Minimize Modeling Time!

Source:
Dr. Richard Lowrey, Mechanical CAD Consultant.

The ideal histry tree will have several levels of constructions as shown by the model classes in Fig. 1 bekow. The Class 1 tree should never be used for complex models. Modeling and editing time will both be greatly reduced with the use of multi-level constructions. Because the model is built in levels, deleting the Boolean associations of constructions to the model, or of individual shapes to each construction, will only delete a few specific Boolean or surface intersection associations.

Figure 2 illustrates the amount of modeling time required verses the number of operations in a Class 1 model. Fig. 3 shows the same model sub-divided into six constructions of a Class 2 model. Note the amount of CPU time saved during construction.

Figure 1 - Classes 1-4 history trees.
The subdivision of operations into discrete constructions within a solid model creates four distinct classes. (a) Class 1 models have a single level of constructions; (b) Class 2 models have two levels of constructions; (c) Class 3 models have three levels; (d) Class 4 models have four levels.

Figure 2 - CPU time verses the number of operations in a Class 1 model.
The amount of time necessary to perform a Boolean or surface intersection operation has a small overhead time "b." If the operation fails immediately, the CPU time "b" is still expended. Each subsequent operation of the same complexity takes a little longer ("a") because more model surfaces must be checked for possible intersection with the current shape. If all operations are to a single root object, then the total CPU time is equal to the area under the curve shown. Note that the curve can rise dramatically in later operations due to data handling.

Figure 3 - CPU time verses the number of operations in a Class 2 model.
In a Class 2 model like the one represented in Figure 2 above, six equal subdivision constructions and then joined at the end. The subdivision process causes each construction to require relatively equal CPU time. Since the accumulative CPU time "a" is not encountered, a large portion of the CPU time required in the Class 1 model is saved.