The challenge of integrating offshore wind power in the U.S. electric grid. Part I: Wind forecast error
Not Available
SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Not Available
As electricity systems transition to higher levels of solar and wind generation, electric system operators will likely need to hold additional reserves to manage solar and wind forecast error. Because solar and wind forecast errors tend to be weakly correlated across space, system operators can reduce their reserve requirements by sharing reserves. This paper examines the benefits of forecast error reserve sharing among balancing areas in the Southeastern United States, in scenarios in which solar and wind generation ranges from 34% to 65% of total generation. It finds that day-ahead forecast error reserve requirements increase linearly with growth in solar and wind generation capacity (6%-10% of total capacity), but that reserve sharing can significantly reduce these requirements (by 6%-29%). It finds that, in economic terms, the value of forecast error reserve sharing ($\$$0.09-$\$$1.24 billion per year, $\$$0.12-$\$$1.68/MWh of load across scenarios) tends to decline with higher levels of solar and wind generation, due to lower reserve and energy prices. Even with declines in reserve prices, forecast error reserve sharing can still provide substantial value, though with higher levels of solar, wind, and electricity storage this value is increasingly tied to avoiding scarcity prices.
As electricity systems transition to higher levels of solar and wind generation, electric system operators will likely need to hold additional reserves to manage the forecast error associated with these resources. Because wind and solar forecast errors tend to be poorly correlated across space, system operators can reduce their reserve requirements by sharing reserves. This paper examines the value of forecast error reserve sharing among balancing areas in the Southeast United States. It finds that forecast error reserve requirements increase linearly with growth in solar and wind generation capacity but that reserve sharing can significantly reduce physical (MW) reserve requirements (from 25%-26% to 18%-19% of average load in high solar scenarios). It finds that the value of forecast error reserve sharing declines with higher levels of solar and wind generation, due to lower wholesale energy and reserve prices. Even with declines in wholesale prices, forecast error reserve sharing can still provide substantial value (as much as $\$$400 million per year in a high solar scenario), though with higher levels of solar, wind, and electricity storage, this value is increasingly tied to avoiding scarcity prices. The results suggest the importance of coordinated capacity expansion planning for forecast error reserve sharing.
In this paper, we examine the shape of the persistence model error distribution for ten different wind plants in the ERCOT system over multiple timescales. Comparisons are made between the experimental distribution shape and that of the normal distribution.
Wind power forecasting is one important tool in the integration of large amounts of renewable generation into the electricity system. Wind power forecasts from operational systems are not perfect, and thus, an understanding of the forecast error distributions can be important in system operations. In this work, we examine the errors from operational wind power forecasting systems, both for asingle wind plant and for an entire interconnection. The resulting error distributions are compared with the normal distribution and the distribution obtained from the persistence forecasting model at multiple timescales. A model distribution is fit to the operational system forecast errors and the potential impact on system operations highlighted through the generation of forecast confidenceintervals.
Dispersion and aggregation of wind power plants lower the uncertainty of wind power by reducing wind power forecasting errors. Using quantitative methods, this paper studies the dispersion's impact on the uncertainty of the aggregated wind power production. A method to simulate day-ahead forecast error distributions at different dispersion and forecasting skill scenarios is presented. The proposed method models the uncertainty of wind power forecasting on an annual basis and at different levels of production. As a result, the uncertainty in the forecasting of spatially dispersed wind power plants is modelled using two continuous distributions: Laplace and beta distribution. The analysis shows that even the production level uncertainty can be modelled in various dispersion and forecasting skill scenarios. The model for aggregated forecast error distributions requires only four variables: capacity-weighted distance of the total wind power plant fleet, mean of the site-specific mean absolute errors (MAEs), number of aggregated wind power plants, and the mean variability of the elevations from the proximities of the aggregated wind power plants. Here, the results are especially promising when the number of aggregated wind power plants exceeds five, the terrain complexity is low or moderate, and the aggregation region is large. This is demonstrated through a case study for Texas, United States.
