*Figure 9. Simulation results showing the MV moves as a function of MV horizon for a given disturbance (MV shown in blue, CV in red and DV in purple). *
图9：当给定一干扰后MV时域函数下MV动作仿真结果（蓝线是MV，红线是CV，紫线是DV）
各个控制器的仿真运行结果如上图所示。该图显示了改变MV时域时控制器响应的差别，最顶部是最短的时域，图中越往下时域越长，在图的底部达到最大的时域。该图表明最短的MV时域将导致过于激烈的控制作用，其中MV计划动作实际上已经触碰到用户定义的最大动作限制。显而易见的是仿真中在蓝色曲线（MV）中绘制的直线。当MV时域变得更长时，SMOCPro有更长的时间达到其目标值。这些到达目标值的额外时间使得MV动作更为平滑。同样明显的是，MV时域越短，被控的CV越不易被扰动，对干扰的动态响应更快，CV驱动回设定值的速度也越快。
POV Impulse Factor（脉冲因子）
本教程将说明改变POV的Impulse Factor（脉冲因子）对控制器行为的影响。
流程模型
在这个例子中，我们利用之前练习中的单输入单输出液位控制应用，但在模型中引入可测量噪声。SMOCPro控制器模型将包含1个操作变量（出料流量阀位OP）、1个过程输出变量（储罐液位）和2个可测量干扰（进料流量和可测量噪声）。控制器模型编译后周期为1.0min，其结构由下图给出。模型块的颜色代码表示了所关注的不同变量，并且与仿真曲线中的颜色相同：可测量干扰变量是紫色，不可测（传感器）干扰是绿色，被控变量是红色，操作变量是蓝色。
Figure 10. Modified process model for the tank vessel control problem.
图10：油罐控制问题的修正过程模型
控制器设计
让我们根据前面的例子建立缺省MV时域（100）的控制器，其中单个POV（储罐液位）定义为CV，并具有以下调整权重：
|名称 |Damping（阻尼）| Weight（权重）| 名称 |Deviation（偏差）| Weight（权重）|
| ------------- |:-------------:| -----:|
|Outlet Flow OP |1.0 |0.5 | Vessel Level |1.0| 1.0|
原文：
The results of the simulation runs under the various controllers are shown on the figure above. The figure shows the difference in controller responses due to changing MV horizon, starting with the shortest horizon at the top, getting successively longer down the figure and culminating with the longest horizon at the bottom of the figure. The illustration shows that the shortest MV horizon results in overly aggressive control action where the MV planned moves actually hit the Maximum Move limit defined by the user. This is evident in the simulation by the straight line being drawn in the blue curve (MV). As the MV horizons get longer, SMOCPro has more time to reach its targets. This extra time to reach targets results in smoother MV action. It is also evident that the shorter the MV horizon the tighter the CV is controlled both in terms of its dynamic response to the disturbance as well as how quickly the CV is driven back to its setpoint.
POV Impulse Factor
This tutorial illustrates the effect that changing a POV’s Impulse Factor has on controller performance.
Process Model
In this example, we utilize the same single-input single-output Level control application from the previous exercise but introduce a way to incorporate measurement noise into the model. The SMOCPro controller model has one manipulated variable (outlet flow valve position OP), one process output variable (vessel level) and two measured disturbances (inlet flow and measurement noise). The controller model is compiled with a period of 1.0 minute and the structure is given in the figure below. Again, the color coding in the model blocks indicates the different variables of interest and will be the same color of the simulation curves: measured disturbance variable in purple, unmeasured (sensor) disturbance in green, controlled variable in red and manipulated variable in blue.
***Controller Design ***
Let us build the controller with the default MV horizon (100) from the previous example, containing the single POV (Vessel Level) defined as a CV and with the following tuning weights:
| Name | Damping| Weight | Name | Deviation| Weight|
| ------------- |:-------------:| -----:|
| Outlet Flow OP | 1.0| 0.5 | Vessel Level| 1.0 | 1.0|
2016.5.21
