System for controlling the charge/discharge cycles of a rechargeable battery and host device including an intelligent battery

摘要

The control system is connected to a rechargeable battery (110) having charge phases alternating with discharge phases in a charge discharge cycles, The system includes neural networks formed from microprocessors (160) and memory zones(170). A first neural network (NN1) is arranged to collect, at the start of a discharge phase in a charge/ discharge cycle, a group of operating parameters called first parameters (WjA) and to receive as input a predetermined value for a critical step discharge voltage (VTH). The network is arranged to produce a predictive output calculated for the instant (tTH) when the battery reaches the critical step (VTH) corresponding to the end of the discharge phase, The system has also: second and third neural networks (NN2,NN3), connected to the first (NN1), which are arranged to receive as input, at an initial instant, of the aforesaid discharge phase, 1 - a battery voltage value, called the initial voltage (Vo), a variation value for this initial voltage of ( delta Vo) after a short time lapse from the instant of the initial voltage and 2 - a value (No) for a number of initial discharge/charge cycles of the battery, to be carried out prior to the aforesaid discharge phase. The neural networks are arranged to provide as output, at the instant of the discharge phase, where the initial values are available, a set of approximate operating parameters corresponding to a correction (WjC) which are added so as to provide the first set of operating parameters (WjA) to be input to the first network (NN1).