A Risk Assessment Method for Ramping Capability Shortage Caused by Generator Failure and Net Load Forecast Error

Abstract

Increasing variable generation (such as wind power and photovoltaics) in power systems has required the system operators to secure more ramping capability (RC) to satisfy variations in net load. The more RC that is secured, the better the system is able to deal with the variations in net load. However, uncertainties inherent in generation resources (i.e., failure) as well as in net load forecasts make it harder to secure the RC. These uncertainties may trigger RC shortage, which can result in a reliability crisis. When considering the impact of the uncertainties, there exists no clear RC requirement. To determine how much RC is sufficient, the trade-off relationship between the risk of RC shortage and the operating costs of providing RC should be first taken into account. However, unlike the operating cost, the risk of RC shortage is not easily quantifiable. Although critical uncertainties such as failure of generators and NLFE have been mentioned as the cause of the risk of RC shortage, their causal relationship has not been explicitly addressed in previous researches on the risk of RC shortage. This dissertation proposes a risk assessment method for ramping capability (RC) shortage. Two major uncertainties in power systems are considered: generator failure and net load forecast error (NLFE). The failure probability of generators is calculated using a Markov-chain-based capacity state model, where two types of generator were considered depending on its initial state: one is a two-state model for generators that are initially committed, and another is a four-state model for generators that are not initially committed. The failure probability is calculated based on a simple matrix multiplication technique. Meanwhile, the NLFE is modeled as a normal distribution, which is represented using a seven-step approximation. The risk of RC shortage is evaluated using an index termed RC shortage expectation (RSE), which is defined as the sum of the probabilities that the RC requirement will be not satisfied by the system. Numerical examples were then presented to describe the RSE calculation procedure with and without NLFE and the generation schedules. A case study was carried out using a modified IEEE-RTS-96 to show the applicability of the method. The sensitivity analysis was also performed to find the relationship between the RSE and influencing parameters. The failure rate of the generators, the spinning reserve requirement, the installed capacity of wind farms were chosen as the parameters to represent the uncertainty, the operating scheme of the system operator, and the variability due to wind power, respectively. The reliance of the RSE on these parameters can enable the system operator to obtain useful information. If this method is combined with a well-developed cost estimation for generation schedules, then it is expected to be particularly useful for the system operator to determine the adequate RC level for a given power system.