In this paper, PID neural network, which is an adaptive controller, has analyzed and compared with two other conventional PID algorithms through computer simulation and experimental study. Cancellation and pole placement are the two selected conventional algorithms. In the simulation study, the effects of factors such as non-minimum phase behavior and model changes on the performance of schemes are investigated. In the experimental study, performance of controllers on pressure control of two serial tanks is investigated. Simulation and experimental results demonstrate that PID neural network can be tuned easily and has better performance in compare with two conventional schemes.

Keywords