Synthesis of multiarea grid power systems
Department of Electrical Engineering
Doctor of Engineering Science
Meyer, Andrew Ulrich
This dissertation presents improved development in the formation of a generalized transmission loss (B)-matrix for a multiarea grid power system. In the procedure, the individual tie powers of each area are replaced by the net interchange, sneak and circulating powers. The latter two variables are directly eliminated in the power reference frame using actual impedances, unlike current methods that require the elimination of sneak and circulating currents, the formation of complex tie current model and the complex tie power model. Consequently, manipulation of large complex current, power and impedance matrices is avoided reducing both computer time and memory requirement. Further, the procedure not only provides a model for predicting individual tie powers, given generator and net interchange powers, but also provides coefficients that reflect the changes in the tie power flows with respect to the changes in generator and net interchange flows.
The dissertation also presents a modified pool lambda dispatch method that could be used on-line for optimal coordination of generating sources in a multiarea grid power system. The classical fuel cost minimization problem is modified with the addition of a constraint equation that forms the basis for the definition of a common pool reference running cost. The solution algorithm is in a closed form rather than iterative and explicitly provides the individual area running costs in terms of the pool reference cost and the desired generation of each area. Thus, individual areas can be dispatched in a multiarea grid power system in the same manner as individual generators are dispatched in a single area.
Finally, a procedural method of selecting and designing an acceptable optimum power system configuration from a group of system alternatives, in terms of a generalized conductance (G)-matrix is presented. Analysis of an arbitrary N area power system by the method presented herein can be very economical, since the dimension (2N-1)X(2N-1) of the (G)-matrix is substantially smaller than the actual network. Once optimal (G)-matrix is identified, the actual network in reference frame one, can be designed by a reverse transformation, reflecting the constraints set by members of the power pool.
njit-etd1979-006 (225 pages ~ 5,167 KB pdf)
Please complete this Feedback Form to inform us about your experience using this website. It will assist us in better serving your information needs in the future. Thank You!
Created March 27, 2009