Building a Process Model-Using transforms for model building-2

AIDAPro可用于鉴定转换域（线性）的模型。AIDAPro拥有一个可执行各种变换的计算引擎。然而，AIDAPro不具有直接将变换表（查找表）自动传递给SMOCPro的能力。因此，用户必须将AIDAPro和SMOCPro中的变换表保持一致。此外，必须仔细确认变换表囊括了整个操作范围。变换表中的值可以被复制粘贴到Microsoft Excel®电子表格。
注：我们建议不要将转换应用于中间POVs。

Figure Input Transform 输入变换图

Figure Input-Output Transform 输入输出变换图

为了说明SMOCPro中的变换功能，我们考虑自AIDAPro中导入一个精馏塔模型。所得模型如下图所示。

在本例中，再沸和回流量通过调节顶部和底部温度中间变量来调节塔顶和塔釜质量。我们需要对回流进行适当的变换（本例中是对数）以更好地实现我们的控制目标。建立模型后，你需要对感兴趣的变量指定变换。本例中要做到这一点，可双击回流模块，并单击Add Transform（添加变换）按钮。
接下来你需要在变换模块中输入分段线性变换。下图显示了变换表。记住变换表格可以从Microsoft Excel®电子表格复制粘贴。值得注意的是在进行这些变换时，用户必须保持值与所辨识的模型一致。最后，在转化的MV,DV和/或POV模块的左上角将显示一个绿色三角形，以证明其是变换变量。

该变换表具有以下重要特征：
•坐标X是工程单位→即测量值；
•坐标Y是变换单位→即变换值；
•每对X-Y都定义了一个内插值（或外插值）数据点；
•变换表最多可包含10个点；
•可通过设置点数为0删除变换；
•两个相同的连续值是不允许的，因为它们的逆变换不能确定；
•最后非常重要的是X和Y坐标系的所有值必须严格单调（请参阅以下的允许函数例子）。
变换表的最后一个要求是单调性。对给定的函数，单调性具有当自变量（X）增加时，因变量始终保持持续增加或持续减小的特性。变换表允许的函数如下图所示。

---

原文：
To identify the model in the transformed (linearized) domain, AIDAPro can be used. AIDAPro has a calculation engine that can perform all kinds of transforms. However, AIDAPro does not have the ability to automatically pass the transform table (lookup table) directly to SMOCPro. Therefore, the user must keep the transform tables in AIDAPro and SMOCPro consistent. In addition, careful consideration must be taken so that the transform table covers the entire range of operations. The values in the transform table can be copied and pasted from a Microsoft Excel® spreadsheet.
NOTE: It is our recommendation that the transforms NOT be applied to intermediate POVs.
To illustrate the transform functionality in SMOCPro, consider the distillation column example as imported from AIDAPro. The resulting model is shown in the figure below.
In this example, the reboil and reflux regulate the column’s top and bottom qualities with the top and bottom temperatures being intermediate variables. We need to transform the reflux with a suitable transformation (in this example, logarithmic) to better achieve our control objectives. After building the model, you need to specify the transform for the variables of interest. To do this in this example, double click on the reflux block and click on the Add Transform button.
Next, you need to enter the piecewise linear transform in the transform block. The figure below shows the transform table. Remember that the transform table can be copied and pasted from a Microsoft Excel® spreadsheet. It is worth noting that while entering these transforms the user must keep the values consistent with the identified model. Lastly, a green triangle is added to the left corner of the transformed MV, DV and/or POV blocks to indicate a transformed variable.
The transform table has the following important characteristics:
•The X-coordinate is in engineering units → this is the measurement
•The Y-coordinate is in transformed units → this is the transformation
• Each X-Y pair defines an interpolation (or extrapolation) data point.
• The transform table can have a maximum of 10 points.
• A transform can be deleted by setting the number of the points to zero.
• Two identical consecutive values are not allowed because they make the inverse transform undefined.
• Lastly, (extremely important) all values in the X- and Y-coordinates must be strictly monotonic (see below for examples of allowed functions).
The last requirement in the transform table is that of monotonicity. For a given function, monotonicity is having the property of never increasing or of never decreasing as the values of the independent (X) variable increases. The allowed functions for the transform table are illustrated in the following figure.

---

2016.9.8