During the spreading of liquid manure, the manure how must be controlled ac
cording to the position-dependent reference inputs and the tractor speed. T
he speed varies during starting up or stopping the tractor or by wheelslip,
particularly in hilly regions. Magnetic inductive flowmeters proved to be
the most precise and robust devices for measuring the manure's flow rate; h
owever, due to built-in measurement pre-processing, their signals suffer fr
om lags and delays. This means that feedback control of the liquid's flow r
ate is either very slow or must be operated with a small stability margin o
r even cannot be realized at all with conventional equipment. Two new contr
ol sq:stems are presented that incorporate model-based adaptive control. On
e is equipped with a supervisory adaptation loop, the other is based on ext
ended Kalman filtering with a Smith predictor. (C) 2001 Elsevier Science Lt
d. All rights reserved.