dirC = np.zeros_like(xPhysBndC)
XbndC = np.concatenate((xPhysBndC, yPhysBndC, xNormC, yNormC, dirC),
axis=1).astype(data_type)
# boundary for x=0, include only the y direction
xPhysBndD, yPhysBndD, xNormD, yNormD = geomDomain.getUnifEdgePts(
numPtsU, numPtsV, [0, 0, 1, 0])
dirD = np.ones_like(xPhysBndD)
XbndD = np.concatenate((xPhysBndD, yPhysBndD, xNormD, yNormD, dirD),
axis=1).astype(data_type)
# concatenate all the boundaries
Xbnd = np.concatenate((XbndA0, XbndA1, XbndB0, XbndB1, XbndC, XbndD), axis=0)
# plot the collocation points
plot_pts(Xint, Xbnd[:, 0:2])
# define loading
YbndA0 = -model_data["inner_pressure"] * XbndA0[:, 2:3]
YbndA1 = -model_data["inner_pressure"] * XbndA0[:, 3:4]
YbndB0 = -model_data["outer_pressure"] * XbndB0[:, 2:3]
YbndB1 = -model_data["outer_pressure"] * XbndB0[:, 3:4]
YbndC = np.zeros_like(xPhysBndC).astype(data_type)
YbndD = np.zeros_like(xPhysBndD).astype(data_type)
Ybnd = np.concatenate((YbndA0, YbndA1, YbndB0, YbndB1, YbndC, YbndD), axis=0)
#define the model
tf.keras.backend.set_floatx(data_type)
l1 = tf.keras.layers.Dense(20, "swish")
l2 = tf.keras.layers.Dense(20, "swish")
l3 = tf.keras.layers.Dense(20, "swish")
Xint = np.concatenate((xPhys,yPhys),axis=1).astype(data_type)
Yint = np.zeros_like(xPhys).astype(data_type)
#boundary conditions at x=0
xPhysBnd, tPhysBnd, _, _ = domainGeom.getUnifEdgePts(numPtsU, numPtsV, [0,0,0,1])
Xbnd = np.concatenate((xPhysBnd, tPhysBnd), axis=1).astype(data_type)
Ybnd = np.where(tPhysBnd<=1, -np.sin(np.pi*tPhysBnd), 0).astype(data_type)
#initial conditions (displacement and velocity) for t=0
xPhysInit, tPhysInit, _, _ = domainGeom.getUnifEdgePts(numPtsU, numPtsV, [1,0,0,0])
Xinit = np.concatenate((xPhysInit, tPhysInit), axis=1).astype(data_type)
Yinit = np.zeros_like(Xinit)
#plot the collocation points
plot_pts(Xint, Xbnd)
#define the model
tf.keras.backend.set_floatx(data_type)
l1 = tf.keras.layers.Dense(30, "tanh")
l2 = tf.keras.layers.Dense(30, "tanh")
l3 = tf.keras.layers.Dense(30, "tanh")
#l4 = tf.keras.layers.Dense(20, "tanh")
#l5 = tf.keras.layers.Dense(20, "tanh")
l4 = tf.keras.layers.Dense(1, None)
train_op = tf.keras.optimizers.Adam()
train_op2 = "BFGS-B"
num_epoch = 1000
print_epoch = 100
pred_model = Wave1D([l1, l2, l3, l4], train_op, num_epoch, print_epoch)