Predicting AC power consumption using a stochastic Markov chain simulating weather conditions

In most electricity systems the residential sector is one of the main contributors to system peaks. Hot and humid summer seasons cause a significant proportion of the supplied power to be used on air-conditioning. Indeed, outdoor weather conditions are crucial in determining the level of residential energy consumption for heating, ventilation and air-conditioning (HVAC) household appliances. In this paper, we address the problem of determining the expected power demand under projected weather conditions. Stochastic Markov process is used to simulate future humidity and temperature levels during each month of the hot season. The probability transition matrix of the stochastic process is constructed from field measurements of outdoor humidity and temperature. Simulations show that a good prediction of the average power demand is achievable.