Detailed discussion of the methodology used to generate Ansergy's forecasts for electric utilities
The core of Ansergy's Electric Utility product offering are proprietary, plant level dispatch models for 100% of the United State's generating capacity. Using these models we provide our clients with daily updates as to each plant's, and hence each company's, operational earnings and cash flows. Our models are based on a series of proprietary algorithms to take real market data and process it into useful, investable outputs. In addition to having a clear picture, at both the company and generating facility level, of operating earnings and cash flows, the models produce a clear picture as to the drivers of period-on-period variances from the main performance drivers of pricing, weather and outages.
Ansergy's models were built with an inside the company view of how power generation really functions. This allows our clients to understand, interpret and make trading and investment decisions based on the company�s real operational financial performance, taking the guesswork out of the process.
Ansergy's Power Generation Models
Dispatch, which underlies the models, is the process of determining if a power plant can economically generate into a power market. The North American electric grid utilizes many different generating technologies. Each of those types requires a unique dispatch algorithm. Ansergy deploys 8 generation dispatch models to reflect the unique requirements of each generating technology. As a general rule, a power plant will be dispatched into the grid whenever it is economically feasible.
Ansergy maps 100% of the USA's generating capacity into 57 power price points and 23 natural gas hubs. Each power and gas hub has historical pricing data back to 2009. All data is updated continuously from over 100 market sources, insuring accurate and timely dispatch. Ansergy's electric dispatch forecasts looks backwards using actual market prices; the model uses Day Ahead hourly LMP prices and Day Ahead daily spot gas prices.
One of the primary sources for electric utility earnings volatility is power plant outages, caused either from forced outages (unplanned) or maintenance (planned). A major outage in a high price wholesale market can cause a utility to dramatically miss earnings. Ansergy has teamed up with IIR Energy, the world's leading provider of power plant outage intelligence, to incorporate the affect of power plant outages in its earnings forecasts. Outage statuses for over 5,000 major power plants are incorporated daily into Ansergy's forecasts.
As stated earlier, Ansergy has a unique, proprietary algorithm for each generating technology. The following section summarizes our approach to modeling each of them.
- Natural Gas - Gas plants burn natural gas in either gas turbines or boilers. Heat rates are used to determine if a plant is economical to operate based on current market prices. A heat rate is the efficiency factor of a power plant. Heat rates vary by plant depending upon the equipment configuration, elevation, and ambient air temperature. Ansergy has compiled all of the USA's gas-fired power plant's heat rates. Plants are dispatched, produce and sell power, whenever the plant heat rate is less than the market heat rate. Market heat rates are computed using actual hourly power prices and actual daily gas prices. Gas Plants are noted for their wide range in efficiencies and the high volatility of natural gas prices. These power plants can be economical one hour, or day, and uneconomical the next. Therefore, Ansergy models these plants hourly to capture that volatility and return operational earnings and cash flows in near real-time.
- Coal - Coal is an economical and relatively non-volatile fuel supply. Coal fired plants are often called "base load" meaning they nearly always run. Ansergy has assigned a fixed and variable fuel cost component to each coal power plant. The dispatch algorithm compares the market power price to the plant's variable fuel and operating costs. When the cost to generate is less than the market price the plant is dispatched into the grid.
- Hydro - One of the more complex electric generating technologies to model. Ansergy accurately computes the amount of energy generated, by hour, at each plant using actual stream flows and the plant's H/K. Ansergy�s maps each hydro plant to the relevant USGS river stations. These stations report hourly stream flows and each station is immediately below the plant's turbines. Each hydro power plant has a unique Head Over Constant factor (H/K), which is the efficiency measurement of converting water to power.
- Oil and Fuel/Switching - Much of the nation's older generating fleet uses oil-based fuels. These plants are noted for their high variable costs and poor efficiencies (high heat rates). Only rarely will oil-fired plants generate power into the US grid as energy plants; instead they are often used as capacity resources. Modeling an oil plant requires first converting the oil into BTUs, and then dispatching using the natural gas model's algorithm. Some oil plants can burn either gas or oil. These plants require an additional calculation to determine the most economical fuel. The most economical fuel is selected and determines the plant's dispatch price.
- Pumped Storage - These plants are essentially physical swing options and a small portion of the nation's generating capacity. When electric prices are low water is pumped uphill into storage. When prices are high that stored water is released and power is generated for sale. These plants can only operate when the spread between low prices (off peak hours and weekends) is less than the high price hours (on peak) minus the energy lost in pumping and generating. Ansergy deploys a look ahead algorithm, using actual power prices, to compute the hours the plant pumps, and the hours the plant generates.
- Wind - The fastest growing source of electric generating capacity in the US is wind. Wind turbines produce electricity without burning hydro carbons. Wind only creates power when there is wind in the right speed ranges; they do not operate if the wind speed is either too high or too low. To model these plants Ansergy uses actual hourly wind speed and converts MPH to MWh using standard conversions appropriate for that turbine.
- Nuclear - The Nuclear Regulatory Agency requires significant disclosure on the costs and operating characteristics of the nation's fleet of nuclear reactors with each plant being required to report its daily operating status. Ansergy dispatches the nuclear power plants using the NRC daily status report regardless of market prices or costs to generate.
- Other - This catch-all category includes bio-mass, solar, and other relatively smaller producing plants. Ansergy assigns operating costs to each plant and dispatches them into the market against those prices.
Ansergy's Retail Load Model
Electric utilities receive operating revenues from both generating electric power and selling that power into the wholesale market and/or buying power in the wholesale market and reselling that power to retail customers. Though some utilities only generate, or only serve load, most do both. This leaves the utility net short or net long power, almost every hour of the year, and the price differentials between the wholesale and retail markets being a primary source of earnings volatility.
Ansergy determines a utility's net position by first computing total generation by hour then computing its hourly load. Loads are driven by weather and macro economic factors. Ansergy deploys two distinct methodologies in determining a utility's hourly loads, depending on where the utility serves its load.
- ISO Utilities - Most of the US power grid operates under an Independent System Operator, or ISO. An ISO is responsible for managing the power markets within a region. The US currently has six ISO serving about 70% of the US grid. ISOs report hourly power prices and hourly loads. A utility's retail load will be a component of the total ISO load. Ansergy computes the ratio of a utility to ISO using actual loads reported by the utility. Hourly utility loads are extrapolated using actual ISO loads and the utility's share of total ISO load. Ansergy tracks an ISO utility's actual loads hourly with less than a 2-3 hour lag.
- Non-ISO Utilities - The other 30% of the nation's retail load is served by utilities that do not participate in an ISO. Forecasting loads for these utilities requires regressing actual hourly loads against actual weather (temperature, humidity, cloud cover, etc) using a series of proprietary equations. Summer loads behave differently than winter loads; week end loads behave differently than week day loads; night time loads behave differently than day time loads. Ansergy uses 176 load shapes by utility. Loads are forecasted using actual weather and the 176 load regression equations.
Utilities sell power to retail customers that include residential, commercial, industrial and other types. The rates charged vary by customer and often vary by usage and time of day. Ansergy uses a system wide retail rate compiled from filed financial statements.
All hourly loads are purchased in the appropriate wholesale market. All generation is sold in the appropriate hourly wholesale market. Often the two wholesale markets are not the same. Ansergy's solution accurately computes the affect of being short or long by turning all generation and retail load positions into cash flows. There are researchers who simplify the process by examining one, or perhaps a few, of the components. Ansergy's solution accurately determines how the utility operates its system and how that system drives the bottom line.