- Home
- NIA_NGSO0008
Solar PV Monitoring Phase 3
- Status:
- Complete
- Project Reference Number:
- NIA_NGSO0008
- START DATE:
- END DATE:
Project summary
- Funding mechanism:
-
- Network Innovation Allowance
- Expenditure:
- £690,000
The vast majority of PV generation is connected to the distribution network and currently cannot all be monitored directly by National Grid as the GB System Operator. In the last 7 years, approximately 12 GW of Solar PV systems have been integrated into the GB distribution network.
The ability of the System Operator to forecast the generation from these systems is now a critical component of managing the network. This project will focus on addressing several key challenges identified in monitoring solar PV generation and its operational impacts on the GB electricity transmission system.
This project will build on methods and approaches for estimating PV generation and data provision to develop, test and validate the capabilities of a proof of concept system which would be suitable for integration and use within the System Operator’s forecasting and control room operations.
Objectives
The Solar PV Monitoring Phase 3 innovation project aims to investigate novel methods for monitoring PV generation on the GB electricity system to support control room operations and decision-making.
The project will run over a period of three years with the following key objectives targeted through delivery of its four work packages:
- Develop and validate a prototype 5-minutely national (GB scale) PV generation forecast
- Develop and validate a prototype 5-minutely regional (GSP scale) PV generation forecast
- Identify relevant and suitable national PV data sources/feeds to use in defining optimum approaches to extract PV capacity estimate directly from such PV feeds.
- Investigate suitable methods to estimate PV capacity directly from a combination of PV_Live, demand outturns and meteorological observations instead of the current method which relies primarily on PV installation registration data. Further refine existing methods of capacity determination.
- Evaluate the resilience of PV data feeds used and the sensitivity of the accuracy of the PV_30-minutely and PV_5-minutely monitoring prototypes to geographic distribution, mix of system sizes and mix of rooftop or ground-mount.
- Devise and test suitable validation methodologies for validating the nationally and regional aggregated PV outturn.
pdf
pdf
pdf
pdf
docx
pdf
pdf
pdf
pdf
docx
Learnings
Outcomes
There are several outputs that have been published as a result of the project. They are listed here:
On GitHub
https://github.com/SheffieldSolar/PV-Deployment-Tracker
Private repository of Python code to compile a complete list of deployed PV systems in GB. Awaiting sign off to be made public.
https://github.com/SheffieldSolar/Geocode
Public Python library Geocode various geographical entities including postcodes and LLSOAs, and to reverse-geocode to LLSOA or GSP/GNode (interfacing directly with the ESO Data Portal).
https://github.com/SheffieldSolar/GSP-Region-Mapping
Private repository of Python code to generate GIS boundaries associated with GSPs using data from DNOs. Under active development and awaiting sign off to be made public.
https://github.com/SheffieldSolar/SS_NG
Private repository containing Python code for the PV_Live outturn models. Work is underway to generalise the code such that it can be made public without exposing sensitive information.
https://github.com/SheffieldSolar/GDPR-Location-Anonymiser
Public repository containing a Python implementation of a location anonymization algorithm that is useful when receiving or sharing datasets containing GDPR-sensitive location data.
https://github.com/SheffieldSolar/sp2ts
Public repository containing a Python library to reliably convert between settlement periods and Unix timestamps.
https://github.com/SheffieldSolar/MySQL-DBConnector
Public repository containing a mature Python library for interfacing with a MySQL database with additional resilience and redundancy. Developed to increase resilience of PV_Live service.
https://github.com/SheffieldSolar/Carbon-Intensity-API
Public repository containing a Python library to conveniently access data from ESO’s carbon intensity API.
https://github.com/SheffieldSolar/OSM-PV
Public repository containing validation code developed as part of a collaboration with The Turing Institute and Open Climate Fix to leverage the Open Street Map community, Machine Learning and Machine Vision to exhaustively map GB PV systems.
https://github.com/SheffieldSolar/PV-Forecast-Validation
A public repository containing a Python library to produce standardised reports of PV forecast accuracy by comparing with historical PV_Live data.
On Sheffield Solar website
https://www.solar.sheffield.ac.uk/pvlive/
https://www.solar.sheffield.ac.uk/pvlive/regional/
https://www.solar.sheffield.ac.uk/pvlive/api/
Near-real-time and historical half-hourly PV outturns for GB, GSP, Gnode and DNO License Area continue to be published graphically on the Sheffield Solar website and via API.
On ESO Data Portal
https://data.nationalgrideso.com/system/gis-boundaries-for-gb-dno-license-areas
Geospatial boundaries for the 14 DNO License Areas of the GB transmission network. Useful for aggregating DER energy flows and for visualising regionally aggregated DER.
https://data.nationalgrideso.com/system/gis-boundaries-for-gb-grid-supply-points
Geospatial boundaries for GSPs/Gnodes on the GB transmission network. Useful for aggregating DER energy flows and for visualising regionally aggregated DER.
Two academic publications relating to the errors and uncertainties involved in PV Live:
Huxley et al. June 2021. The Uncertainties in measuring national photovoltaic electricity generation. In preparation for submission to Renewable and Sustainable Energy Reviews. Pre-print available on request.
Huxley. The Uncertainty in measuring national level PV electricity generation. The case study of Great Britain. Thesis being prepared for submission to The University of Sheffield for confirmation of degree of PhD in July 2021.
Lessons Learnt
We have learned of the high value of adopting incremental improvements and releases throughout the project both in terms of developing our community of users but also in order to learn of benefits and problems quickly.
Recommendations:
Implement a second source of real time PV sample data to increase the resilience and accuracy of the service.
Include an algorithm to correct the bias error in intraday out turn when compared with day+1 out turn
Include the MCS database as a primary source of capacity data alongside, and in future instead of, the FIT register.
Complete the Electralink validation work.
Lobby for ENA system wide resource register to include systems significantly less than the current 1 MW capacity.
Work with BEIS, Solar UK and Solar Media to consolidate learning on PV capacity registers and converge on a single national capacity tracking methodology, service and capacity value.