Abstract
In developing an automated intelligent system for a cold water supply dispatching center, the object model in certain cases consists of a set of models having the same general external structure (type) but different parameters. The most commonly used model is a cascade connection of a first- or second-order filter with a time-delay element. As an example, this approach can be applied to the water supply system of a residential complex or a group of residential buildings. According to materials published in scientific literature, such an object can be approximated by a simplified model: a set of externally similar models with different parameters may be replaced by a single model of the same type with averaged parameters. However, using simple arithmetic averaging of numerous parameters appears to be an unjustified approach. Therefore, it is more appropriate to apply numerical optimization methods to determine the parameters of the approximating model based on signals obtained from water pressure sensors. For linear models, the most suitable test signal is a single step input (unit step response). This article presents an analysis based on the above assumption. The simulation results of the step response for third-, fourth-, fifth-, and sixth-order filters are provided. Additionally, the results of modeling the response of ninth- and twelfth-order filters to two rectangular pulses with durations of 0.4 s and 0.6 s are presented; the number indicated near each graph represents the operating mode of the filter. Furthermore, the modeling results for the response of ninth- and twelfth-order filters to two rectangular pulses with durations of 0.2 s and 0.7 s are shown, where the number adjacent to each graph also indicates the filter operating mode.