Error when running grid refinement with solve_segments='forward'

[kaushikponnapalli]

Issue Type

• Bug
• Enhancement
• Docs
• Miscellaneous

Description

Using grid refinement on a problem with run_driver=True and solve_segments='forward' causes an array size mismatch error

Example

In test_aircraft_steady_flight.py changing line 105 to dm.run_problem(p, refine_iteration_limit=1) causes the test case test_ex_aircraft_steady_flight_solve to fail. The error is

File "/Users/kponnapa/codes/dymos/dymos/examples/aircraft_steady_flight/test/test_ex_aircraft_steady_flight.py", line 119, in test_ex_aircraft_steady_flight_solve p = ex_aircraft_steady_flight(optimizer='SLSQP', solve_segments='forward',

File "/Users/kponnapa/codes/dymos/dymos/examples/aircraft_steady_flight/test/test_ex_aircraft_steady_flight.py", line 105, in ex_aircraft_steady_flight dm.run_problem(p, refine_iteration_limit=1)

File "/Users/kponnapa/codes/dymos/dymos/run_problem.py", line 85, in run_problem failed = _refine_iter(problem, refine_iteration_limit, refine_method, case_prefix=case_prefix,

File "/Users/kponnapa/codes/dymos/dymos/grid_refinement/refinement.py", line 70, in _refine_iter failed = problem.run_driver(case_prefix=f'{_case_prefix}{refine_method}_{i}_')

File "/Users/kponnapa/codes/openmdao/openmdao/utils/hooks.py", line 130, in execute_hooks ret = f(*args, **kwargs)

File "/Users/kponnapa/codes/openmdao/openmdao/core/problem.py", line 856, in run_driver return self.driver.run()

File "/Users/kponnapa/codes/openmdao/openmdao/drivers/pyoptsparse_driver.py", line 321, in run coloring = self._get_coloring(run_model=not model_ran)

File "/Users/kponnapa/codes/openmdao/openmdao/utils/hooks.py", line 130, in execute_hooks ret = f(*args, **kwargs)

File "/Users/kponnapa/codes/openmdao/openmdao/core/driver.py", line 1290, in _get_coloring coloring = c_mod.dynamic_total_coloring(self, run_model=run_model,

File "/Users/kponnapa/codes/openmdao/openmdao/utils/coloring.py", line 1977, in dynamic_total_coloring coloring = compute_total_coloring(problem, num_full_jacs=num_full_jacs, tol=tol, orders=orders,

File "/Users/kponnapa/codes/openmdao/openmdao/utils/coloring.py", line 1911, in compute_total_coloring J, sparsity_info = _get_bool_total_jac(problem, num_full_jacs=num_full_jacs, tol=tol,

File "/Users/kponnapa/codes/openmdao/openmdao/utils/coloring.py", line 1649, in _get_bool_total_jac prob.run_model(reset_iter_counts=False)

File "/Users/kponnapa/codes/openmdao/openmdao/core/problem.py", line 810, in run_model self.model.run_solve_nonlinear()

File "/Users/kponnapa/codes/openmdao/openmdao/core/system.py", line 3992, in run_solve_nonlinear self._solve_nonlinear()

File "/Users/kponnapa/codes/openmdao/openmdao/core/group.py", line 2595, in _solve_nonlinear self._nonlinear_solver._solve_with_cache_check()

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/nonlinear/nonlinear_runonce.py", line 26, in _solve_with_cache_check self.solve() # don't use caching

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/nonlinear/nonlinear_runonce.py", line 45, in solve self._gs_iter()

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/solver.py", line 787, in _gs_iter subsys._solve_nonlinear()

File "/Users/kponnapa/codes/openmdao/openmdao/core/group.py", line 2595, in _solve_nonlinear self._nonlinear_solver._solve_with_cache_check()

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/solver.py", line 825, in _solve_with_cache_check self.solve()

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/solver.py", line 583, in solve raise err

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/solver.py", line 579, in solve self._solve()

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/solver.py", line 656, in _solve self._single_iteration()

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/nonlinear/newton.py", line 241, in _single_iteration self.linesearch.solve()

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/solver.py", line 583, in solve raise err

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/solver.py", line 579, in solve self._solve()

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/linesearch/backtracking.py", line 233, in _solve self._enforce_bounds(step=du, alpha=1.0)

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/linesearch/backtracking.py", line 176, in _enforce_bounds _enforce_bounds_scalar(system._outputs, step, alpha, lower, upper)

File "/Users/kponnapa/codes/openmdao/openmdao/solvers/linesearch/backtracking.py", line 568, in _enforce_bounds_scalar change_lower = 0. if lower_bounds is None else np.maximum(u_data, lower_bounds) - u_data
ValueError: operands could not be broadcast together with shapes (4319,) (3752,)

Environment

Operating System: OS X 12.4
Python environment: Anaconda Python 3.9
Packages: Dymos v1.5.1dev0, OpenMDAO v3.19.1.dev0, pyoptsparse v2.8.3