Toward combining intra-real time dispatch (RTD) and AGC for on-line power balancing

In this paper we propose first a predictive adaptive intra-RTD dispatch comprising: (1) Intra-RTD non-zero mean predictor which utilizes historic data about various uncertainties to predict the most likely non-zero mean imbalances; and, (2) Adaptive controller for non-zero mean intra-RTD imbalances using two-stage robust optimization. The result of applying these methods sequentially at the beginning of each RTD interval is the least-cost robust adaptation of RTD schedules in anticipation of the most likely non-zero mean imbalances. We illustrate how this proposed predictive adaptive intra-RTD dispatch can be implemented for a typical hour. Second, only hard-to-predict near-zero-mean imbalances are compensated using Automatic Generation Control (AGC). The basic conclusion is that a combination of predictive adaptive intra-RTD and recently proposed Enhanced-AGC (E-AGC) should be used for the least cost on-line balancing of non-zero mean and zero mean imbalances, respectively.