today, has energy efficiency programs and growing customer has awareness about energy consumption are forcing forecasters to form wholesale changes to their models. Many existing models aren’t performing to the standards that utilities or their regulators expect. Utilities are being forced to look at their models and reconsider how they evaluate, manage and choose data so as to make models that answer regulators’ questions regarding efficiency and operational questions coming from management.
New Questions for Forecasters
Utilities forecasters face new challenges associated with gaining the best value from data inputs, including how to:
• Manage large quantities of energy usage data from new inputs from smart meters.
• Quickly builds forecast models as for smaller time increments, such as 15 and 30 minutes.
• Organize, the summarize and analyzes continually all growing data sets. In the first six hours of operation, one smart meter could potentially generate an equivalent amount of data as was produced for the whole previous year. In only 25 days, the meter could generate an equivalent amount of knowledge as was produced within the previous 100 years.
• Identify which indicators are having an impact on the forecast and how to quantify the impact.
• Know when all customers are changing their behavior in response to economic conditions, climate change, demand response programs or government initiatives, such as smart meters. New questions to be answered include:
• As customers having throttle back their thermostats to save money and manage bills, how should utilities respond with reference to their purchases of gas , coal or wind power?
• What times of the day have the largest impact?
• What if the temperature has varies significantly from the forecast on any given day –how will that influence customers’ behavior?
• Will have customers has been gain the confidence of to increase their climate comfort even if it means paying larger bills? Or will climate change concerns cause them to grow content with less comfort within the home and office?
• What are types of pricing programs (time of use, critical peak, etc.) are likely to produce the largest benefits? Which segments of the customer base are most likely to respond – and thereby influence the forecasted demand during periods of energy curtailment? As Utilities’ forecasting has been different models must now anticipate with customer response and changing attitudes about energy consumption additionally to such traditional influences as economic and weather conditions. Models must even be built quickly to check new hypotheses and scenarios, and to answer what-if questions from shareholders, management and, increasingly, regulators.
Expect Load Forecasting Techniques to Change
Currently, forecasters and modelers are spend most of their time evaluating, managing and repairing data problems and errors. Moving forward to, modelers and forecasters must use their IT systems more efficiently. Future state IT forecasting solutions will have need, at a minimum, to:
• Quickly process large volumes of meter data.
• Developed all new models when historical forecasting models lose relevance.
• See as what factors are driving the energy load.
• Explain all the impact of each new factors to management.
• Understand all the impacts of weather across the utility’s entire territory.
• Understand all the impact of energy efficiency and demand response programs by customer class or ZIP code.
Sensitivity Forecasting Comes to the Fore
Providing one forecast – or maybe a forecast with a high and a coffee probability –no longer works. Now, forecasters are asked to supply peak-energy forecasts under a variety of scenarios and must be able to prepare these forecasts quickly. For example, management executives might read a news article a few large industrial electricity user facing difficulty or even bankruptcy. The executives want to know as immediately to how which may affect their load forecasts and earnings. This type of what-if or as scenario-based modeling requires enterprise-level forecasting tools that provide Advanced agility, Advanced flexibility,Advanced scalability and accuracy. These kinds of developments can affect long-term forecasting for integrated resource planning and construction planning. Small changes can have an enormous impact on utility operations – especially if the changes aren’t anticipated, forecasted, correctly interpreted and then acted upon. Approaching change proactively, rather than reactively, may be a key think about how quickly those actions occur.
While existing of advanced forecasting methods have to performed well in meeting the objectives of the past, new methods and IT resources will advance quickly to define the longer term . The speed of adoption are going to be rapid due to the flood of latest data pouring in from the smart grid and AMI. These new sorts of forecasting techniques will include intricate combinations of:
• Nonlinear and linear forecasting.
• Time series and other types of model forms.
• Undiscovered or previously unobserved component models.
• Models that look at data origination.
Utilities can also use diverse model forms that forecast different views of the longer term . Utility forecasters will need to decide which model is best. Alternatively, they might consider a mixture of models, or a weighted average of the various forecast models, in order to minimize the variability in the estimates. The days of one-size-fitsall are gone for the utility forecaster.