Apr 2016
Electricity Distribution
Solar PV Forecasting Phase 1
Apr 2016
Apr 2018
National Grid Electricity System Operator
David Lenaghan
Click here to send a question to the contact.
Network Innovation Allowance
ED - Network improvements and system operability
Carbon Emission Reduction Technologies, Distributed Generation and Modelling
The scope of the project is to both support Numerical Weather Prediction (NWP) improvements and establish effective techniques to improve upon raw NWP output, both of which are areas the Met Office have scientific expertise in.

The objectives of the project are to reduce demand forecast error attributable to embedded solar PV as a result of researching improvements in solar irradiance forecasting methods.

In order to achieve the objectives, the  following work packages will be pursued and delivered:

Post Processing Work Packages:

WP 1 – Refinements to the existing solar radiation forecasting capability to best suit National Grid’s requirements

Clear evidence shows the Met Office Global and Regional Ensemble Prediction System (MOGREPS) UK ensemble has a tendency to be too cloudy, which has a particularly strong negative impact upon solar radiation forecasts. For Day +1 forecasts, the Met Office will test replacing MOGREPS-UK with the global MOGREPS-G (Global) ensemble output which demonstrably performs well from Day +2 onwards. They will also test replacing the MOGREPS-UK ensemble mean with a blended forecast using multi-model solar radiation, including output from the Met Office’s high resolution deterministic Numerical Weather Prediction (NWP). If testing demonstrates these changes are positive, the Met Office will implement these improvements by Summer 2016.

WP 2 – Investigation and development of statistical post-processing techniques

A number of techniques exist that can add value to raw forecasts through learning systematic error characteristics from making use of existing radiation observations. The Met Office currently deploy Kalman Filtering to temperature and wind speed.  Application of a Kalman Filter; perhaps to normalised solar radiation, will be tested and assessed.  The Met Office will also explore other techniques such as Model Output Statistics (MOS) and Neural Nets and potential combinations of techniques. Where these demonstrate added value, the aim will be to develop them into the Met Office operational environment by Summer 2017.   A report of these methods will be published as part of the project deliverables.

WP 3 – Development of a gridded solar radiation Nowcast 

Accuracy in the 4-6 hour period is operationally significant for National Grid. The Met Office have an hourly updating Nowcast capability that combines observational data from both weather stations and satellite imagery with the latest high resolution NWP for a number of key parameters. This work package would extend that capability to solar radiation, for which both station and satellite measurements offer the clear potential for improving the fit to reality in the earliest hours of the forecast. The Met Office would also explore whether real-time available PV readings could be ‘reverse-engineered’ to give an additional observational source of solar radiation, with good spatial coverage, that could be integrated into the Nowcast.  The aim would be to operationalise by Summer 2017.

Core Science Work Package:
WP 4 – Focussed development of core NWP cloud/radiation schemes
This work package aims to:

  • Perform research into model development that could lead to improvements in the forecasting of cloud and radiation at the surface.
  • Deliver a report on research and development aimed at improving the representation of clouds, radiation and their interaction within the forecasting system.
  • Provide a reference for future improvements in the MOGREPS forecasting system that would directly benefit the irradiance forecasts National Grid receive.
This project’s aim is to drive improvement in the key climatological variable that feeds National Grid’s PV generation forecast model. These Improvements can be more accurately measured by comparing National Grid PV forecasts derived on project completion with the PV outturn data calculated using the method given by the NIA PV Monitoring Phase 2 project (NIA_NGET0170).

Post Processing Work Packages

  • Immediate improvement in National Grid’s operational PV power forecasts, including those published for industry visibility.

  • Reduction of demand forecast error. This would lead to reduced balancing costs for managing the current level of solar PV on the system and reducing the costs to consumers of introducing further solar PV on the power system.

Core Science Work Package:

This work package would support and boost research and potential development in the fundamental cloud/radiation physics at the heart of our Numerical Weather Prediction (NWP) models. Where, as a result of this focus, the Met Office identify improvements, they will work to develop these into their operational capability as means to provide improved solar irradiance forecasting. However, the Met Office has to ensure that enhancements do not adversely affect other applications of NWP and for this reason cannot commit to specific timelines. A progress report will be provided in the research and development pertinent to solar radiation forecasting at appropriate intervals through the life of the project.