Experimental and Practical Experience With the Gibson Method of Water Measurement: A Discussion of Prof. D. Thoma’s Paper on “The Degree of Accuracy of the Gibson Method of Water Measurement”

Abstract

Abstract The Gibson method is based on the equation of impulse and momentum applied to an enclosed column of water in motion. It is applicable in testing hydraulic power plants where the turbine is supplied with water through a closed conduit and means, such as turbine gates, are available for interrupting the flow of the water. To apply the method, it is necessary to obtain pressure-time diagrams, which show the changes with respect to time that occur in the conduit during and after the closure of the turbine gates. There are two kinds of diagrams: (a) simple diagrams, in which the changes of pressure at one point in the conduit are recorded, and (b) differential diagrams, in which the difference between the changes of pressure at two points in the conduit are recorded. Professor Thoma (1) has discussed quite thoroughly, in an accompanying paper (Trans. A.S.M.E., 1935, HYD-57-4), the conditions of water flow which may affect the degree of accuracy of the Gibson method of water measurement, when using the simple application method. In general, he has considered four errors due to (No. 1) accessory motions, or extra currents due to turbulence, (No. 2) false valuation of the friction in the conduit during closure, (No. 3) friction of the mercury column, and (No. 4) the inertia of the mercury column. In discussing the Thoma article, the authors have first given a somewhat more detailed proof of the fundamental equation applying to the method and have supplemented the derivation with a graphical representation of the terms of the equation on the diagram itself. Test results on a long pipe are shown where both kinds of diagrams were taken and where the agreement between the two sets of diagrams is within 0.2 per cent. The test results are discussed and a detailed study is made of the effects of the four errors. The authors conclude that the residual error is probably within the limits of precision possible when large quantities of water are being measured.