# Simulink Model Inventory

> **Markdown report of non-markdown source files.**
>
> | | |
> |---|---|
> | **Sources** | Every `.slx`, `.slxc` and `.mat` under `Shift Matlab Drive/Shift Matlab Drive/Simulations/` (excluding the vendored [EST model](existing-est-model.md)) |
> | **Format** | Simulink models — **binary**, not readable or diffable outside MATLAB |
> | **Status** | Current inventory |
> | **Report generated** | 2026-07-25 |

Simulink `.slx` files are opaque to everyone without MATLAB and carry no readable git history.
This inventory makes their lineage legible.

> [!IMPORTANT]
> **Filesystem timestamps are useless here.** Every file in `Shift Matlab Drive/` carries an
> mtime inside a 4-second window on `2026-06-16 10:15` — the whole tree was bulk-copied from
> MATLAB Drive in one operation.
>
> **The lineage below comes from inside the files.** An `.slx` is a ZIP archive; its
> `metadata/coreProperties.xml` records Simulink's own **revision counter**, **release
> version**, **creation date**, and **last-modified-by username**. That is authoritative
> version data, and it is what every ordering claim on this page rests on.

## The current model

**`Current Framework/Bobert0206_Initial_Simulation_Framework.slx`** — revision **2.7**,
MATLAB **R2026a**, last saved by **`robert`**.

Three independent lines of evidence agree:

1. It carries the highest Simulink revision anywhere in the tree.
2. It sits in the folder literally named `Current Framework`.
3. `Current Framework/` is the only folder that also holds the live `Constants.m` (the complete
   variant — see [below](#constantsm--the-top-level-copy-is-stale)) and the input workbook.

`Bobert0206` reads as *Bobert* (robert) + `0206` — a date-stamped personal save.

## The `Initial_Simulation_Framework` family

Nine files, **seven distinct byte-states**. All share `dcterms:created = 2025-11-30T10:58:42Z`,
so every one descends from a single original model.

| Rev | Release | Saved by | Size | MD5 | Path (under `Simulations/`) |
|:---:|---|---|---:|---|---|
| **2.7** | R2026a | `robert` | 173,799 | `ba7f9f6e` | **`Current Framework/Bobert0206_Initial_Simulation_Framework.slx`** ← current |
| 2.6 | R2026a | `chal` | 157,508 | `306c11f9` | `Current Framework/Initial_Simulation_Framework_CLEAR_CONNECTIONS.slx` |
| 2.2 | R2026a | `20243274` | 169,553 | `95d08300` | `Old Frameworks/Copy_of_Initial_Simulation_Framework.slx` |
| 2.0 | R2026a | `20243274` | 162,353 | `b766d7d6` | `Old Frameworks/Initial_Simulation_Framework.slx` |
| 2.0 | R2026a | `20243274` | 162,353 | `b766d7d6` | `Old Simulations/Initial_Simulation_Framework.slx` |
| 1.3 | R2025b | `robert` | 105,571 | `e30dbf8a` | `Initial_Simulation_Framework.slx` *(top level)* |
| 1.3 | R2025b | `robert` | 105,571 | `e30dbf8a` | `Old Simulations/Initial_Simulation_Framework.slx.r2025b` |
| 1.3 | R2025b | `20243274` | 1,033,972 | `46010d0d` | `Big Ugly Folder/Initial_Simulation_Framework.slx` |
| 1.2 | R2025b | `20243274` | 1,007,485 | `541b7087` | `Old Simulations/Initial_Simulation_Framework.slx.autosave` |

### Reconstructed history

```
rev 1.2  (autosave, 20243274, R2025b)
   │
rev 1.3  ── forks ──┬── 20243274's copy → Big Ugly Folder/        (1.03 MB)
   │                └── robert's copy   → top level + .r2025b     (106 KB)
   │
   │  [upgraded R2025b → R2026a]
rev 2.0  (20243274)  → Old Frameworks/ and Old Simulations/
rev 2.2  (20243274)  → Old Frameworks/Copy_of_...
rev 2.6  (chal)      → Current Framework/..._CLEAR_CONNECTIONS
rev 2.7  (robert)    → Current Framework/Bobert0206_...           ← CURRENT
```

### Findings

**The top-level model is *not* the newest.** `Simulations/Initial_Simulation_Framework.slx` is
byte-identical to `Old Simulations/Initial_Simulation_Framework.slx.r2025b` and is revision
**1.3** — five revisions behind current. Its prominent top-level position is misleading.

