def test_rk_k_comp_rk4_scalar(self):
state_options = {'y': {'shape': (1, ), 'units': 'm'}}
p = om.Problem(model=om.Group())
ivc = p.model.add_subsystem('ivc',
om.IndepVarComp(),
promotes_outputs=['*'])
ivc.add_output('h', shape=(4, ), units='s')
ivc.add_output('f:y', shape=(4, 4, 1), units='m/s')
p.model.add_subsystem(
'c',
RungeKuttaKComp(num_segments=4,
method='RK4',
state_options=state_options,
time_units='s'))
p.model.connect('f:y', 'c.f:y')
p.model.connect('h', 'c.h')
p.setup(check=True, force_alloc_complex=True)
p['h'] = [0.5, 0.5, 0.5, 0.5]
p['f:y'] = np.array([[[1.50000000], [1.81250000], [1.89062500],
[2.19531250]],
[[2.17513021], [2.40641276], [2.46423340],
[2.65724691]],
[[2.63960266], [2.73700333], [2.76135349],
[2.77027941]],
[[2.75681897], [2.63352371], [2.60269990],
[2.30816892]]])
np.set_printoptions(linewidth=1024)
p.run_model()
expected = np.array([[[0.75000000], [0.90625000], [0.94531250],
[1.09765625]],
[[1.087565104166667], [1.203206380208333],
[1.232116699218750], [1.328623453776042]],
[[1.319801330566406], [1.368501663208008],
[1.380676746368408], [1.385139703750610]],
[[1.378409485022227], [1.316761856277783],
[1.301349949091673], [1.154084459568063]]])
assert_near_equal(p.get_val('c.k:y'), expected, tolerance=1.0E-9)
p.model.list_outputs(print_arrays=True)
cpd = p.check_partials(method='cs', out_stream=None)
assert_check_partials(cpd)
def test_rk_state_advance_comp_rk4_matrix(self):
num_seg = 1
num_stages = 4
state_options = {'y': {'shape': (2, 2), 'units': 'm'}}
p = Problem(model=Group())
ivc = p.model.add_subsystem('ivc',
IndepVarComp(),
promotes_outputs=['*'])
ivc.add_output('h', shape=(num_seg, ), units='s')
ivc.add_output('f:y', shape=(num_seg, num_stages, 2, 2), units='m/s')
p.model.add_subsystem(
'c',
RungeKuttaKComp(num_segments=num_seg,
method='RK4',
state_options=state_options,
time_units='s'))
p.model.connect('f:y', 'c.f:y')
p.model.connect('h', 'c.h')
p.setup(check=True, force_alloc_complex=True)
p['h'] = [0.5]
p['f:y'] = np.array([[[[1.50000000, 2.63960266],
[2.17513021, 2.75681897]],
[[1.81250000, 2.73700333],
[2.40641276, 2.63352371]],
[[1.89062500, 2.76135349],
[2.46423340, 2.60269990]],
[[2.19531250, 2.77027941],
[2.65724691, 2.30816892]]]])
p.run_model()
expected = np.array([[[[0.75000000, 1.319801330566406],
[1.087565104166667, 1.378409485022227]],
[[0.90625000, 1.368501663208008],
[1.203206380208333, 1.316761856277783]],
[[0.94531250, 1.380676746368408],
[1.232116699218750, 1.301349949091673]],
[[1.09765625, 1.385139703750610],
[1.328623453776042, 1.154084459568063]]]])
assert_rel_error(self, p.get_val('c.k:y'), expected, tolerance=1.0E-9)
cpd = p.check_partials(method='cs', out_stream=None)
assert_check_partials(cpd)