Flexiramp market design for real-time operations: Can it approach the stochastic optimization ideal?

Some ISOs are creating short-run markets for “flexiramp” products to ensure that sufficient flexible capacity is on-line to manage increased volatility of net loads expected under high levels of renewable integration. We evaluate some basic features of flexiramp market designs by comparing the solutions of (1) a deterministic dispatch model with a flexible ramping constraint with (2) a stochastic dispatch model. Our fundamental assumption is that the purpose of the flexiramp product is to improve the expected performance, in terms of costs, prices, and reliability, of the deterministic market models used by ISOs, with a stochastic model being ideal. A highly simplified market is considered so that fundamental reasons for successes (and failures) of flexiramp markets can be readily understood and replicated. The deterministic and stochastic models are compared considering the dispatch, prices, settlements, and market efficiency resulting from the actual realizations of net load. Our preliminary results illustrate the contribution of the flexiramp product to market efficiency if the optimal amount of flexiramp is secured, but they also clearly show that flexiramp constraints are insufficient to produce the least expected cost solution. Accordingly, we recommend careful design of the ramp product market and, ultimately, transition to the use of full stochastic dispatch and commitment models.