[WIP][Add]Add forward analysis of control arm example

README.md

@@ -65,6 +65,7 @@ PaddleScience 是一个基于深度学习框架 PaddlePaddle 开发的科学计
 | 受力分析 | [3D 连接件变形](https://paddlescience-docs.readthedocs.io/zh/latest/zh/examples/bracket) | 机理驱动 | MLP | 无监督学习 | [Data](https://paddle-org.bj.bcebos.com/paddlescience/datasets/bracket/bracket_dataset.tar) | [Tutorial](https://docs.nvidia.com/deeplearning/modulus/modulus-v2209/user_guide/foundational/linear_elasticity.html) |
 | 受力分析 | [结构震动模拟](https://paddlescience-docs.readthedocs.io/zh/latest/zh/examples/phylstm) | 机理驱动 | PhyLSTM | 监督学习 | [Data](https://paddle-org.bj.bcebos.com/paddlescience/datasets/PhyLSTM/data_boucwen.mat) | [Paper](https://arxiv.org/abs/2002.10253) |
 | 受力分析 | [2D 弹塑性结构](https://paddlescience-docs.readthedocs.io/zh/examples/epnn.md) | 机理驱动 | EPNN | 无监督学习 | [Train Data](https://paddle-org.bj.bcebos.com/paddlescience/datasets/epnn/dstate-16-plas.dat)<br>[Eval Data](https://paddle-org.bj.bcebos.com/paddlescience/datasets/epnn/dstress-16-plas.dat) | [Paper](https://arxiv.org/abs/2204.12088) |
+| 受力分析和逆问题 | [3D 汽车控制臂变形](https://paddlescience-docs.readthedocs.io/zh/latest/zh/examples/control_arm.md) | 机理驱动 | MLP | 无监督学习 | - | - |
 | 拓扑优化 | [2D 拓扑优化](https://paddlescience-docs.readthedocs.io/zh/latest/zh/examples/topopt.md) | 数据驱动 | TopOptNN | 监督学习 | [Data](https://paddle-org.bj.bcebos.com/paddlescience/datasets/topopt/top_dataset.h5) | [Paper](https://arxiv.org/pdf/1709.09578) |
 
 <br>

docs/index.md

@@ -60,6 +60,7 @@
 | 受力分析 | [3D 连接件变形](./zh/examples/bracket.md) | 机理驱动 | MLP | 无监督学习 | [Data](https://paddle-org.bj.bcebos.com/paddlescience/datasets/bracket/bracket_dataset.tar) | [Tutorial](https://docs.nvidia.com/deeplearning/modulus/modulus-v2209/user_guide/foundational/linear_elasticity.html) |
 | 受力分析 | [结构震动模拟](./zh/examples/phylstm.md) | 机理驱动 | PhyLSTM | 监督学习 | [Data](https://paddle-org.bj.bcebos.com/paddlescience/datasets/PhyLSTM/data_boucwen.mat) | [Paper](https://arxiv.org/abs/2002.10253) |
 | 受力分析 | [2D 弹塑性结构](./zh/examples/epnn.md) | 机理驱动 | EPNN | 无监督学习 | [Train Data](https://paddle-org.bj.bcebos.com/paddlescience/datasets/epnn/dstate-16-plas.dat)<br>[Eval Data](https://paddle-org.bj.bcebos.com/paddlescience/datasets/epnn/dstress-16-plas.dat) | [Paper](https://arxiv.org/abs/2204.12088) |
+| 受力分析和逆问题 | [3D 汽车控制臂变形](./zh/examples/control_arm.md) | 机理驱动 | MLP | 无监督学习 | - | - |
 | 拓扑优化 | [2D 拓扑优化](./zh/examples/topopt.md) | 数据驱动 | TopOptNN | 监督学习 | [Data](https://paddle-org.bj.bcebos.com/paddlescience/datasets/topopt/top_dataset.h5) | [Paper](https://arxiv.org/pdf/1709.09578) |
 
