Planning for dispatch efficiency
Using health checks on dispatch practices to ensure efficiency, reduce costs and ramp up renewable energy use
Background, challenges, and context
Efficiency of dispatch in power systems is an important health check to ensure that existing resources are being used properly and low carbon resources can be accommodated going forward.
The World Bank’s Energy Sector Management Assistance Program (ESMAP) has found that even in a relatively small system such as in Bangladesh (11 GW at the time of the analysis), over a billion dollars could be saved through more efficient dispatch by reducing unnecessary reliance on oil and using available domestic gas more efficiently.
In countries where wholesale electricity markets are absent, it is critical that health checks are performed on dispatch practices not only to reduce costs, but also to explore ways to introduce much-needed flexibility into systems so that variable renewable energy can ramp up rapidly without jeopardising system reliability. Yet, dispatch diagnosis is very rarely performed in developing countries.
Research overview and objectives
The Power Systems Planning Group, part of ESMAP, conducted a dispatch optimisation study in Pakistan and Nigeria, replicating a previous study it carried out in Bangladesh.
The team established a simple but robust methodology that could work with existing modelling tools and available data and could be applied to a number of systems.
The research questions included:
How can we set an efficient benchmark, namely an optimised generation dispatch?
Are existing generators being used optimally to meet demand?
Is the system ready to accommodate a significant level of renewable entry?
What explains the divergence from an efficient benchmark?
What needs to change to ensure improved utilisation of existing generators? What additional measures need to be put in place to increase penetration of solar and wind?
An efficient benchmark was established by developing a dispatch optimisation model that dispatched all generating units in the power system on an hourly basis for one or more years, and reflected the constraints the system faced. This optimised dispatch was then compared with the actual dispatch. The objective was to understand the deviations of the actual dispatch from the efficient benchmark model and to identify plausible reasons for these deviations.
To assess renewable readiness, the team also tested the adequacy of spinning reserve* for each hour (the availability of spinning reserve is critical to accommodate variability in solar and wind resources).
The study was conducted in close cooperation with power system operators, which provided data on demand, actual dispatch, transmission, etc, and with the central utilities in charge of Power Purchase Agreements (PPAs).
Research results, key messages, and recommendations
The main deviations between actual and modelled dispatch were found to be:
In the actual dispatch the (costlier) combustion engines and steam turbines have a generation share far beyond their optimised generation in the modelled dispatch.
Instead, generation from combined cycle gas turbines (CCGTs) and coal-fired power plants is higher in the modelled dispatch than in the actual dispatch.
Modelled dispatch costs are 17.8% lower than the actual (estimated) cost.
The model shows more dispatch in the South, balanced by less dispatch in the Midlands (approximately the Punjab region).
As a result of the above, the carbon dioxide emissions are lower in the modelled dispatch despite a higher share of coal-based generation.
It appears likely that the following drivers are responsible for deviations between the actual and modelled dispatch:
It is likely that fewer economical plants (combustion engines and steam turbines) are dispatched due to capacity constraints or voltage regulation issues, which require the injection of reactive power in a certain grid area to keep the voltage in its admissible bandwidth. This cause for inefficient dispatch is beyond the control of the dispatcher and can only be eliminated with reinforcement measures for the transmission system.
Lack of automatic generation control and fully working Supervisory Control and Data Acquisition (SCADA) necessitate a manual optimisation, which is not as efficient as a computer-based generation control, e.g. respecting minimum up and down times and optimum and full reserve provisioning. Such equipment would allow for timely improvement of the dispatch efficiency. Further, the manual dispatch is likely influenced by established historic dispatching schedules and standards, which may not be the most efficient.
Another effect could derive from the considerable increase of total capacity just before the analysis period. The many large new power plants (regasified liquified natural gas, CCGTs, and coal power plants) may not be fully introduced into daily dispatch routines as some challenges of commissioning may persist (e.g. increased generation in the South requires higher transfer to the Midlands, fuel contract consideration is manual, and newly commissioned plants may still need some testing and improvements). As new capacity is committed and under construction mainly in the South (nuclear, local coal, and wind farms), any such issue may increase in the future and should be addressed with proper planning and coordination.
Another potential inefficiency that is not shown in the dispatch data refers to considerable curtailment of wind generation due to transmission congestion derived from various generation sources.
A set of prioritised recommended actions and measures has been prepared, aiming to:
Use resources efficiently.
Eliminate dispatch inefficiencies.
Provide the required level of flexibility (operating and contingency reserve) for enabling the integration of large-scale variable renewable energy, existing and committed capacities, and planned further expansion.
The team’s dispatch analysis model showed that:
Merit-order based dispatch within the current constraints can yield an annual benefit of $29 million or 4% of variable costs.
Short-term efforts to improve gas supply and infrastructure yield the largest benefits, especially if they are combined with improved plant-level availability. If the existing gas supply bottlenecks are moved and average plant availability can be improved from the current level of 27% to 50%, $115 million could be saved.
With the removal of physical availability constraints around plants and gas supply, benefits increase by an order of magnitude to $309 million, or 19% of total annual costs.
Even larger savings could be achieved if commercial contracts could be streamlined to reduce reliance on generators that have very large fixed charges and take or pay (ToP) obligations. If uneconomic ToP obligations could be waived, benefits would further increase to $579 million or over 30% of total annual costs.
Removal of operational and commercial constraints could also allow the Nigerian power system to meet increased demand to the benefit of unserved residents or businesses (the existing gas generation fleet could easily meet 25% more demand) and pave the way towards increased penetration of variable renewable energy resources. An addition of 500 MW solar capacity in the current system would earn reasonable return on investment between 4% and 9% and help to meet increased demand. However, an important precondition for adding large-scale solar is to raise the spinning reserve capability in the system through additional investments in flexible resources, namely open cycle gas turbines, pumped storage hydro, or battery storage systems.
The project did not estimate the cost of implementing these measures, but very little hard investments are needed for some of them (such as establishing a strict merit-order based dispatch or enhancing operational practices to improve plant and gas availability). When it comes to renegotiating take or pay contracts however, there will be difficult commercial and institutional issues – albeit the huge financial savings should make it a worthwhile undertaking.
Improving the utilisation of existing generation assets and making operational and commercial improvements would bring significant benefits to the Nigerian power system. These are major findings, especially if considered in the context of the tremendous financial stress in the power sector.
However, the team highlights that parallel measures are needed to improve the financial position of the sector. A comprehensive reform programme needs to include gradual implementation of cost-reflective tariffs together with improved collection efficiency and installation of (smart) meters to reduce non-technical losses.
Data, tools, analysis, and training has been made available to relevant groups, with a key objective of the project being to ensure system operators and utilities can perform health checks routinely and apply the findings to ensure more efficient dispatch going forward.
*Spinning reserve in this case refers to the sum of the currently unused net generating capability of all power supplies currently connected to the grid that are synchronised to the power distribution system and which are capable of immediately accepting additional load in the event of, for example, the failure of a fossil fuel generator plant or short-term variability in renewable generation.