L'optimisation du réglage de vitesse des groupes hydroélectriques

Optimized speed governing for hydro-electric power units

¹ Electricité de France - Direction des Etudes et Recherches.

Résumé

The complex difficulties involved in choosing a characteristic speed-governing "quality criterion" are briefly reviewed and the cases are considered for which statistical information on network load fluctuations is available. General principles of the determination methods used. The speed variation range of a unit gives a good indication of "governing quality" for practical purposes. The criterion used in the determination is directly related to this, the aim there being to reduce the standard deviation of speed fluctuations due to random network load fluctuations to a minimum. The principle of Wiener's optimization method is reviewed in this connection. Application to the speed governing of a hydroelectric power unit. The pattern of electricity network power fluctuations found experimentally and theoretically is defined, also the transfer functions for the governed network, for which relevant notations and basic equations are given in the Appendix. Figure :1 shows the general system layout and Figure 4 refers directly to the application of Wiener's method. Results obtained with Wiener's method. The principal closed-loop transfer functions for the optimum system are given by formulae (11) and (12), and formulae (13) gives the transfer function form the governor requires so that this opotimum system may result. The minimum standard deviation for speed variations is then found by formulae (14). Figure 5 gives this standard deviation in terms of acceleration constant for two special cases. Optimum governor design. The possibilities and limitations are compared of three different special govern or systems obtained with a accelero-tachometric governor equipped to detect the derivative of power unit acceleration, electric power and net head respectively, with suitable signal-connecting filters in the last two cases. Optimum regulation corresponding to a system stability limit cannot be achieved with the first and third methods, but can with the second in certain cases. Application of the optimization method to accelero-tachometric governing. As a definite transfer function is specified for the governor in this case, the method can be used to determine optimum regulation. Figure 10 gives the values of a parameter m0 in terms of two values u and v depending directly on the characteristics of the system. The standard deviation for speed variations σn, the promptitude time constant Tn and the derivative time constant of the speed sensor Tn are found from m0 by simple formulae. Differences in performance between this adapted form of accelero-tachometric governing and ideal optimum governing are shown for individual applications in Figure 11.