 <br>

examples/control_arm/conf/forward_analysis.yaml

@@ -0,0 +1,89 @@
+hydra:
+  run:
+    # dynamic output directory according to running time and override name
+    dir: outputs_control_arm/forward/${now:%Y-%m-%d}/${now:%H-%M-%S}/${hydra.job.override_dirname}
+  job:
+    name: ${mode} # name of logfile
+    chdir: false # keep current working direcotry unchaned
+    config:
+      override_dirname:
+        exclude_keys:
+          - TRAIN.checkpoint_path
+          - TRAIN.pretrained_model_path
+          - EVAL.pretrained_model_path
+          - mode
+          - output_dir
+          - log_freq
+  sweep:
+    # output directory for multirun
+    dir: ${hydra.run.dir}
+    subdir: ./
+
+# general settings
+mode: train # running mode: train/eval
+seed: 2023
+output_dir: ${hydra:run.dir}
+log_freq: 100
+
+# set working condition
+NU: 0.3
+E: 1
+# T: [0, 0, 0.0025]   # +Z axis
+T: [-0.0025, 0, 0]  # -X axis
+
+# set geometry file path
+GEOM_PATH: ./datasets/control_arm.stl
+
+# set geometry parameter
+CIRCLE_LEFT_CENTER_XY: [-4.4, 0]
+CIRCLE_LEFT_RADIUS: 1.65
+CIRCLE_RIGHT_CENTER_XZ: [15.8, 0]
+CIRCLE_RIGHT_RADIUS: 2.21
+
+# model settings
+MODEL:
+  disp_net:
+    input_keys: ["x", "y", "z"]
+    output_keys: ["u", "v", "w"]
+    num_layers: 6
+    hidden_size: 512
+    activation: "silu"
+    weight_norm: true
+  stress_net:
+    input_keys: ["x", "y", "z"]
+    output_keys: ["sigma_xx", "sigma_yy", "sigma_zz", "sigma_xy", "sigma_xz", "sigma_yz"]
+    num_layers: 6
+    hidden_size: 512
+    activation: "silu"
+    weight_norm: true
+
+# training settings
+TRAIN:
+  epochs: 2000
+  iters_per_epoch: 1000
+  save_freq: 100
+  eval_freq: 100
+  eval_during_train: false
+  eval_with_no_grad: false
+  lr_scheduler:
+    epochs: ${TRAIN.epochs}
+    iters_per_epoch: ${TRAIN.iters_per_epoch}
+    learning_rate: 1.0e-4
+    gamma: 0.95
+    decay_steps: 15000
+    by_epoch: false
+  batch_size:
+    arm_left: 128
+    arm_right: 256
+    arm_surface: 4096
+    arm_interior: 2048
+    visualizer_vtu: 100000
+  weight:
+    arm_right: {"u": 1, "v": 1, "w": 1}
+  pretrained_model_path: null
+  checkpoint_path: null
+
+# evaluation settings
+EVAL:
+  eval_with_no_grad: true
+  pretrained_model_path: null

examples/control_arm/conf/inverse_parameter.yaml

@@ -0,0 +1,101 @@
+hydra:
+  run:
+    # dynamic output directory according to running time and override name
+    dir: outputs_control_arm/inverse/${now:%Y-%m-%d}/${now:%H-%M-%S}/${hydra.job.override_dirname}
+  job:
+    name: ${mode} # name of logfile
+    chdir: false # keep current working direcotry unchaned
+    config:
+      override_dirname:
+        exclude_keys:
+          - TRAIN.checkpoint_path
+          - TRAIN.pretrained_model_path
+          - EVAL.pretrained_model_path
+          - mode
+          - output_dir
+          - log_freq
+  sweep:
+    # output directory for multirun
+    dir: ${hydra.run.dir}
+    subdir: ./
+
+# general settings
+mode: train # running mode: train/eval
+seed: 2023
+output_dir: ${hydra:run.dir}
+log_freq: 100
+
+# set working condition
+NU: 0.3
+E: 1
+
+# set geometry file path
+GEOM_PATH: ./datasets/control_arm.stl
+
+# set geometry parameter
+CIRCLE_LEFT_CENTER_XY: [-4.4, 0]
+CIRCLE_LEFT_RADIUS: 1.65
+CIRCLE_RIGHT_CENTER_XZ: [15.8, 0]
+CIRCLE_RIGHT_RADIUS: 2.21
+
+# model settings
+MODEL:
+  disp_net:
+    input_keys: ["x", "y", "z"]
+    output_keys: ["u", "v", "w"]
+    num_layers: 6
+    hidden_size: 512
+    activation: "silu"
+    weight_norm: true
+  stress_net:
+    input_keys: ["x", "y", "z"]
+    output_keys: ["sigma_xx", "sigma_yy", "sigma_zz", "sigma_xy", "sigma_xz", "sigma_yz"]
+    num_layers: 6
+    hidden_size: 512
+    activation: "silu"
+    weight_norm: true
+  inverse_lambda_net:
+    input_keys: ["x", "y", "z"]
+    output_keys: ["lambda_",]
+    num_layers: 6
+    hidden_size: 512
+    activation: "silu"
+    weight_norm: true
+  inverse_mu_net:
+    input_keys: ["x", "y", "z"]
+    output_keys: ["mu",]
+    num_layers: 6
+    hidden_size: 512
+    activation: "silu"
+    weight_norm: true
+
+# training settings
+TRAIN:
+  epochs: 100
+  iters_per_epoch: 100
+  save_freq: 1
+  eval_freq: 1
+  eval_during_train: true
+  eval_with_no_grad: false
+  lr_scheduler:
+    epochs: ${TRAIN.epochs}
+    iters_per_epoch: ${TRAIN.iters_per_epoch}
+    learning_rate: 1.0e-4
+    gamma: 0.95
+    decay_steps: 100
+    by_epoch: false
+  batch_size:
+    arm_interior: 2000
+    visualizer_vtu: 100000
+  pretrained_model_path: null
+  checkpoint_path: null
+
+# evaluation settings
+EVAL:
+  pretrained_model_path: null
+  eval_with_no_grad: true
+  total_size:
+    validator: 20000
+  batch_size:
+    validator: 1024
+    visualizer_vtu: 100000