X-Git-Url: http://git.salome-platform.org/gitweb/?a=blobdiff_plain;f=Cases%2Fcrue-ADAO-3DVAR-1parametre.py;fp=Cases%2Fcrue-ADAO-3DVAR-1parametre.py;h=373c175377887299c07a68c87938c552ed9c7ace;hb=c4e852224680b3f1c8dccdd73388430222fe79ab;hp=0000000000000000000000000000000000000000;hpb=227a599c1e22d20e960d325e93f44b2934cc40f3;p=tools%2Fadao_interface.git

diff --git a/Cases/crue-ADAO-3DVAR-1parametre.py b/Cases/crue-ADAO-3DVAR-1parametre.py
new file mode 100644
index 0000000..373c175
--- /dev/null
+++ b/Cases/crue-ADAO-3DVAR-1parametre.py
@@ -0,0 +1,75 @@
+# -*- coding: utf-8 -*-
+# Copyright (C) 2019 - Jean-Philippe Argaud
+
+import numpy as np
+import unittest
+np.set_printoptions(precision=3)
+from adao import adaoBuilder
+
+# Les entrées
+nbobs = 4
+Qobs = np.array([10.,20.,30.,40.])
+Hobs = np.array([0.19694513, 0.298513, 0.38073079, 0.45246109])
+
+# Définit les paramètres de l'AD
+KsInitial = 20.
+ZvInitial = 49.
+ZmInitial = 51.
+thetaB = [KsInitial,]
+
+def functionCrue(Q, K_s):
+    L = 5.0e3
+    B = 300.0
+    Z_v = ZvInitial
+    Z_m = ZmInitial
+    alpha = (Z_m - Z_v)/L
+    H = (Q/(K_s*B*np.sqrt(alpha)))**(3.0/5.0)
+    return H
+
+def obsFunction(theta):
+    # Evaluation de la sortie du modèle
+    K_s = float(theta)
+    sortie = np.zeros((nbobs,))
+    for i in range(nbobs):
+        sortie[i] = functionCrue(Qobs[i],K_s)
+    return sortie
+
+def obsFunctionMulti(thetas):
+    ret = []
+    for elt in thetas:
+        ret.append( np.array(obsFunction(elt)) )
+    return ret
+
+class TestRefFloodTest(unittest.TestCase):
+
+    def test0(self):
+        # define the problem bounds
+        boundsMin = [20.,]
+        boundsMax = [40.,]
+
+        # Ecart-type des paramètres
+        sigmaKs = 5.e10
+        sigmaZv = 1.
+        sigmaZm = 1.
+        sigmaTheta = [sigmaKs,]
+
+        # Ecart-type des observations
+        sigmaH = 0.5 # (m^2)
+
+        case = adaoBuilder.New()
+        case.set( 'AlgorithmParameters', Algorithm='3DVAR' )
+        case.set( 'Background',          Vector=thetaB)
+        case.set( 'BackgroundError',     DiagonalSparseMatrix=sigmaTheta )
+        case.set( 'Observation',         Vector=Hobs)
+        case.set( 'ObservationError',    ScalarSparseMatrix=sigmaH )
+        case.set( 'ObservationOperator', OneFunction= obsFunctionMulti, InputFunctionAsMulti = True)
+        case.execute()
+        res = case.get("Analysis")[-1][0]
+        self.assertAlmostEqual(res,25.,6)
+        pass
+    pass
+
+if __name__ == "__main__":
+    unittest.main()
+    pass
+