我们来考虑基准情景，在工作区中该液位CV是优先级被定义为最低优先级的被控变量。
上图显示了“Baseline_Level_Prio_50”情景的仿真结果。在TDZ干扰引入过程后，SMOCPro马上按最高优先级变量到最低优先级变量的顺序实施动作以满足其控制目标。首先，它解决了skin temperatures（表层温度）（Skin Temps – 优先级 = 1），其次将不同的flows（流股）渐渐地纳入控制之内(DFs – 优先级 = 10)，紧接着是delta temperatures（增量温度）（TDZs-优先级= 30）。在这时SMOCPro意识到它不能满足优先级从1跨越到30的所有控制目标，因此其根据各自的稳态增益和用户指定的CV deviations（误差）（权重）放松了TDZs。在此阶段只需要计算一个剩余可到达目标，即液位CV，记住该斜坡CV有1个斜坡稳定约束和1个目标计算。然而，SMOCPro为更高优先级的CV占用了所有的MVs。如图右下象限所示，这导致SMOCPro放弃稳定和控制Level。
接下来，我们来关注当Level CV的优先级定为30的情况。下图显示了此方案的仿真结果。显而易见的是在这里SMOCPro选择控制Level CV到设定点，而放弃了TDZ CVs。要知道为什么会发生这种情况，让我们来分析SMOCPro在每个优先级基础上的计算。正如在基准情况下，优先级最高的CV是skin temperatures，而这些目标将首先达到。接下来DFs在慢慢地引向设定点。现在剩余的所有CVs都具有相同的优先级（30）。然而，斜坡CV（Level）也包含了在CVs目标被解决之前的稳定约束。鉴于在该点可以达到稳定，SMOCPro有能力稳定level。这种稳定在稳态处理MVs的值，并给Level CV留下空间以达到其设定点。剩下的CVs（TDZ）根据其工作点，稳态增益和CV偏差（权重）适当放宽。这个例子强调了一个在准则文档中提到的条款—即：面对稳定的斜坡CVs都是相同优先级时，和可能受限的MVs，由于稳态约束的缘故，斜坡CV将对MV动作计算有更强的影响。最后值得注意的是图的右下角象限也显示了Level的Priority = 20时的情况。因为priority = 30或20时的两种情况稳态约束都首先得到解决，Level CV的行为是相同的。这是因为原来的不可行性位于优先级=30的TDZ CVs。
原文：
Consider the baseline scenario, where the Level CV priority has been defined as the lowest priority controlled variable in the workspace.
The figure above displays the simulation results for the “Baseline_Level_Prio_50” scenario. Immediately after the TDZ disturbances are introduced into the process SMOCPro implements moves to meet its control objectives starting with the highest priority variables and moving down to the lowest priority variables. First, it addresses the skin temperatures (Skin Temps -Priority = 1), next the differential flows are slowly brought within control (DFs - Priority = 10), followed by the delta temperatures (TDZs - Priority = 30). At this point SMOCPro realizes that it cannot meet all the control objectives spanning Priority 1 to 30 and relaxes the TDZs based on their steady state gains and the user-specified CV deviations (weights). At this stage there is only one remaining reachable target to calculate and that is the Level CV, remembering that this ramp CV has a ramp stabilization constraint and a target calculation. However, SMOCPro has already tied up all MVs with the higher priory CVs. This results in SMOCPro giving up on stabilizing and controlling the Level as seen on the bottom-right quadrant of the figure.
Next, we look at the case where the Level CV has been defined at Priority 30. The figure below shows the simulation results for this scenario. Here it is clearly evident that SMOCPro controls the Level CV to setpoint and gives up on the TDZ CVs. To see why this happens let us analyze the SMOCPro calculations on a per priority basis. Just as in the baseline case the highest priority CVs are the skin temperatures and those objectives are met first. Next the DFs are brought slowly to their setpoints. Now all the remaining CVs are at the same priority (30). However, the ramp CV (Level) also contains the stabilization constraint which is solved before the targets for the CVs. Since at this point the stabilization can be met SMOCPro is able to stabilize the level. This stabilization fixes the value of the MVs at steady state but leaves room for the Level CV to reach its setpoint. The remaining CVs (TDZ) are relaxed based on their operating points, their steady state gains and the CV deviations (weights). This example highlights one of the items that is mentioned in the guidelines document—that is, in the presence of stable and ramp CVs at the same priority and with possibly constrained MVs the ramp CV will have a greater influence on the MV move calculation due to the stabilization constraint.Lastly, it is worth noting that the bottom right quadrant in the figure also shows Priority = 20 for the Level. The behavior with the Level CV at priority = 30 or 20 is the same since the stabilization constraint is solved first for both cases. This happens because the original infeasibility lies in the TDZ priority 30 CVs.
2016.6.6