**`_CLEAR_CONNECTIONS` is a parallel branch, not a predecessor.** Revision 2.6 by `chal`, one
revision behind `robert`'s 2.7. The name suggests a signal-routing cleanup fork that was not
merged before 2.7 was saved. Whether that cleanup still needs merging is an open question for
the Simulations cluster — it is the one piece of work in this family that may have been lost.

**`Old Frameworks/` and `Old Simulations/` hold the same rev-2.0 file** (identical MD5). One of
the two folders is redundant.

**The two ~1 MB outliers** (`Big Ugly Folder` rev 1.3, and the rev 1.2 `.autosave`) are 6×
larger than every other variant and are the two oldest revisions. Likely they retain cached
simulation output that later saves discarded.

### Contributors

Four usernames appear across the tree: **`robert`**, **`chal`**, **`nallu`**, and
**`20243274`** (a TU/e student number).

## `Big Ugly Folder/` versus the two "Ready Made" folders

**21 of the 22 files** in `Big Ugly Folder/` are byte-identical to a file in
`Ready Made PEM Electrolyzer/` or `Ready Made PEM fuel cell/` — the `.ssc` sources, the `.svg`
icons, the `.slx` models, the `.slxc` caches, `PEMFuelCellSystemDriveCycle.mat`, all the
`PEM*System*.m` scripts, and all six `slprj/` cache files.

The set difference is exactly one file each way:

| Only in `Big Ugly Folder` | Only in the clean folders |
|---|---|
| `Initial_Simulation_Framework.slx` (`46010d0d`, rev 1.3, 1.03 MB) | `Ready Made PEM fuel cell/PEMElectrolysisSystem.slxc` (`f8c088e4`) |

**Verdict: `Big Ugly Folder` is the unsplit predecessor.** Someone downloaded both MathWorks
example packages into one directory alongside the team's own framework model, then later split
the vendor code into two tidy folders. Supporting evidence:

1. Its content is the **exact union** of the two clean folders, with zero drift across 21 files
   — consistent with copy-then-split, not independent work.
2. Its unique extra file is the framework at **revision 1.3 / R2025b**, the *oldest* era. A
   folder assembled later by merging the clean folders would not contain an obsolete rev-1.3
   model.
3. The stray `.slxc` is a **newer build** of the electrolysis cache (5,779 B) than the one in
   both `Big Ugly Folder` and `Ready Made PEM Electrolyzer` (5,777 B) — so the clean folders
   were opened and re-simulated *after* the split, while `Big Ugly Folder` was frozen.

**This folder cannot simply be deleted.** Its rev-1.3 `Initial_Simulation_Framework.slx`
(`46010d0d`) is unique — it exists nowhere else. Roughly 470 kB of the folder is pure
redundancy, but that one file is not. See [`superseded/README.md`](../../superseded/README.md)
for the disposition.

