Kngwledge-Based Qualitative Modelling and Adaptive Distribution of Power

Abstract

Sudden surge in power demand or reduction in power supply aboard a vehicle typically require the removal of the less critical loads to maintain the operation of the more critical ones. Conventional approaches assign a priority value to each load and distribute power to the loads in the order of priority, with each load receiving its nominal rated power. As a result, many lower priority loads may be unserviced. However, a large class of loads can operate within a range of input power below the nominal value, albeit with reduced performance. For this class of loads, the degradation in functionality with decreasing input power a continuous, not discrete. The optimal power distribution is conceivably one in which these loads receive only partial power, with a resultant increase in the number of loads serviced and an increase in the overall vehicle functionality. We have developed a method that automates the distribution of partial power to maximize overall vehicle functionality. In addition to the use of priority values, we use fuzzy sets to represent the relationships between the fully functional state of a load and its input power. Using this description, power is iteratively distributed to the loads using fuzzy logic decision rules, taking into account the priority of each load and its power requirement relative to the remaining loads in the set. This algorithm represents an initial alternative to the conventional all-or-none power distribution method. The algorithm is also efficient and practical for real-time power redistribution during contingencies.