Operation Risk Constraint Resolution. Lagrangian Method For Unit

Spinning reserve (SR) is needed in the system to cover for unforeseen events such as sudden increase in demand and loss of generators/lines. SR allocation often has important bearing on the dispatch and unit commitment decision, because it comes at some cost. This can be achieved by adjusting the SR on various generating units to keep the total start-up and operating cost impacts at the minimum. As competition intensifies, this cost is bound to come under scrutiny and a fundamental question will need to be revisited: what is the optimal amount of spinning reserve that should be scheduled?Spinning reserve requirement is set in most traditional UC models using various deterministic criteria, because they are easily understood and implemented, but they do not match the stochastic nature. A probabilistic reserve criterion represents the complete system outage probability distribution and enables dispatch of reserve to meet an acceptable risk level (e.g., maximum Loss of Load Probability-LOLP). This thesis analyzes and discusses the method of dealing with spinning reserve constraints based on probabilistic analysis.However, the relation between unit commitment risk and forced outage capacity is a discrete distribution, the Lagrangian Relaxation unit commitment algorithm isn't used directly. So this thesis analyzes two kinds of curve fitting from the unit commitment risk, which are proposed Gauss function and previous exponential function. It shows that Gauss function is better than exponential function. And this thesis modifies this reserve constraints which can be introduced into models of unit commitment. The cases of 26-units,110-units have been studied and the test results are satisfying . This method can organically handle the unit commitment risk constraint under the optimal conception and realize the real optimization.