Replace scaling profiler tests with more detailed convergence analysis #1701
Description
Description
In #1695 I commented out some tests that were using the scaling profiler. The intention of these tests was originally to keep track of the number of iterations that it takes to solve the model, but using the scaling profiler for this purpose was 1) performance intensive and 2) made us track a lot of additional information about autoscaling with 7 different methods. Since autoscaling isn't recommended for anything but the simplest models, those tests require a lot of effort to maintain without much benefit.
We should keep track of how many iterations it takes to solve a problem, though, but we should do that through Pyomo's IPOPT log parser, not the scaling profiler.