The power system balancing process, which includes the scheduling, real time dispatch (load following) and regulation processes, is traditionally based on deterministic models. Since the conventional generation needs time to be committed and dispatched to a desired megawatt level, the scheduling and load following processes use load and wind power production forecasts to achieve future balance between the conventional generation and energy storage on the one side, and system load, intermittent resources (such as wind and solar generation) and scheduled interchange on the other side. Although in real life the forecasting procedures imply some uncertainty around the load and wind forecasts (caused by forecast errors), only their mean values are actually used in the generation dispatch and commitment procedures. Since the actual load and intermittent generation can deviate from their forecasts, it becomes increasingly unclear (especially, with the increasing penetration of renewable resources) whether the system would be actually able to meet the conventional generation requirements within the look-ahead horizon, what the additional balancing efforts would be needed as we get closer to the real time, and what additional costs would be incurred by those needs. In order to improve the system control performance characteristics, maintain system reliability, and minimize expenses related to the system balancing functions, it becomes necessary to incorporate the predicted uncertainty ranges into the scheduling, load following, and, in some extent, into the regulation processes. It is also important to address the uncertainty problem comprehensively, by including all sources of uncertainty (load, intermittent generation, generators’ forced outages, etc.) into consideration. All aspects of uncertainty such as the imbalance size (which is the same as capacity needed to mitigate the imbalance) and generation ramping requirement must be taken into account. The latter unique features make this work a significant step forward toward the objective of incorporating of wind, solar, load, and other uncertainties into power system operations. In this report, a new methodology to predict the uncertainty ranges for the required balancing capacity, ramping capability and ramp duration is presented. Uncertainties created by system load forecast errors, wind and solar forecast errors, generation forced outages are taken into account. The uncertainty ranges are evaluated for different confidence levels of having the actual generation requirements within the corresponding limits. The methodology helps to identify system balancing reserve requirement based on a desired system performance levels, identify system “breaking points”, where the generation system becomes unable to follow the generation requirement curve with the user-specified probability level, and determine the time remaining to these potential events. The approach includes three stages: statistical and actual data acquisition, statistical analysis of retrospective information, and prediction of future grid balancing requirements for specified time horizons and confidence intervals. Assessment of the capacity and ramping requirements is performed using a specially developed probabilistic algorithm based on a histogram analysis incorporating all sources of uncertainty and parameters of a continuous (wind forecast and load forecast errors) and discrete (forced generator outages and failures to start up) nature. Preliminary simulations using California Independent System Operator (California ISO) real life data have shown the effectiveness of the proposed approach. A tool developed based on the new methodology described in this report will be integrated with the California ISO systems. Contractual work is currently in place to integrate the tool with the AREVA EMS system.
Wind power forecasting is expected to be an important enabler for greater penetration of wind power into electricity systems. Because no wind forecasting system is perfect, a thorough understanding of the errors that do occur can be critical to system operation functions, such as the setting of operating reserve levels. This paper provides an international comparison of the distribution of wind power forecasting errors from operational systems, based on real forecast data. The paper concludes with an assessment of similarities and differences between the errors observed in different locations.
This is a report for task one of the tail event analysis project for BPA. Tail event refers to the situation in a power system when unfavorable forecast errors of load and wind are superposed onto fast load and wind ramps, or non-wind generators falling short of scheduled output, the imbalance between generation and load becomes very significant. This type of events occurs infrequently and appears on the tails of the distribution of system power imbalance; therefore, is referred to as tail events. This report analyzes what happened during the Electric Reliability Council of Texas (ERCOT) reliability event on February 26, 2008, which was widely reported because of the involvement of wind generation. The objective is to identify sources of the problem, solutions to it and potential improvements that can be made to the system. Lessons learned from the analysis include the following: (1) Large mismatch between generation and load can be caused by load forecast error, wind forecast error and generation scheduling control error on traditional generators, or a combination of all of the above; (2) The capability of system balancing resources should be evaluated both in capacity (MW) and in ramp rate (MW/min), and be procured accordingly to meet both requirements. The resources need to be able to cover a range corresponding to the variability of load and wind in the system, additional to other uncertainties; (3) Unexpected ramps caused by load and wind can both become the cause leading to serious issues; (4) A look-ahead tool evaluating system balancing requirement during real-time operations and comparing that with available system resources should be very helpful to system operators in predicting the forthcoming of similar events and planning ahead; and (5) Demand response (only load reduction in ERCOT event) can effectively reduce load-generation mismatch and terminate frequency deviation in an emergency situation.
The examination of wind power forecasting errors is crucial for optimal unit commitment and economic dispatch of power systems with significant wind power penetrations. This scheduling process includes both renewable and nonrenewable generators, and the incorporation of wind power forecasts will become increasingly important as wind fleets constitute a larger portion of generation portfolios. This research considers the Western Wind and Solar Integration Study database of wind power forecasts and numerical actualizations. This database comprises more than 30,000 locations spread over the western United States, with a total wind power capacity of 960 GW. Error analyses for individual sites and for specific balancing areas are performed using the database, quantifying the fit to theoretical distributions through goodness-of-fit metrics. Insights into wind-power forecasting error distributions are established for various levels of temporal and spatial resolution, contrasts made among the frequency distribution alternatives, and recommendations put forth for harnessing the results. Empirical data are used to produce more realistic site-level forecasts than previously employed, such that higher resolution operational studies are possible. This research feeds into a larger work of renewable integration through the links wind power forecasting has with various operational issues, such as stochastic unit commitment and flexible reserve level determination.
