控制器设计
输出选项：
如下所示，4个控制器都被考虑为单个POV（储罐液位）定义CV而构建：
惩罚项：
4个控制器包含了以下调整权重：
| 名称 | Damping（阻尼） | Weight（权重） |
|---|---|---|
| Outlet Flow OP | 1.0 | 0.5 |
| 名称 | Deviation（偏差） | Weight（权重） |
| Vessel Level | 1.0 | 1.0 |
压缩点：
我们所考虑的4个控制器都搭建有以下的压缩点：
|控制器名称 |输入时域：最近的压缩点 | 稳定时间 | 输出时域：最近的压缩点 |
| ------------- |:-------------:|: -----:|-----:|
| 缺省| 100| 1 |101|
|Short_MV_Horizon| 10| 1| 11|
|Medium_MV_Horizon| 50| 1 |51|
|Long_MV_Horizon| 150 |1 |151|
需要注意的是自动计算（默认）出的MV时域被随机指定为100 × dT。这是当没有正确的瞬态动力学时，必须指定对MV时域作一个决定的情况下。在这种情况下，我们感兴趣的是理解当改变MV时域时，对控制器性能的影响。仔细选择时域涉及到的一些附加输入/输出压缩点以允许控制器增大/减小它的预测范围，同时计算动作计划。
为了研究不同MV时域的影响，将默认控制器拷贝到一个新的名称为“Short_MV_Horizon”的控制器。定位到压缩点选项卡并单击Calculator（计算器）按钮。在这里人们可以方便地访问子控制器的输入时域和稳定时间。计算器窗口被设定为如下所示，以便允许动作计划的一些自由度，以及更多的控制误差预测贡献。如下所示，对Short_MV_Horizon的输入时域指定为10：
然后生成一个压缩列表，Default in Use（默认使用）按钮被切换到Reset to Default（重置为默认值）。
继续搭建中/长MV时域的两个控制器。
仿真
仿真的目的是为了检查当改变MV时域时对控制器性能的影响。
所有的仿真场景的初始工艺操作条件如下：
储罐液位起始于50（Setrange High设定高限 = Setrange Low设定低限 = 50）。
MV的初始值稳定在50，最大值为100，最小值为0，最大动作步幅为2。
进口流量Inlet Flow (DV)的初始操作点为0。
在第10步我们对进口流量Inlet Flow引入一个10单位的斜坡干扰。我们运行100步仿真，并观察SMOCPro计算的将储罐液位带回设定点的动作计划。这4种情况下最后执行步(101)的值都是一样的，MV=60, CV=50 ,DV=10。
原文：
***Controller Design ***
**Output Selection: **
Four controllers are considered all built with only the single POV (Vessel Level) defined as a CV as such:
**Penalties: **
The four controllers containg the following tuning weights:
|Name| Damping| Weight| Name |Deviation| Weight|
| ------------- |:-------------:| -----:|
|Outlet Flow OP| 1.0 |0.5 | Vessel Level| 1.0 |1.0|
Compaction points:
The four controllers under consideration are built with the following compaction points:
|Controller Name| Input Horizon (Last Comp. Point) |Settling Time | Output Horizon (Last Comp. Point)|
| ------------- |:-------------:|: -----:|-----:|
| Default| 100 |1| 101|
|Short_MV_Horizon |10 |1| 11|
|Medium_MV_Horizon| 50| 1 |51|
|Long_MV_Horizon| 150 |1 |151|
Notice that the automatically computed (default) MV horizon is arbitrarily specified to be 100 × dT. This is the result of not having true transient dynamics and having to make a decision as to what to specify the MV horizon to be. In this case, it is of interest to understand the effect that changing MV horizon has on controller performance. Carefully selecting the horizon involves appending some input/output compaction points to allow the controller to increase/decrease its prediction horizon while computing the plan of action.
To study the effect of varying the MV horizon, copy the Default controller onto a new controller called “Short_MV_Horizon.” Navigate to the Compaction Points tab and click on the Calculator button. Here one has easy access to the input horizon and the settling time of the sub-controller. The Calculator window is set as follows to allow some freedom in the plan of action as well as some more prediction contribution of the control errors. For the Short_MV_Horizon case specify 10 for the Input Horizon as follows:
Then, a list of compaction points is generated and the Default in Use button is toggled to Reset to Default.
Continue to build the remaining 2 controllers with the Medium and Long MV horizons.
Simulation
The objective of the simulation is to examine the effect of changing MV horizon on controller performance.
The initial process operation conditions for all the simulation scenarios are the following:
The Level in the vessel starts at 50 with a Setrange High = Setrange Low = 50.
The MV is initially holding steady at 50 with a Maximum value of 100, Minimum value of 0 and Maximum Move Size of 2.
The Inlet Flow (DV) starts at an operating point of 0.
At step 10 we introduce a ramp disturbance of 10 units into the inlet flow. We run the simulation for 100 steps and observe the planned moves that SMOCPro calculates to bring the vessel level back to setpoint. For all four cases the values at the last execution step (101) are the same, MV=60, CV=50 and DV=10.
2016.5.21
