degradation model cleanup #443
Conversation
Moving old degradation branch changes to this "cleaned up" branch
Remove "replacement" strategy test because it cannot be run using open source solvers.
this should fix failing/erroring tests
There was a problem hiding this comment.
Thanks for re-initiating @rathod-b!
I checked this against the conversation threads in the previous repo and everything checks out. I caught a couple things that I tried to implement directly. Assuming I didn't break anything in the tests, the changelog needs to be updated and then this is ready to merge.
| @@ -7,6 +7,7 @@ function add_degradation_variables(m, p) | |||
| @variable(m, Eplus_sum[days] >= 0) | |||
| @variable(m, Eminus_sum[days] >= 0) | |||
| @variable(m, EFC[days] >= 0) | |||
| @variable(m, SOH[days]) | |||
| end | |||
|
|
|||
|
|
|||
There was a problem hiding this comment.
we should refactor constrain_degradation_variables() so that the constraints are generated in the set days instead of in a for loop
There was a problem hiding this comment.
this is addressed in 8563bb3, pending tests passing
|
|
||
| # test the replacement strategy | ||
| d["ElectricStorage"]["degradation"] = Dict("maintenance_strategy" => "replacement") | ||
| m = Model(optimizer_with_attributes(HiGHS.Optimizer, "output_flag" => false, "log_to_console" => false)) | ||
| set_optimizer_attribute(m, "mip_rel_gap", 0.01) | ||
| r = run_reopt(m, d) | ||
| @test occursin("are not supported by the solver", string(r["Messages"]["errors"])) || occursin("Unable to use IndicatorToMILPBridge", string(r["Messages"]["errors"])) | ||
| # test the replacement strategy ## Cannot test with open source solvers. | ||
| # d["ElectricStorage"]["degradation"] = Dict("maintenance_strategy" => "replacement") | ||
| # m = Model(optimizer_with_attributes(HiGHS.Optimizer, "output_flag" => false, "log_to_console" => false)) | ||
| # set_optimizer_attribute(m, "mip_rel_gap", 0.01) | ||
| # r = run_reopt(m, d) | ||
| # @test occursin("not supported by the solver", string(r["Messages"]["errors"])) | ||
| # #optimal SOH at end of horizon is 80\% to prevent any replacement | ||
| # @test sum(value.(m[:bmth_BkWh])) ≈ 0 atol=0.1 | ||
| # @test sum(value.(m[:dvSOHChangeTimesEnergy])) ≈ 68.48 atol=0.01 | ||
| # # @test r["ElectricStorage"]["maintenance_cost"] ≈ 2972.66 atol=0.01 | ||
| # # the maintenance_cost comes out to 3004.39 on Actions, so we test the LCC since it should match | ||
| # @test r["Financial"]["lcc"] ≈ 1.240096e7 rtol=0.01 | ||
| # @test last(value.(m[:SOH])) ≈ 66.633 rtol=0.01 | ||
| # @test r["ElectricStorage"]["size_kwh"] ≈ 83.29 rtol=0.01 |
There was a problem hiding this comment.
should we just remove this test if we can't solve it with an open-source solver?
There was a problem hiding this comment.
I thought it would be helpful to keep it here for documentation? And in case someone wants to test these locally using a different solver, they can uncomment and test?
There was a problem hiding this comment.
makes sense to me. Maybe we add a note to that effect at the start of the commented test, that if you're modifying degradation this should be tested?
|
@rathod-b I added some notes to the changelog, can you review for completeness? |
rm documentation to check failing test
…into degradation-cleanup
Moving old degradation branch changes to this "cleaned up" branch