## The "Ready Made" folders are unmodified MathWorks code

> [!WARNING]
> **Every `.m` and `.ssc` file in `Ready Made PEM Electrolyzer/`, `Ready Made PEM fuel cell/`
> and `Big Ugly Folder/` carries a MathWorks copyright line.** Not one file lacks it, and there
> is no SHIFT attribution anywhere.

```
Electrolyzer.ssc:9                        % Copyright 2021-2024 The MathWorks, Inc.
PEMElectrolysisSystemExample.m:57         % Copyright 2021 The MathWorks, Inc.
PEMElectrolysisSystemParameters.m:6       % Copyright 2021 The MathWorks, Inc.
FuelCell.ssc:10                           % Copyright 2020-2024 The MathWorks, Inc.
PEMFuelCellSystemExample.m:64             % Copyright 2020-2021 The MathWorks, Inc.
PEMFuelCellSystemParameters.m:6           % Copyright 2020 The MathWorks, Inc.
…and 8 more Plot*.m files
```

`Constants.m` records how they were obtained:

```matlab
%openExample('simscape/PEMElectrolysisSystemExample')
%openExample('simscape/PEMFuelCellSystemExample')
```

These are the canonical Simscape shipped-example IDs, and `Constants.m` calls the electrolyser
*"Ready made electrolyser"* — the origin of the folder names.

Unlike the [EST model](existing-est-model.md), which carries its MIT LICENSE and upstream URL,
**this vendored code is marked nowhere**. A short `PROVENANCE.md` in each folder recording
"unmodified MathWorks Simscape example, obtained via `openExample(...)`" would close the gap.

See [PEM Electrolyser and Fuel Cell Models](pem-electrolyser-and-fuel-cell-models.md) for the
parameter sets and the unfinished sizing work.

## `FC_DigitalTwin_MathModel.slx`

SHIFT-authored (no MathWorks markers), separate from the Simscape package. Two revisions, same
creation date `2026-03-18T18:14:27Z`:

| Rev | Saved by | Size | MD5 | Path |
|:---:|---|---:|---|---|
| **1.2** | `robert` | 84,858 | `3b17283e` | **`Fuelcell/FC_DigitalTwin_MathModel.slx`** ← current |
| 1.1 | `nallu` | 75,673 | `d0dcd69c` | `Old Frameworks/FC_DigitalTwin_MathModel.slx` |

`Fuelcell/` holds the newer revision, so the `Old Frameworks/` label is accurate for this file.
`Fuelcell/` is a bare single-file directory — no `Constants.m`, no data, no README.

## Electrolyser models in `Old Frameworks/`

| File | Rev | Release | Saved by | Created | Size |
|---|:---:|---|---|---|---:|
| `EL20N_Electrolyser_Model.slx` | 1.3 | R2025b | `chal` | 2026-02-11 | 116,355 |
| `electrolyserNEW.slx` | 1.1 | R2025b | `chal` | **2026-05-26** | 119,836 |

Both by `chal`. `electrolyserNEW.slx` was **created three months later** but carries a *lower*
revision — it is a fresh model, not a continuation of `EL20N`. `EL20N` encodes a specific
commercial unit choice.

Both sit in `Old Frameworks/` and are superseded in practice by the MathWorks PEM package that
`Constants.m` points at. They are retired rather than deleted — `EL20N` in particular records a
real product decision worth recovering.

## `Constants.m` — the top-level copy is stale

| MD5 | Size | Path |
|---|---:|---|
| `4a4bac92` | 839 B | `Simulations/Constants.m` |
| `b3cf89c4` | 1,006 B | `Simulations/Current Framework/Constants.m` |
| `b3cf89c4` | 1,006 B | `Simulations/Old Simulations/Constants.m` |

**`Simulations/Constants.m` is a strict byte-prefix of the other two** — its first 839 bytes are
bit-for-bit identical, and it is simply missing the final 167 bytes:

```matlab
%%-------------------------------------------------------------------------
H2_leak    = 0.001;   %Fraction (%)
H2_cap     = 128;     %m3
E_H2_vol_h = 3000;    %Wh/m3
```

> [!WARNING]
> Despite occupying the most prominent position, the top-level `Constants.m` is an **older,
> truncated snapshot**. Anyone who runs it gets a workspace silently missing `H2_leak`,
> `H2_cap`, and `E_H2_vol_h`.
>
> The corroboration is decisive: `controller-claude.m`'s commented-out hydrogen block
> references `H2_leak` and `E_H2_vol_h` — variables that exist **only** in the longer version.
> The controller was written against `b3cf89c4`.

