This paper addresses the problem of 'quickest possible' online transient stability assessment, by minimizing the decision time of combined event detection that might lead to a system split and unstable generator group prediction, from real-time wide area power system measurements. More importantly it does so by respecting predefined probabilistic error constraints for the prediction. The statistical theory of optimal detection is applied, firstly to choose the detection threshold and secondly to select a flexible assessment time, after using probabilistic neural networks to provide a temporal representation of the data. On simulated wide area measurements from the interconnected New England test system and New York power system this approach is between two and three times faster on average than strategies based on fixed assessment times, despite having comparable error rates.

We find empirical evidence that mean-reverting jump processes are not statistically adequate to model electricity spot price spikes but independent, signed sums of such processes are statistically adequate. Further we demonstrate a change in the composition of these sums after a major economic event. This is achieved by developing a Markov Chain Monte Carlo (MCMC) procedure for Bayesian model calibration and a Bayesian assessment of model adequacy (posterior predictive checking). In particular we determine the number of signed mean-reverting jump components required in the APXUK and EEX markets, in time periods both before and after the recent global financial crises. Statistically, consistent structural changes occur across both markets, with a reduction of the intensity and size, or the disappearance, of positive price spikes in the later period. All code and data are provided to enable replication of results.

This paper proposes a novel methodology for characterization of nonlinear low frequency oscillations. Selecting a window of the speed machine signals, a polynomial in the z-domain of the samples is obtained, and then Pade approximation is used to obtain a rational equivalent of the polynomial. This equivalent has a direct relationship with a discrete transfer function which captures the dominant parameters embedded in the speed machine signals. The method is tested with corrupted noisy signals showing that it may be applied to real signals. The developed methodology is applicable to determine the oscillation parameters of speed signals of the 16-Machine 68-bus NPCC power system.

We address the problem of predicting the transient stability status of a power system as quickly as possible in real time subject to probabilistic risk constraints. The goal is to minimise the average time taken after a fault to make the prediction, and the method is based on ideas from statistical sequential analysis. The proposed approach combines probabilistic neural networks with dynamic programming. Simulation results show an approximately three-fold increase in prediction speed when compared to the use of pre-committed (fixed) prediction times.

We propose a new practical methodology for risk-sensitive stochastic optimal control, showing that material decreases in risk indices are possible with relatively little loss of average case performance. Cost optimisation of energy system assets has typically been carried out under the assumption of risk-neutrality, minimising average operational costs. The risk profile of control strategies is thereby ignored, despite the fact that liberalised electricity markets can be highly volatile and financial risk is a material consideration. In a flexible energy system with cogeneration and heat storage, however, it is possible to exploit variation in wholesale price level or volatility (or both) by shifting heat demand through time and varying electricity demand, achieving a balance between low average cost and low volatility which takes account of risk preferences. Based on least squares Monte Carlo regression, our proposed method optimises an exponential objective function containing a risk sensitivity parameter which may then be tuned to achieve the desired tradeoff. We provide a realistic case study of a flexible district energy system, where local heat and electricity demand must be satisfied at minimum cost subject to stochastic price dynamics and the physical constraints of the system. In this example we compare risk-neutral and risk-sensitive optimal strategies and show consistent changes in economic risk under two different risk measures.

This paper proposes a novel methodology for characterization of nonlinear low frequency oscillations. Selecting a window of the speed machine signals, a polynomial in the z-domain of the samples is obtained, and then Pade approximation is used to obtain a rational equivalent of the polynomial. This equivalent has a direct relationship with a discrete transfer function which captures the dominant parameters embedded in the speed machine signals. The method is tested with corrupted noisy signals showing that it may be applied to real signals. The developed methodology is applicable to determine the oscillation parameters of speed signals of the 16-Machine 68-bus NPCC power system.