Machine learning is quickly becoming a commonly used technique for wind speed and power forecasting. Many machine learning methods utilize exogenous variables as input features, but there remains the question of which atmospheric variables are most beneficial for forecasting, especially in handling non-linearities that lead to forecasting error. This question is addressed via creation of a hybrid model that utilizes an autoregressive integrated moving-average (ARIMA) model to make an initial wind speed forecast followed by a random forest model that attempts to predict the ARIMA forecasting error using knowledge of exogenous atmospheric variables. Variables conveying information about atmospheric stability and turbulence as well as inertial forcing are found to be useful in dealing with non-linear error prediction. Streamwise wind speed, time of day, turbulence intensity, turbulent heat flux, vertical velocity, and wind direction are found to be particularly useful when used in unison for hourly and 3 h timescales. The prediction accuracy of the developed ARIMA–random forest hybrid model is compared to that of the persistence and bias-corrected ARIMA models. The ARIMA–random forest model is shown to improve upon the latter commonly employed modeling methods, reducing hourly forecasting error by up to 5 % below that of the bias-corrected ARIMA model and achieving an R 2 value of 0.84 with true wind speed.
This paper presents four algorithms to generate random forecast error time series, including a truncated-normal distribution model, a state-space based Markov model, a seasonal autoregressive moving average (ARMA) model, and a stochastic-optimization based model. The error time series are used to create real-time (RT), hour-ahead (HA), and day-ahead (DA) wind and load forecast time series that statistically match historically observed forecasting data sets, used for variable generation integration studies. A comparison is made using historical DA load forecast and actual load values to generate new sets of DA forecasts with similar stoical forecast error characteristics. This paper discusses and compares the capabilities of each algorithm to preserve the characteristics of the historical forecast data sets.
This presentation presents some statistical analysis of wind power forecast errors and error distributions, with examples using ERCOT data.
Given the multi-scale variability and uncertainty of wind generation and forecast errors, it is a natural choice to use time-frequency representation (TFR) as a view of the corresponding time series represented over both time and frequency. Here we use wavelet transform (WT) to expand the signal in terms of wavelet functions which are localized in both time and frequency. Each WT component is more stationary and has consistent auto-correlation pattern. We combined wavelet analyses with time series forecast approaches such as ARIMA, and tested the approach at three different wind farms located far away from each other. The prediction capability is satisfactory -- the day-ahead prediction of errors match the original error values very well, including the patterns. The observations are well located within the predictive intervals. Integrating our wavelet-ARIMA (‘stochastic’) model with the weather forecast model (‘deterministic’) will improve our ability significantly to predict wind power generation and reduce predictive uncertainty.
This paper presents a first-order autoregressive algorithm to generate real-time (RT), hour-ahead (HA), and day-ahead (DA) wind and load forecast errors. The methodology aims at producing random wind and load forecast time series reflecting the autocorrelation and cross-correlation of historical forecast data sets. Five statistical characteristics are considered: the means, standard deviations, autocorrelations, and cross-correlations. A stochastic optimization routine is developed to minimize the differences between the statistical characteristics of the generated time series and the targeted ones. An optimal set of parameters are obtained and used to produce the RT, HA, and DA forecasts in due order of succession. This method, although implemented as the first-order regressive random forecast error generator, can be extended to higher-order. Results show that the methodology produces random series with desired statistics derived from real data sets provided by the California Independent System Operator (CAISO). The wind and load forecast error generator is currently used in wind integration studies to generate wind and load inputs for stochastic planning processes. Our future studies will focus on reflecting the diurnal and seasonal differences of the wind and load statistics and implementing them in the random forecast generator.
Abstract not provided
This presentation offers new data and statistical analysis of wind power forecasting errors in operational systems.
One of the critical challenges of wind power integration is the variable and uncertain nature of the resource. This paper investigates the variability and uncertainty in wind forecasting for multiple power systems in six countries. An extensive comparison of wind forecasting is performed among the six power systems by analyzing the following scenarios: (i) wind forecast errors throughout a year; (ii) forecast errors at a specific time of day throughout a year; (iii) forecast errors at peak and off-peak hours of a day; (iv) forecast errors in different seasons; (v) extreme forecasts with large overforecast or underforecast errors; and (vi) forecast errors when wind power generation is at different percentages of the total wind capacity. The kernel density estimation method is adopted to characterize the distribution of forecast errors. The results show that the level of uncertainty and the forecast error distribution vary among different power systems and scenarios. In addition, for most power systems, (i) there is a tendency to underforecast in winter; and (ii) the forecasts in winter generally have more uncertainty than the forecasts in summer.