Utilization of Statistical Techniques in a Two Step Parameter Estimation for a Hydraulic Valve

This paper investigates two well known parameter estimation techniques, the ordinary least squares and the maximum likelihood methods, for parameter estimation in the main spool of a solenoid proportional valve. The methods are relatively easy to apply and are applied in two steps. The first step consists of operation with closed outlet ports. This eliminates flow forces on the spool enabling the spring constant and precompression to be estimated. These values are then used in the next step, which is the open port case, to estimate valve orifice area gradient and dead band. The results agreed quite well with experimental values determined using a graphical technique for spring constant and pre-compression. However, estimations of the dead band and orifice area gradient showed a strong temperature dependency, contrary to physical expectation.