The MSEPS System and its Applications
The MSEPS SystemThe Multi-Scheme Ensemble Prediction System (MSEPS) was designed with the aim of developing a real-time ensemble prediction system to assist in the deployment and large-scale integration of enewables, especially wind energy and provide future compatible forecasting services to those companies that deal with weather sensitive goods or services. WEPROG's founders had the vision of finding an answer to a problem that has been researched extensively over the past two decades, both in meteorology and wind engineering. With the development of the MSEPS system, this vision has come to life.
The MSEPS is a forecasting approach that provides information about a crucial parameter: physical uncertainty of the weather.
The potential of the approach and need for improvements in the forecasting of wind power is the background for the fast operationalisation of the MSEPS in the autumn of 2003 after 9 months of intensive programming. The MSEPS is using 75 forecasts to produce physical uncertainty parameters of the weather situation for the next 6 days. The MSEPS system is based on experience in meteorology, engineering and super-computing made in the past 25 years.
The MSEPS system's 75 ensemble members are generated with a so-called multi-scheme approach, which uses one numerical weather prediction (NWP) model kernel and generates perturbations by varying the formulation of physical processes inside the NWP model. In the MSEPS it is the initial conditions, advection and fast physical processes that form the differences in the ensemble members, because these processes are most relevant for renewables application such as wind power, solar power, hydropower and hydrological and oceanographic modelling.
State of the Art
It has often been mentioned that ensemble forecasts are more accurate than deterministic forecasts. Although WEPROG is provider of ensemble forecast we do regard this statement as inaccurate and valueless from a forecasting perspective, because it is our belief that in reality it is never possible to hit right with a single forecast at any time.
Especially in wind power context there has to be balance on second basis and every error counts. It is the cost of these errors that matters and this is where intelligent usage of forecasts is the crucial requirement.
To make use of ensemble forecasts adds another dimension to the operation, because it provides the operator with the information about the uncertainty of the weather development and thereby enables him to schedule the dispatch in the most secure and yet cost efficient way by pre-allocating reserves according.
In the market place, the ensemble forecasts assist in generating higher competition, because the knowledge about the uncertainty of the weather development and hence wind power production provides the possibility of more intelligent and cost effective planning and trading.
The ensemble approach is essentially a robustification of the forecasting with embedded uncertainty estimates. The robustification is achieved via a more conservative ensemble mean forecast in which nearly all uncertain information is filtered out. The deviation between the ensemble average and the truth is something that has low predictability. It has been demonstrated over the past few years that an operator cannot rely on one forecast as single forecasts cannot be trusted.
A security margin is required and this is where the ensemble technique adds most value. As the cost of these securities are significant and weather dependent, there is reason to use an objective time dependent quantification of the security requirements derived from ensemble forecasts rather than a constant value determined by past forecast error and average market prices.
In addition, the MSEPS system provides a range of probability of given events and given parameters. The system provides probabilities for all relevant weather related quantities such as precipitation, temperatures, wind conditions and wind power. One of the key features of the approach is that the approach predicts the uncertainty changes down to the shortest possible time scale.
The MSEPS System
The MSEPS system has a wide spectra of applications, because it is developed as a modular system, where an application that requires input from any kind of weather parameters can be integrated in the system in a plug & play manner as a 3rd party application (see figure).
WEPROG has developed a number of applications/modules specifically designed to take full advantage of the ensemble weather data. These are for example:
- Wind Power and Ramp Rate Prediction Modules (WPPM/RRPM)
- Short-term Forecasting with Measurements (iEnKF)
- Data Quality Control and Measurement handling
- Optimisation Module for Trading of Energy (MSOPTI)
- Combined Wind Power and Demand forcast Module (CWPDM)
- Enhanced Transmission Line Management Module (eTLM)