Methodology for Continuous Infeed Cylindrical Plunge Grinding

本计画将发展一种新的外圆切入磨削循环最佳控制方法。 在这种类型的研磨中，一个循环通常由顺序施加的几个离散的横向进给速率组成。 考虑有限数量的离散横向进给速率使得该问题易于处理并且允许该循环容易地用较旧的机器控制技术（例如，梯形逻辑）。 但它限制了优化减少周期时间的程度。 该方法的基础将是模型预测控制（MPC），其中每个采样时刻的进料速率值根据其对过程和零件质量约束的预计影响来确定。 然而，由于研磨过程的固有不确定性，MPC的基本形式将不足以用于该应用。 为了增强其鲁棒性，该控制将与递归约束定界相结合，以便将保守性纳入优化过程。最优控制方法能够提供连续变化的进给速率的循环。 这将导致周期时间和成本的显著节省，并且更适合于使用开放式架构控制器来实现控制。

This project will develop a new methodology for optimal control of cylindrical plunge grinding cycles. In this type of grinding, a cycle generally consists of a few discrete infeed rates applied sequentially. The consideration of a limited number of discrete infeed rates makes the problem tractable and allows the cycle to be readily implemented with older machine control techniques (e.g., ladder logic). But it limits the extent of cycle-time reduction by optimization. The foundation of the methodology will be Model Predictive Control (MPC) where the value of infeed rate at each sampling instant is determined according to its projected effect on the process and part quality constraints. However, the basic form of MPC will not be adequate for this application because of the inherent uncertainty of the grinding process. To enhance its robustness, this control will be integrated with Recursive Constraint Bounding so as to incorporate conservatism into the optimization process.The optimal control methodology is able to provide a cycle with continuously varying infeed rates. This should lead to significant savings in cycle time and cost, as well as being more suited to implementation of the control with open-architecture controllers.