Identification and validation of dynamic global load model parameters for use in power system frequency simulations

In small to medium sized isolated power systems, the provision of emergency reserve to arrest the fall in frequency following the loss of a generator is imperative in preventing system black out. This reserve largely comes from three sources, generators, load shedding and demand relief. The estimation of this demand relief has proved difficult. This paper reports on a series of tests made on a 2800 MW system to identify parameters for a dynamic load model which could represent this relief. These tests measured load behaviour indirectly by monitoring system generation and frequency under specific operating conditions. Due to the discontinuous nature of the data, direct search methods had to be employed for the identification algorithm. The parameters were-cross validated by use in a system simulation of an actual contingency and the results obtained were encouraging.