Artificial representation of power systems

Abstract

Synopsis.-The size and complexity of present day power systems have increased to the point where the prediction of the behavior of the system by analytical methods is more and more difficult. The solution of commercial networks by Kirchhoff s Laws or by cut and try methods, even with the help of star-delta trans formations, leads to such involved equations that the need for simpler methods is keenly felt. Increasing attention has been given to various methods of representing power systems in miniature so that an experimental solution may be substituted for an algebraic one. The d-c. short circuit calculating table is a satisfactory and relatively simple means of determining short circuit currents in net works, but is too inaccurate to give satisfactory solution under normal operating conditions. An a-c. artificial representation of power networks in miniature has been developed by 0. R. Schurig of the General Electric Company who used 3.75-kw. 110-volt three-phase generators as power stations. Actual transformers are used, while lines and loads are made up of conveniently arranged lumped units of inductance, capacity, and resistance. This apparatus has been in satisfactory operation for several years. Evans and Bergvall of the Westinghouse Electric and Manu facturing Company used a test floor setup to check experimentally the theory of long line stability. Powers of about 500 kv-a. were used. The present paper presents a method of artificial representation on a laboratory scale, decreasing the size of the apparatus and in creasing the precision of the results. All rotating apparatus has been eliminated. Generators are represented by phase shifting transformers; transformers by their equivalent circuits, and lines by lumped constants. A description of the apparatus used by the writers is presented, together with the results which were obtained by its use in the solution of several typical problems. An analytical check on one of the examples is given, showing a precision of better than 1 per cent. PURPOSE AND SUMMARYT HE purpose of this paper is to describe briefly the work done by the authors under the direction of Dr. V. Bush, in the Research Laboratory of the Masschusetts Institute of Technology in designing, building, and testing apparatus for setting up miniature networks, using generating station powers of 100 watts or less. Complete voltage, current, and power solutions were made on several arbitrary networks. The repre sentation was single-phase, using phase-shifting trans formers for generating stations, and a vacuum tube voltmeter drawing absolutely negligible current for measuring potentials.

Artificial Neural Network for Short-Term Load Forecasting in Distribution Systems

“…For example, Spencer and Hazen [1] present in 1925 an artificial system to simulate and study Kirchhoff's laws with different loads, and with alternating and direct current.…”

Perpetually Laborious: Computing Electric Power Transmission Before The Electronic Computer

“…After introducing their first calculating board in 1916 as a solution to computing ''complexity'', 42 and after justifying a series of modifications as solutions to the fact that ''the mathematical methods were very laborious and often impossible of application'', 43 they and their MIT partners would, by 1925, find that ''the size and complexity of present-day power systems have increased to the point where the prediction of the behavior of the system by analytical methods is more and more difficult'' and the calculating table is ''too inaccurate''. 44 By 1930, they were prepared to introduce their own alternating current calculating board -baptized the ''network analyzer'' and ceremoniously presented at MIT to emphasize how revolutionary a machine it was. It too was offered as the solution to ''lengthy and usually impracticable mathematical calculations''.…”