Model investigation of the automatic regulator of the IBR-2 reactor

Abstract

The IBR, IBR-30, and IBR-2 periodic-pulse fast reactors operated successively in Dubna from 1960 to 2006. The IBR-2 reactor began operating in 1984 at average power 2 MW and then 1.5 MW. There are still no analogous reactors in the world, and there are only few aperiodic-pulse reactors. Work is now being done to convert IBR-2 into IBR-2M. Therefore, a careful analysis of its regulation system is topical. This is especially important because IBR-2 is more sensitive to the activity changes than stationary reactors.The automatic regulator of an IBR-2 reactor was moved with an electric stepping motor, as result of which reactivity changed abruptly (continually operating motors are used in IBR and IBR-30, as a result of which the reactivity changed smoothly). A characteristic feature of the IBR-2 regulator was that its input signal was pulsed (a discrete signal with time quantization) while the output signal (the reactivity introduced by the regulator) was a discrete signal with level quantization. There was no direct relation between the moments in time when level (output) and time (input) quantization were performed.Considering this feature of the regulator, a special mathematical model was developed for calculating transient power processes. In the model used previously for IBR-2 dynamics [1-3], level discreteness was neglected and the regulator was considered to be a continually operating component. Just as the previous model, the latest model took account of the nonlinear dependence of the energy of the power pulse on the reactivity and the effect of reactor heating on the reactivity by introducing power pulse energy -reactivity nonlinear feedback.The relative amplitude of the power pulse p = P m /P 0 m , where P m and P 0 m are the amplitude and base value of the power pulse, respectively, was adopted as the regulated parameter of the reactor. For IBR-2, the relative amplitude of the power pulse is equal to the relative energy e of the power pulse: p = e = E/E 0 , where E 0 is the base energy E of a pulse. The regulator motor generated a control signal in the following sequence. First, the amplitude P m of each power pulse was recorded and its relative value p was generated. The sequence of relative amplitudes was fed into the smoothing block, which generated a continuous signal { with a stepped shape. The step height and the time interval between the nth and (n + 1)st pulses was generated according to the law { n = { n-1 + (p n − { n-1 )/q, where q is a smoothing factor, which can take on one of four values: 4, 8, 16, and 32 (standard values). The model also allows for the value q = 1, which corresponds to no smoothing of the signal p. The smoothing block, operating according to this law with q > 1, consists of an aperiodic (slow) unit with input signal p and output signal { [4]. Its time constant and the transfer coefficient are, respectively,