Shared content (the common 839-byte prefix):

```matlab
%Energy production
E_PV            =  460000;   %kWh
E_Grid          =  140000;   %kWh
E_ExternalGrid  =  400000;   %kWh

E_HospitalLoad  = -1000000;  %kWh
```

followed by the developer's notes, verbatim:

> `%Ready made electrolyser -`
> `%Good news: it has a solar profile.`
> `%Bad news: random constants - we want it at 128m3 hydrogen production a day`
> `%I have attempted to find volume via the volumetric flow rate, however this`
> `%uses "standard conditions". THIS BRINGS IT TO 12.8m3, JUST ADJUST (power?)`
> `%SO ITS 10 TIMES BIGGER AND WE ARE GOOD!`
> …
> `%To do: Check if PEM is ok, change parameters for electrolyzer so that`
> `%production is exactely 10times bigger, adjust fuel cell, create grid,`
> `%ensure its power output is appropriate`

**That TODO is still unresolved** — see
[PEM Electrolyser and Fuel Cell Models](pem-electrolyser-and-fuel-cell-models.md).

## `PV+Battery Simulink/`

| File | Size | Purpose |
|---|---:|---|
| `SolarPanel.slx` | 108,605 | PV model |
| `Battery.slx` | 89,343 | Battery model |
| `SolarPanel_n_Battery.slx` | 126,034 | Combined model |
| `SolarPanel.slxc` | 5,141 | Build cache |
| `slprj/` | 2,488 | Build cache |

Documented in full in [PV + Battery Simulink Model](pv-battery-simulink.md).

## Zero-byte placeholders

An exhaustive `find . -type f -size 0` over `Simulations/` returns exactly three files:

- `Simulations/Some.m`
- `Simulations/Old Simulations/Some.m`
- `Simulations/Current Framework/TEST.txt`

Empty placeholders carrying no content or structural role. Safe to delete.

## Build artefacts

`.slxc` files and `slprj/` directories are machine-generated Simulink caches — regenerated on
demand, no information content, spurious diffs.

**All 15 `slprj/` files are committed to git**, across four directories:

| Directory | Cached models | Files | Bytes |
|---|---|---:|---:|
| `Big Ugly Folder/slprj/` | `PEMElectrolysisSystem`, `PEMFuelCellSystem` | 6 | 38,176 |
| `Ready Made PEM fuel cell/slprj/` | `PEMFuelCellSystem` | 3 | 21,776 |
| `Ready Made PEM Electrolyzer/slprj/` | `PEMElectrolysisSystem` | 3 | 16,776 |
| `PV+Battery Simulink/slprj/` | `SolarPanel` | 3 | 2,488 |

Six of the fifteen are byte-identical duplicates between `Big Ugly Folder/` and the clean
folders — further corroboration of the split.

The root `.gitignore` covers `*.asv` and `*.autosave` but **has no `slprj/` rule**. That is a
genuine gap: `slprj/` is by far the more common MATLAB build artefact. Suggested addition:

```gitignore
slprj/
*.slxc
*.slx.r20*
```

`Old Simulations/Initial_Simulation_Framework.slx.autosave` (1 MB) is a MATLAB crash autosave.
It is correctly ignored by `*.autosave` and **untracked** — but still sitting in the working
tree. It is the only surviving trace of revision 1.2.

## Related

- [PEM Electrolyser and Fuel Cell Models](pem-electrolyser-and-fuel-cell-models.md) — vendor parameter sets
- [Rule-Based Controller Implementation](rule-based-controller-implementation.md) — `controller-claude.m`
- [PV + Battery Simulink Model](pv-battery-simulink.md)
- [PowerWorld Microgrid Model](powerworld-microgrid.md)
- [Existing EST Model](existing-est-model.md) — the vendored model, excluded here
- [`superseded/README.md`](../../superseded/README.md) — disposition of retired folders