A method for distributed simulation of a continuous control system running on a single machine

Abstract

The invention discloses a distributed simulation method for a continuous control system in unit operation, and belongs to the technical field of computer simulation. The method comprises the following steps: firstly, dividing sub-models forming an initial system into two types namely straight-through models and non-straight-through models; secondly, splitting the initial system and deploying each split part on a separate simulation node to run to form a distributed simulation structure, wherein the straight-through model is prevented from being deployed on the separate simulation node during the splitting; finally, setting a simulation period in the distributed environment according to a calculated maximum simulation period. According to the method, the initial system is not required to be modified; data delay generated when the continuous control system in unit operation is converted into one in distributed operation due to the characteristics of parallel computation is reduced by the first two steps; the maximum simulation period is determined to enable the converted distributive system to keep the stability characteristic of the initial system; the data exchange frequency among all distributed simulation nodes is reduced to the greatest extent.

Description

technical field [0001] The invention belongs to the technical field of computer simulation, and in particular relates to a method for distributed simulation of a continuous control system operated by a single machine. Background technique [0002] Traditionally developed continuous control systems generally run on a single machine. However, in some specific situations, it is necessary to run these systems in a distributed environment. A typical example is a flight modeler with a distributed structure, its attitude controller runs on one simulation node, and the controlled object (aircraft) model runs on another simulation node. [0003] For the convenience of description, the continuous control system running on a single machine is recorded as CS, or the initial system; the transformed control system running in a distributed environment is called DS, or the target system. According to the general form of "controller-controlled object", CS can be expressed as follows: [0...

AI Technical Summary

Problems solved by technology

[0023] The conversion from CS to DS can be completed through parallel programming, but it takes a lot of time and cost to modify the model code

But on the other hand, if the direct conversion without any processing, as shown in equations (1a), (1b) and equations (2a), (2b), will cause the initial system to be different from the target due to the introduction of discretization and delay factors There is an error between the state trajectories of the system

Images