Page 83 - Biosystems Engineering
P. 83

64    Chapter  Two

                     40



                     30

                   Magnitude (dB)  20



                     10



                      0


                    –10
                                      10 –1                 10 0
                                         Frequency (Hz)
               FIGURE 2.17  The frequency response averaged over four periods for ground
               speed. The response on the excited frequencies is divided by the normalized
               response on the pump setting. The solid diamonds (♦) indicate the noise
               level; the gray open circles ( ) indicate the excited odd frequencies; the open
               squares ( ) indicate the nonexcited odd frequencies; and the open diamonds
               (◊) indicate the (non excited) even frequencies. The solid line is a second-
               order system fi tted to this data set.


                   The response of the ground speed is shown in Fig. 2.17. Several
               observations can be made from this response.  At low frequencies
               (approximately 0.1 Hz), the response on the even frequencies, which
               are not excited, is considerable compared to the noise level. Even some
               of the odd nonexcited frequencies show a considerable response.
                   Also note that the response on the excited frequencies increases
               again at higher frequencies (more than 1 Hz). This is probably caused
               by propulsion system dynamics. The engine control unit determines
               the fuel injection of the diesel engine as a function of engine speed.
               There are no details about this factory-fitted controller available, but
               a differential action on this controller might result in such a bump on
               the ground speed spectrum.


               2.4.4 Parameter Estimation
               Now that we have empirical data for the system, we can use them to
               estimate the parameters of our system model. The objective is to find
               the model parameters for which the simulated model behavior cor-
               responds best with the measured model behavior.
   78   79   80   81   82   83   84   85   86   87   88