# amidouetude/ml-mcp-wt architecture

> Architecture diagram of the GitHub repository amidouetude/ml-mcp-wt, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-02). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/amidouetude/ml-mcp-wt
- Repository: https://github.com/amidouetude/ml-mcp-wt

## Overview

This MATLAB research pipeline studies real-time ML-surrogate MPC for wind turbines, diagnoses MATLAB predict() overhead, and supplies manual-forward bypasses plus reproducibility and sensitivity checks. The main workflow runs stages 1–6, builds/trains surrogate models, designs controllers, and evaluates closed-loop behavior. Track B adds residual-learning experiments and alternate plant integration; the repository also documents campaign provenance checks and paper outputs. Relationships below follow README statements; sampled source was limited to the Python digest reference, so finer call-level wiring is not asserted.

Main entry points: [`common/sf_gp.m`](https://github.com/amidouetude/ml-mcp-wt/blob/main/common/sf_gp.m) (Surrogate functions), [`common/stage3_run_simulation.m`](https://github.com/amidouetude/ml-mcp-wt/blob/main/common/stage3_run_simulation.m) (Closed-loop simulation), [`piste_B_nominal_correction/manual_forward_generic.m`](https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/manual_forward_generic.m) (Manual forward passes), [`run_all.m`](https://github.com/amidouetude/ml-mcp-wt/blob/main/run_all.m) (Stage runner)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_pipeline["Core pipeline"]
  node_orchestrator["Stage runner<br/>[run_all.m]"]
  node_plant["Turbine plant<br/>[wt_step.m]"]
  node_trainingdata[("Training data")]
  node_surrogatetraining["V1/V2 training<br/>[stage2_main.m]"]
  node_surrogates["Surrogate functions<br/>[sf_gp.m]"]
  node_controller["Controller design"]
  node_simulation["Closed-loop simulation"]
end

subgraph group_bypass["Bypass and residuals"]
  node_trueplant["Unmodeled plant<br/>[wt_step_true.m]"]
  node_residualdata[("Residual data")]
  node_residualtraining["Residual training"]
  node_manualforward["Manual forward passes"]
  node_residualcontroller["Residual controllers"]
end

subgraph group_verification["Verification"]
  node_domaincheck["Domain checks"]
  node_sensitivity["Sensitivity studies"]
  node_campaignaudit["Campaign verification"]
  node_digest["Manifest digest<br/>[digest_manifeste.m]"]
  node_digestreference["Digest reference"]
end

subgraph group_research["Research outputs"]
  node_montecarlo["Monte Carlo"]
  node_paper["Paper results<br/>[main.tex]"]
end

node_researcher(("Researcher"))

node_researcher -->|"runs stages"| node_orchestrator
node_orchestrator -->|"runs pipeline"| node_trainingdata
node_trainingdata -.->|"trains from"| node_surrogatetraining
node_surrogatetraining -->|"produces models"| node_surrogates
node_surrogates -.->|"supplies models"| node_controller
node_controller -.->|"evaluates closed loop"| node_simulation
node_plant -.->|"simulates turbine"| node_simulation
node_trueplant -->|"generates residual data"| node_residualdata
node_residualdata -->|"trains residual model"| node_residualtraining
node_residualtraining -->|"provides model"| node_residualcontroller
node_manualforward -->|"bypasses predict"| node_surrogates
node_manualforward -.->|"supports inference"| node_residualcontroller
node_surrogates -->|"checks operating domain"| node_domaincheck
node_simulation -.->|"tests robustness"| node_sensitivity
node_researcher -.->|"verifies campaign"| node_campaignaudit
node_digest -->|"cross-checks convention"| node_digestreference
node_simulation -->|"extends evaluation"| node_montecarlo
node_montecarlo -.->|"reports results"| node_paper

