Curtailment and Dispatch Estimation Toolkit
|Funding mechanism||Network Innovation Allowance (NIA)|
|Duration||Oct 2018 - Oct 2019|
|Research area||Network Improvements and System Operability|
The main objective of the project will be the development of customer behaviour models for all types of demand, generation and storage that can be used as an input to the Energy Curtailment (and/or dispatch) Estimation techniques that WPD is developing.
The development of distributed energy resources termed DER (such as wind, solar photovoltaic, hydro, landfill gas, CHP etc) will displace existing conventional methods of generation which historically have provided the necessary response characteristics to maintain the overall integrity of the network. The output of the DER is both variable and uncertain because of its intermittency. On the demand side too, uncertainties are growing due to changes in consumption patterns (e.g. with electric vehicles etc). This scenario is thus transitioning to a less predictable and more stochastic world.
The most common form of analysis undertaken by any DNO for network planning is a load-flow study taking into account the maximum coincident load. With more DER (particularly wind and solar photovoltaic generation), the snapshot of the maximum coincident generation, will also become relevant for network planning. Currently carrying out a multitude of modelling analysis using half-hourly flows at each network point, for different types of DER and demand technology presents WPD with a scaling problem – doing this for an historical year is manageable, but trying to do this for forward planning purposes, using a further number of future energy scenarios would be impracticable using the current software. It is therefore imperative that new modelling capabilities are required to address these challenges.
It has been recognized that WPD will need to develop “forecasting future energy volumes across the network (under different scenarios) to highlight opportunities for flexibility, operability issues and to identify when strategic reinforcement will be needed”(see bullet point 2 on page 10 of the DSOF document) in order to facilitate its transition from DNO to DSO. The purpose of this project is to develop customer behaviour models for all types of demand, generation and storage that can be used as an input to the Energy Curtailment (and/or dispatch) Estimation techniques that WPD is developing. There are about 17,520 half-hours in a year; preferably this project will identify sufficient commonality that needs to be assessed to arrive at MWh figures for a year to perhaps several hundred. Specifically, this project will provide a more accurate visibility of future curtailment and potential for flexibility required under modelled scenarios.
The work will be carried out in three phases covering the four WPD’s licence areas, comparing two innovative methods. Methods Alpha and Beta have different first phases, but share the replication phases 2 and 3.
Method Alpha Phase 1: Analysis of half-hourly data for a summer for the South West licence area will be selected for this analysis. The detailed analysis will consist of (a) development of the Stochastic Load Flow (SLF) Algorithm (b) assessment of load profile/correlations with meteorological data (c) assessment of hourly, daily, weekly and monthly groupings.
Method Beta Phase 1: Understand the three main independent variables affecting load profiles; weather, demand growth and generation growth across grouped climatic regions.
Perform a study based on the “Monte Carlo” method – that of repeatedly analysing random inputs (conditions) to establish the distribution of outputs (loadings).
Phase 2: Expansion of Phase 1 techniques across a yearly profile (a) development of the WPD generation/demand clusters (b) for each cluster, repeating the studies described in phase 1, for a set location (c) development of other technology types excluded or minimised under phase 1.
Phase 3: Generalisation of techniques to other locations (a) application of techniques to other licence areas and locations (b) iteration to grouping techniques based on data from other locations.