In order to overcome these disadvantages, a new Model-Driven design  and implementation method is proposed in this paper to enhance the development process of machine tool control system, in which: designers capture the system specification as much as possible and build an executable system model with some Computer-Aided Control System Design (CACSD) tools. The model has formal syntax and semantics, so that it can be simulated to eliminate the design errors before implementation. The control system derived from the model is pided into two portions: a generic process engine which implements the control algorithm and operation rules which are related to inpidual specification. The process engine is used to interpret the operation rules. Using this implementation scheme, only the operation rules need to be re-generated every time when the specification is modified while the engine remains unchanged.

The rest of this paper is organized as follows. Section 2 describes a typical machine tool control system structure. Section 3 discusses the StateChart-based Model-Driven discrete event control system design and implementation. Section 4 presents a case of a Mode

Control Supervisor design and implementation by using the proposed method. The paper concludes with section 5.

2 TYPICAL STRUCTURE OF MACHINE TOOL CONTROL SYSTEM

Figure 1 shows the typical structure of a machine tool control system which applies the foreground/background frame pattern mostly used in industry. The foreground is called real-time system (RTS), while the background is called free-time system (FTS). RTS involves two tasks: TIS and EIS which take care of time-based periodical interrupt service and event-based interrupt service respectively. Usually the PID algorithm of motor control, for example, is handled by TIS. Emergency stop is typically handled by EIS, etc. In the FTS, the system runs repeatedly along the sequence from DIG to OCS task. The external operation request and the internal system status will influence the task execution order. The complete execution sequence of the FTS is shown as follows:

1. The system starts from either ST (normal start-up) or RST (error recovery).

2. INZ task initializes the system, such as default parameter setup, work mode selection, etc.

3. DIG task checks system working status to judge whether the system has fault or not.

4. If the system has no fault, the control goes to SSC task, otherwise to ERH task.

5. If there is no fault after ERR, SSC task collects all the sensors and system status flags at this moment.

6. MCS determines at the current moment which task should be activated.

7. The control enters one of four kinds of work mode related tasks: ACS/MOS/PPS/PGS.

8. OCS task refreshes the entire system indicators and system status flags.

9. The control goes back to DIG task.

摘要：机床控制系统离散运动控制的设计与实现非常复杂。在当前的工业实践中，设计师往往倾向于依据系统规范分析来实现粗略的系统设计。这种设计方法设计出的系统，其系统性能高依赖于设计师的经验。通常一个长的“循环和调试”阶段需要在原型系统已经修复错误后建造。此外，当规格改变时，通过修改现有系统来构建新系统总是困难的。在本文中，作者提出了一个模型驱动方法来模拟机器的离散运动控制的设计和实现工具控制系统。基于系统规范，首先构建一个可执行模型。然后通过模拟仿真来测试该模型，以消除之前设计产生的误差。最后系统完成，一个单独从模型中获得操作规则的流程引擎。机床控制系统的关键模块用于说明所提出的方法。