In this series, we explore some of the important concepts that are used in deciding how planning investment decisions are made and the growing role that demand side management will play into the future.
Accurate network planning is the foundation on which good investment decisions in T&D infrastructure are made. Planning engineers must help network managers allocate their scarce resources and this has very real consequences for end customers. Network planners must help their businesses deliver on network reliability targets while meeting the financial goals of their shareholders. But how should the trade-off between network reliability and economic drivers be balanced? How can networks be augmented in the most cost effective manner?
There are of course many approaches available to network planners and this post will outline and compare the two most common planning standards: Deterministic and Probabilistic planning, as well as exploring some of the alternatives.
Deterministic planning (N-x)
Deterministic planning requires that the network continues to provide an adequate and secure supply of energy to customers after any of a range of contingencies occur (with no probability of a contingency occurring taken into account). Example: for a given constraint it is assumed that the largest network asset (e.g. transformer) is out of service (i.e. the nominal capacity of the constraint minus one contingency: N-1). This results in a network design where if any one asset fails there is no loss of supply.
A simplistic view of an N-1 deterministic planning approach can be applied to a substation which has two 30MVA transformers resulting in a total capacity – N rating – of 60MVA. If one transformer were to fail the substation would have a remaining capacity of 30MVA, its N-1 rating. With deterministic planning if peak load growth forecasts for the substation were expected to reach 31MVA next year, even for just one hour, an augmentation would be required to increase the capacity of the substation and maintain network security to the N-1 deterministic standard.
Deterministic planning approaches are both recognised and applied internationally and they have an advantage for regulators in that the approach is easy to regulate, which this results in quick planning solutions. However, deterministic planning suffers from a number of shortcomings. Deterministic planning delivers a higher level of network redundancy, typically well above the needs of most end customers, and does not consider any unique conditions within a region or the typically low probability of an outage.
Probabilistic standards require the electricity network to provide adequate and secure supplies of energy to customers under a wide range of contingencies each treated as a random event and the probability of each contingency occurring (e.g. transformer failure rates) is taken into account.
Probabilistic standards use economic cost-benefit techniques to determine the economic viability of a proposed augmentation. A network planner assesses the probability that events are likely to cause constraints and load shedding on the network during the planning horizon and then calculates the impact of those load shedding events. The amount of energy which could be lost if a contingency occurs is called the ‘energy at risk’ (EaR). When the EaR is multiplied by the probability of a contingency occurring we have the ‘expected unserved energy (EUSE). This framework provides flexibility for network planners as reliability targets are met through an explicit valuation of customers willingness to pay (WTP) for uninterrupted electricity supply. In Australia this is known as the Value of Customer Reliability (VCR).
What is WTP and VCR?
Willingness to pay (WTP) and the value of customer reliability (VCR) are the same thing: a representation, in dollars per kilowatt hour (kWh), the willingness of customers to pay for the reliable supply of electricity. The WTP/VCR assists electricity planners, asset owners, and regulators to strike a balance between delivering secure and reliable electricity supplies and maintaining reasonable costs for customers.
Probabilistic planning in practise
Using the 60MVA substation in the example above; we can assess the constraint from a probabilistic perspective.
Firstly we have to lay out our assumptions namely:
• The preferred network option to augment the substation costs $5 million
• The WTP/ VCR is $100 per kWh
• The probability of a transformer failing is 0.5%
We can now use these values in line with the load forecasts to calculate that the economic timing of the augmentation occurs in Year 5.
A probabilistic approach enables network development with an optimal level of reliability, rather than a level of pre-determined level of redundancy, and assesses the economic impact of low probability but high impact events. It achieves optimal level of system reliability, security and congestion. While it can suffer from being lengthy, given the number of inputs required, any unforeseen events can be taken into account through appropriate safeguards.
A little bit from column A and a little bit from column B…
There are many different planning approaches which are used in different jurisdictions. Two alternatives utilised in Australia are the Hybrid Planning and Energy Cap approaches.
South Australia uses a probabilistic standard (using economic considerations) expressed as a deterministic standard. The VCR is used to inform the deterministic standard and is taken into account in the reliability standards for longer-term planning requirements. It requires assumptions to be made many years in advance of the likely augmentations to address an emerging constraint.
Load connection points are allocated into one of six reliability categories which are designed to capture the idea that as demand at a connection point increases over time, so does the economic cost of losing the connection point’s supply.
The probabilistic approach is used to compare the cost of increasing reliability standards of a connection point to the next deterministic reliability level with the value of the increased reliability delivered to the connection point.
This approach delivers better outcomes than those of a strict N-X approach because it identifies different levels of reliability according to demand for each connection point and considers a VCR when determining the level of reliability required at a particular connection point. It also allows an independent body to establish and audit the arrangements.
It is difficult to apply in regions with a large number of connection points or a more meshed network where there is additional complexity to determine the base levels of security.
Further, although the VCR is used for level of reliability of a connection point it is not used when considering augmentation options. The result is that the price-service balance is not optimal.
An energy cap approach is applied in Tasmania. The approach limits or caps the size of customer load, expressed as energy or peak demand that can be lost. The planning criteria requires:
• a credible contingency event will not interrupt more than 25 MW of load,
• a single asset failure (e.g. a double circuit transmission line or substation busbar) will not interrupt more than 850 MW or cause a system black-out
• the unserved energy to loads interrupted as a result of a credible contingency event must not exceed 300 MWh
• the unserved energy to loads interrupted as a result of a single asset failure or the failure of equipment to perform as intended following a credible contingency event, must not exceed 3000 MWh.
Similar to the hybrid approach, this approach delivers better outcomes than a strict N-X approach by implicitly recognising that it may not be economic to augment at the time that the N-X standard is breached. However the analysis the analysis still fails to recognise that the cost of some of the solutions may be inappropriate given the level of energy at risk. .
So, how do these planning approaches stack up?
A comparison between the four planning approaches is set out below. It highlights that the probabilistic planning approach delivers system security and performance obligations of a network in the most economical manner.
There are considerable benefits to both end consumers and network operators created by probabilistic planning. The key to creating the best economic outcome with probabilistic planning is to understand the subtleties in the benefits calculation and while this is often well understood for traditional network builds it is often poorly understood for demand side solutions. The ability to understand these subtleties is a key differentiator between networks and regulators who make smart investment decisions and those which don’t.
This blog is part of a series titled “PortfolioCM: Network Planning and Design”
Planning teams within utilities have decades of experience in network augmentation design, with well trodden tools and processes to follow to arrive at a technically sound, economic design. For the modern utility, modelling the costs and benefits of a DSM programs, storage solutions, embedded generation assets and renewables will become a day-to-day activity for planners and the GreenSync Network Planning and Design blog series will explore the key concepts in network planning for modern networks.