Testing the pelicun.control module

These are unit and integration tests on the control module of pelicun.

pelicun.tests.test_control.test_FEMA_P58_Assessment_central_tendencies()

Perform a loss assessment with customized inputs that reduce the dispersion of calculation parameters to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations.

pelicun.tests.test_control.test_FEMA_P58_Assessment_EDP_uncertainty_basic()

Perform a loss assessment with customized inputs that focus on testing the methods used to estimate the multivariate lognormal distribution of EDP values. Besides the fitting, this test also evaluates the propagation of EDP uncertainty through the analysis. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations.

pelicun.tests.test_control.test_FEMA_P58_Assessment_EDP_uncertainty_detection_limit()

Perform a loss assessment with customized inputs that focus on testing the methods used to estimate the multivariate lognormal distribution of EDP values. Besides the fitting, this test also evaluates the propagation of EDP uncertainty through the analysis. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. This test differs from the basic case in having unreliable EDP values above a certain limit - a typical feature of interstory drifts in dynamic simulations. Such cases should not be a problem if the limits can be estimated and they are specified as detection limits in input file.

pelicun.tests.test_control.test_FEMA_P58_Assessment_EDP_uncertainty_failed_analyses()

Perform a loss assessment with customized inputs that focus on testing the methods used to estimate the multivariate lognormal distribution of EDP values. Besides the fitting, this test also evaluates the propagation of EDP uncertainty through the analysis. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. Here we use EDP results with unique values assigned to failed analyses. In particular, PID=1.0 and PFA=100.0 are used when an analysis fails. These values shall be handled by detection limits of 10 and 100 for PID and PFA, respectively.

pelicun.tests.test_control.test_FEMA_P58_Assessment_EDP_uncertainty_3D()

.

Perform a loss assessment with customized inputs that focus on testing the methods used to estimate the multivariate lognormal distribution of EDP values. Besides the fitting, this test also evaluates the propagation of EDP uncertainty through the analysis. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. In this test we look at the propagation of EDP values provided for two different directions. (3D refers to the numerical model used for response estimation.)

pelicun.tests.test_control.test_FEMA_P58_Assessment_EDP_uncertainty_single_sample()

Perform a loss assessment with customized inputs that focus on testing the methods used to estimate the multivariate lognormal distribution of EDP values. Besides the fitting, this test also evaluates the propagation of EDP uncertainty through the analysis. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. In this test we provide only one structural response result and see if it is properly handled as a deterministic value or a random EDP using the additional sources of uncertainty.

pelicun.tests.test_control.test_FEMA_P58_Assessment_EDP_uncertainty_zero_variance()

Perform a loss assessment with customized inputs that focus on testing the methods used to estimate the multivariate lognormal distribution of EDP values. Besides the fitting, this test also evaluates the propagation of EDP uncertainty through the analysis. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. This test simulates a scenario when one of the EDPs is identical in all of the available samples. This results in zero variance in that dimension and the purpose of the test is to ensure that such cases are handled appropriately.

pelicun.tests.test_control.test_FEMA_P58_Assessment_QNT_uncertainty_independent()

Perform loss assessment with customized inputs that focus on testing the propagation of uncertainty in component quantities. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. This test assumes that component quantities are independent.

pelicun.tests.test_control.test_FEMA_P58_Assessment_QNT_uncertainty_dependencies()

Perform loss assessment with customized inputs that focus on testing the propagation of uncertainty in component quantities. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. This test checks if dependencies between component quantities are handled appropriately.

pelicun.tests.test_control.test_FEMA_P58_Assessment_FRAG_uncertainty_dependencies(dep='IND')

Perform loss assessment with customized inputs that focus on testing the propagation of uncertainty in component fragilities. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. Component fragilites are assumed independent in this test.

pelicun.tests.test_control.test_FEMA_P58_Assessment_FRAG_uncertainty_dependencies_PG()

Perform loss assessment with customized inputs that focus on testing the propagation of uncertainty in component fragilities. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. Component fragilites are assumed perfectly correlated between performance groups in this test.

pelicun.tests.test_control.test_FEMA_P58_Assessment_FRAG_uncertainty_dependencies_DIR()

Perform loss assessment with customized inputs that focus on testing the propagation of uncertainty in component fragilities. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. Component fragilites are assumed perfectly correlated between performance groups controlled by the same EDPs in identical directions in this test.

pelicun.tests.test_control.test_FEMA_P58_Assessment_FRAG_uncertainty_dependencies_LOC()

Perform loss assessment with customized inputs that focus on testing the propagation of uncertainty in component fragilities. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. Component fragilites are assumed perfectly correlated between performance groups at identical locations in this test.

pelicun.tests.test_control.test_FEMA_P58_Assessment_FRAG_uncertainty_dependencies_ATC()

Perform loss assessment with customized inputs that focus on testing the propagation of uncertainty in component fragilities. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. Component fragilites are assumed perfectly correlated within performance groups if such correlation is prescribed by ATC in the FEMA P58 document.

pelicun.tests.test_control.test_FEMA_P58_Assessment_FRAG_uncertainty_dependencies_CSG()

Perform loss assessment with customized inputs that focus on testing the propagation of uncertainty in component fragilities. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. Component fragilites are assumed perfectly correlated within a performance group in this test.

pelicun.tests.test_control.test_FEMA_P58_Assessment_FRAG_uncertainty_dependencies_DS()

Perform loss assessment with customized inputs that focus on testing the propagation of uncertainty in component fragilities. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations. Component fragilites are assumed perfectly correlated between damage states in this test.

pelicun.tests.test_control.test_FEMA_P58_Assessment_DV_uncertainty_dependencies()

Perform loss assessment with customized inputs that focus on testing the propagation of uncertainty in consequence functions and decision variables. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations.

pelicun.tests.test_control.test_FEMA_P58_Assessment_DV_uncertainty_dependencies_with_partial_DV_data()

Perform loss assessment with customized inputs that focus on testing the propagation of uncertainty in consequence functions and decision variables when not every component has injury and red tag consequences assigned to it. Dispersions in other calculation parameters are reduced to negligible levels. This allows us to test the results against pre-defined reference values in spite of the randomness involved in the calculations.