Constraints Section
Applicability: Synapse

Every design and selection contains one or more constraints which viable candidates must satisfy. Design constraints can be imposed on a single physical property, such as density or viscosity, or on a complex function of physical properties such as a heat transfer coefficient calculated by a Nusselt number correlation.

Each constraint contains a function name, a minimum value, a goal value and a maximum value. Some example constraints are shown in the following table.

Function Minimum Goal Maximum
Freezing Point [°C] -40 -10 0
Liquid Density at 20°C [kg/m3] 850 1000 1150
Heat Transfer Coefficient at 20°C [W/m2 K] 1000 1200 2000
Example: Entering a design constraint

In this example we will enter a constraint for a graphical chemical design. The same constraint can be used for combinatorial chemical designs and chemical selections. In fact, you can copy and paste the constraints between these documents.

Similarly, the constraints used in a graphical mixture design can also be copied and pasted into combinatorial mixture designs and mixture selections.

  1. Open the MKS Chemical Design Examples document. Save a "working" copy of the document. (See documentation on Copying Documents for details.)
  2. Change to the Graphical Designs chapter and navigate to the page displaying the "RI Matching Fluid Design".
  3. Scroll the datapane and click the left mouse button on the Constraints Section's large table control. The application activates the Edit Constraints dialog.
  4. The example design already has several constraints. Click the left mouse button on the dialog table's first empty row. Then press the dialog's Edit button. The application activates the Edit Constraint dialog.
  5. Click the left mouse button on the List button in the dialog's Constraint Function section. The application activates the Select Function dialog which lists all the design functions in the current document. Select the Melting Point function and press the dialog's OK button.
  6. Our goal is design a refractive index fluid that will not freeze at temperatures above -20°C. Thus, enter a minimum value of -80, a goal value of -20 and a maximum value of 0. (We use a fairly high maximum limit because melting point estimates often contain significant errors.) Optionally, enter values into the Reference and Comment controls. Finally, press the dialog's OK button.
  7. Finally, press the Constraints Dialog Save button to save the entered constraints into the current document.

See documentation on the Constraints Dialog for additional details on editing constraints.

Tip: Compile property values to establish constraint limits

The knowledge base's Compile Property Values command can be helpful in identifying the physical property limits to use for your constraints. (See documentation on Compile Propery Values for details.) For example, the following image shows a compilation of melting points in units of °C.

Using the dialog's various controls and displays you can analyze typical physical property ranges.

Tip: Always consider estimation error when setting constraint limits

The physical properties of most design candidates and many selection candidates will need to be estimated in order to evaluate a constraint. These estimated property values, even those estimated by very accurate techniques, will contain estimation errors. It is important to consider this error when setting contraint limits.

For example, the evaluation of the "Tm: Joback Method" estimation technique is shown in the image below. (See documentation on the Technique Evaluation Dialog for details on evaluating estimation techniques.) The statistics were determined only for those chemicals whose melting points were between -100°C and 50°C.

The average absolute error (Avg Abs Error) shown in the dialog's Statistics section is 26.08. If our constraint is meant to select candidates with freezing points below -20°C, then we might consider an maximum limit on the constraint of 6.08°C, i.e.

limit = goal ± error

max limit = (-20.0) + (26.08) = 6.08

If we do not account for estimation error, then our constraint will be discarding candidates whose actual melting points are satisfactory but whose estimated melting points are not because they are estimated to be too high.

Related Documentation