Physically Based Methodology for Constructing Load Models for Power Systems. Final Report

摘要

The properties of aggregate loads and the estimation of aggregate load properties from realistic observational data have been emphasized. Most of the results deal with loads composed of many smaller loads with steady-state response characteristics. All of the elements of information needed to construct a steady-state load model have been identified in general terms, and much of the methodology has been developed. Network aggregation (Chapter 2) enables stochastic models of loads at one level of the system to be ''lifted'' to the next higher level. It may be used to aggregate from the customer to the distribution level, from distribution to subtransmission, or from subtransmission to the bulk power level. Chapter 3 treats the difficult problem of estimating end-use (device-level) demand by disaggregation when it cannot be measured directly, and measurements must be made at the customer level. The probabilistic argument applied to a large aggregate of steady-state loads in Chapter 2 is extended to an aggregate of linear dynamic loads in Chapter 4 for the special case in which the loads are connected radially to a single bus. The modeling of composite dynamic loads presents many difficulties not encountered in the steady-state case, and the results presented in Chapter 4 are less comprehensive. Chapter 5 deals with device-level load modeling. The work reported in this chapter led to the decision to emphasize the modeling of demand by disaggregation from the customer level. (ERA citation 11:009037)