click node_orchestrator "https://github.com/amidouetude/ml-mcp-wt/blob/main/run_all.m"
click node_plant "https://github.com/amidouetude/ml-mcp-wt/blob/main/common/wt_step.m"
click node_trainingdata "https://github.com/amidouetude/ml-mcp-wt/blob/main/common/stage1_generate_data.m"
click node_surrogatetraining "https://github.com/amidouetude/ml-mcp-wt/blob/main/common/stage2_main.m"
click node_surrogates "https://github.com/amidouetude/ml-mcp-wt/blob/main/common/sf_gp.m"
click node_controller "https://github.com/amidouetude/ml-mcp-wt/blob/main/common/stage3_design_controllers.m"
click node_simulation "https://github.com/amidouetude/ml-mcp-wt/blob/main/common/stage3_run_simulation.m"
click node_trueplant "https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/wt_step_true.m"
click node_residualdata "https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/stage1_generate_residual_data.m"
click node_residualtraining "https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/stage2_train_residual.m"
click node_manualforward "https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/manual_forward_generic.m"
click node_residualcontroller "https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/stage3_design_controllers_residual.m"
click node_domaincheck "https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/check_surrogate_domains.m"
click node_sensitivity "https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/run_sensitivity_noise_delay.m"
click node_campaignaudit "https://github.com/amidouetude/ml-mcp-wt/blob/main/common/verifier_lot_campagne.m"
click node_digest "https://github.com/amidouetude/ml-mcp-wt/blob/main/common/digest_manifeste.m"
click node_digestreference "https://github.com/amidouetude/ml-mcp-wt/blob/main/common/digest_manifeste_reference.py"
click node_montecarlo "https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/run_extended_monte_carlo.m"
click node_paper "https://github.com/amidouetude/ml-mcp-wt/blob/main/paper/main.tex"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_orchestrator,node_plant,node_trainingdata,node_surrogatetraining,node_surrogates,node_controller,node_simulation toneBlue
class node_trueplant,node_residualdata,node_residualtraining,node_manualforward,node_residualcontroller toneAmber
class node_domaincheck,node_sensitivity,node_campaignaudit,node_digest,node_digestreference toneMint
class node_montecarlo,node_paper toneRose
class node_researcher toneIndigo
```

## Components

### Core pipeline

- [Stage runner](https://github.com/amidouetude/ml-mcp-wt/blob/main/run_all.m): `run_all.m`
- [Turbine plant](https://github.com/amidouetude/ml-mcp-wt/blob/main/common/wt_step.m): `common/wt_step.m`
- [Training data](https://github.com/amidouetude/ml-mcp-wt/blob/main/common/stage1_generate_data.m): `common/stage1_generate_data.m`
- [V1/V2 training](https://github.com/amidouetude/ml-mcp-wt/blob/main/common/stage2_main.m): `common/stage2_main.m`
- [Surrogate functions](https://github.com/amidouetude/ml-mcp-wt/blob/main/common/sf_gp.m): `common/sf_gp.m`
- [Controller design](https://github.com/amidouetude/ml-mcp-wt/blob/main/common/stage3_design_controllers.m): `common/stage3_design_controllers.m`
- [Closed-loop simulation](https://github.com/amidouetude/ml-mcp-wt/blob/main/common/stage3_run_simulation.m): `common/stage3_run_simulation.m`

### Bypass and residuals

- [Unmodeled plant](https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/wt_step_true.m): `piste_B_nominal_correction/wt_step_true.m`
- [Residual data](https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/stage1_generate_residual_data.m): `piste_B_nominal_correction/stage1_generate_residual_data.m`
- [Residual training](https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/stage2_train_residual.m): `piste_B_nominal_correction/stage2_train_residual.m`
- [Manual forward passes](https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/manual_forward_generic.m): `piste_B_nominal_correction/manual_forward_generic.m`
- [Residual controllers](https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/stage3_design_controllers_residual.m): `piste_B_nominal_correction/stage3_design_controllers_residual.m`

### Verification

- [Domain checks](https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/check_surrogate_domains.m): `piste_B_nominal_correction/check_surrogate_domains.m`
- [Sensitivity studies](https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/run_sensitivity_noise_delay.m): `piste_B_nominal_correction/run_sensitivity_noise_delay.m`
- [Campaign verification](https://github.com/amidouetude/ml-mcp-wt/blob/main/common/verifier_lot_campagne.m): `common/verifier_lot_campagne.m`
- [Manifest digest](https://github.com/amidouetude/ml-mcp-wt/blob/main/common/digest_manifeste.m): `common/digest_manifeste.m`
- [Digest reference](https://github.com/amidouetude/ml-mcp-wt/blob/main/common/digest_manifeste_reference.py): `common/digest_manifeste_reference.py`

### Research outputs

- [Monte Carlo](https://github.com/amidouetude/ml-mcp-wt/blob/main/piste_B_nominal_correction/run_extended_monte_carlo.m): `piste_B_nominal_correction/run_extended_monte_carlo.m`
- [Paper results](https://github.com/amidouetude/ml-mcp-wt/blob/main/paper/main.tex): `paper/main.tex`

### Other components

- **Researcher**

## Connections

- Researcher → Stage runner: runs stages (evidence: `README.md`)
- Stage runner → Training data: runs pipeline (evidence: `README.md`)
- Training data → V1/V2 training: trains from
- V1/V2 training → Surrogate functions: produces models (evidence: `README.md`)
- Surrogate functions → Controller design: supplies models
- Controller design → Closed-loop simulation: evaluates closed loop
- Turbine plant → Closed-loop simulation: simulates turbine
- Unmodeled plant → Residual data: generates residual data (evidence: `README.md`)
- Residual data → Residual training: trains residual model (evidence: `README.md`)
- Residual training → Residual controllers: provides model (evidence: `README.md`)
- Manual forward passes → Surrogate functions: bypasses predict (evidence: `README.md`)
- Manual forward passes → Residual controllers: supports inference
- Surrogate functions → Domain checks: checks operating domain (evidence: `README.md`)
- Closed-loop simulation → Sensitivity studies: tests robustness (evidence: `README.md`)
- Researcher → Campaign verification: verifies campaign
- Manifest digest → Digest reference: cross-checks convention (evidence: `common/digest_manifeste_reference.py`)
- Closed-loop simulation → Monte Carlo: extends evaluation (evidence: `README.md`)
- Monte Carlo → Paper results: reports results (evidence: `README.md`)
