中文摘要：

An immune algorithm-based approach was developed to optimize a feedforward neural network. The network architecture, activation functions, and training method were encoded as individuals with an ap- propriate method for individual selection. The immune feedforward neural network is then applied to fault detection of water quality monitoring equipment. This gives better performance than a feedforward neural network.

英文摘要：

An immune algorithm-based approach was developed to optimize a feedforward neural network. The network architecture, activation functions, and training method were encoded as individuals with an ap- propriate method for individual selection. The immune feedforward neural network is then applied to fault detection of water quality monitoring equipment. This gives better performance than a feedforward neural network.