CHNN型神经网络在双效溴化锂吸收式制冷机的参数优化设计中的应用

CHNN Neural Network Used in the Optimizate Parameter Design of Double-effect LiBr Absorption Chiller

对于溴化锂制冷机组的参数优化设计,传统的做法是先建立数学模型,然后根据坐标轮换法求得最优解,但对于双效、双级甚至三效制冷机,该方法的运算量大,建模复杂。本文利用CHNN型神经网络建模,把目标函数和约束条件转化为能量函数,进行优化计算,计算速度大大提高,且不受空间维数限制,并且模型可直接转化为电路,求得最优解。

The optimize parameters design of LiBr absorption chiller set formerly up mathematical model before all others, the function of optimize computation is carried out. This method enhances the speed en get the most excellent solutions by coordinates cyclic method, but regarding double-effect, two-stage even three effect refrigerators operand big, the modeling is complex. This article uses the CHNN nerve network, transforms the objective function and the constraint condition as the energy of calculating and isn＇t restricted by dimension of space. Once the model is made successfully, it can be directly transform into electric circuit, getting its the most optimal solutions.