DMCplus中关于定义斜坡变量有几种不同的方式。这些类型的简要描述如下。欲了解斜坡类型的更多细节，请点击链接。
•未失衡的传统斜坡
这是早期DMCplus版本提供的斜坡类型。它允许使用平均液位控制。稳态优化总是试图平衡斜坡，若不能则控制器会处于关闭状态。斜坡变量必须定义为对应用程序非常重要。
•失衡的传统斜坡
该选项提供了用户指定时间间隔允许斜坡失衡，控制失衡不是非常大。
它也允许使用平均液位控制。稳态优化总是试图平衡斜坡;然而若不能平衡，在控制器关闭前，其将在失衡情况下继续运行用户指定的控制时间间隔。而若任意时刻失衡过大，不管其是否已达到用户定义的不平衡周期，控制器都将被关闭。
该方法用于确定操作员提供的操作极限失衡容忍。斜坡变量必须定义为对应用程序非常重要。
·编程失衡斜坡
在此选项中，稳态优化并不试图平衡斜坡;斜坡允许无限期失去平衡。任一方向上失衡容忍量是当前值接近操作限程度的函数。
当值靠近极限时，失衡容忍量降低到零。这个选项主要用于当DMCplus控制器存在一个大的激增容量时，控制器将花费几个稳态时间到达设定液位操作限时。当某时刻流程上下游不在服务时，这是特别有用的。再次强调，斜坡变量必须定义为对应用程序非常重要。
·伪斜坡
这一选项是为那些不是真实的积分变量，但在我们感兴趣的时域里表现出斜坡行为的变量提供的。最常见的例子是高纯精馏塔的塔盘温度。失衡的传统斜坡的稳态优化总是尝试着平衡斜坡，但如果不能平衡，控制器允许保留运行用户指定的时间间隔。
伪斜坡与失衡的传统斜坡的主要差别在于，前者不检查失衡的量。尽管伪斜坡必须明确地被控制，但它们不要求对应用程序非常重要。需要注意的是伪斜坡是在DMCplus Build中定义的。
斜坡变量的RAHEDSUB选项可用于当遇到斜坡失衡时配置子控制器的行为。如果RSHEDSUB选项设置为1，控制器将只留下包含不平衡斜坡的子控制器。如果RSHEDSUB选项设置为0，整个控制器将留下。当RSHEDSUB选项处于启用状态时，控制器将在当前周期执行失衡方案，然后在下一控制周期开始时关闭子控制器。这将使得最新方案得到执行的同时避免重新求解导致的错过周期或组合方案时间的问题。
附原文：
There are several different ways in which a ramp variable can be defined in DMCplus. A brief description of each of these types is listed below. For more detail about a ramp type, click its link.
·Traditional Ramp with No Imbalance
This is the ramp type that is provided in earlier versions of DMCplus.It allows averaging level control. The steady-state optimization is always trying to balance the ramp, and if it cannot, the controller is turned off. The ramp variable must be defined as critical to the application.
·Traditional Ramp with Imbalance
This option is provided to allow the ramp to be out of balance for a user-specified number of intervals, provided the imbalance is not too large.
It also allows averaging level control. The steady-state optimization is always trying to balance the ramp; however,if it cannot, the controller will continue to run out of balance for the user-specified number of control intervals before the controller is turned off.If at any time the imbalance is too large, the controller turns off anyway,regardless of whether it has reached the user-specified number of imbalanced cycles.
The method used to determine what is too much imbalance is based on the operating limits provided by the operator. The ramp variable must be defined as critical to the application.
In this option, the steady-state optimization does not attempt to balance the ramp; the ramp is allowed to be out of balance indefinitely. The amount of imbalance tolerated in either direction is a function of how near the current value is to the operating limit.
As the value approaches the limit, the amount of imbalance allowed decreases to zero. This option is used primarily when there is a large surge capacity in the middle of a DMCplus controller, where it may take several steady-state times of the controller for the level to reach the desired operating limit. This is especially useful when the upstream or downstream section of the process may be out of service sometime. Again, the ramp variable must be defined as critical to the application.
This option is provided for those variables that are not true integrators,but that act like ramps over the time domain of interest. Most commonly, these are tray temperatures in a high purity distillation tower. As with the Traditional Ramp with Imbalance, the steady-state optimization is always trying to balance the ramp, but if it cannot, the controller is allowed to remain on for a user-specified number of intervals.
The primary difference between pseudo-ramps and traditional ramps with imbalance is that for pseudo-ramps, no check is made on the amount of imbalance. And while pseudo-ramps must be explicitly controlled, they are not required to be critical to the application.Note that pseudo-ramps are defined in DMCplus Build.
The RSHEDSUB option on ramp variables can be used to configure the behavior of subcontrollers when a ramp imbalance is encountered. If the RSHEDSUB option is set to 1, the engine will only shed the subcontrollers that contain the imbalanced ramp. If the RSHEDSUB option is set to 0, the entire controller willbe shed. With the RSHEDSUB option enabled, the controller will implement the imbalanced solution at the current cycle and then shut the subcontrollers off at the beginning of the next control cycle. This allows the latest solution toget implemented while avoiding the overhead of re-solving the problem which could cause missed cycles or a Composite Suite timeout.
2015.9.28
