ALLSHIFT/docs/04-simulations/existing-est-model.md

Existing EST Model (vendored)

Markdown report of a vendored third-party subtree.

Location	Shift Matlab Drive/Shift Matlab Drive/Simulations/Exisiting EST Model/est-model-main/
Upstream	https://gitlab.tue.nl/p.b.r.arnaud.de.calavon/est-model.git (TU/e GitLab)
Licence	MIT — © 2023 Energy Storage and Transport
Status	Vendored — do not reorganise. SHIFT modifications live in the outer layer
Report generated	2026-07-25

Important

This subtree is not SHIFT's own work. It is an externally authored, MIT-licensed model reused as a baseline. Its internal structure — including its duplicate data files and its ~100 MATLAB Project bookkeeping XMLs — is upstream's business, not a documentation defect. Nothing inside it has been moved, renamed, or retired.

The folder name Exisiting EST Model contains a typo for Existing. It is left as-is: renaming it changes paths for no benefit.

What EST is

EST — Energy Storage and Transport. From the upstream README:

This project contains the Simulink model for the Energy Storage and Transport (EST) project. This Simulink model contains a simplified version of a real-life energy storage and transport system, which describes the flow of energy in such a system. Supporting MATLAB files are provided which can be used to predefine parameters and to post-process data into figures.

It is a TU/e teaching/reference model for generic energy storage — supply, storage, transport, demand, with dissipation coefficients on each stage. SHIFT uses it as a starting framework, not as a hydrogen model.

The double-nested structure

Exisiting EST Model/
└── est-model-main/              ← SHIFT's working layer (modifications + stock GitLab README)
    ├── README.md                ← stock GitLab template boilerplate, NOT project docs
    ├── ModelDevelopment.prj      ← MATLAB Project file (essentially empty)
    ├── Add_Mu_and_Gen_eff.m      ← SHIFT modification
    ├── Control_to_sell_and_buy.m ← SHIFT modification
    ├── injection.m               ← SHIFT modification
    ├── Injection_run.m           ← SHIFT modification
    ├── run2.m                    ← SHIFT modification
    ├── resources/project/        ← ~100 machine-generated MATLAB Project XMLs
    └── EST-model-main/          ← pristine upstream copy
        ├── README.md            ← the REAL EST documentation
        ├── LICENSE              ← MIT
        ├── EST.slx              ← the model
        ├── preprocessing.m / postprocessing.m
        ├── scripts/             ← constants.m, loadDemandData.m, loadSupplyData.m
        ├── data/                ← Team03_demand.csv, Team03_supply.csv
        ├── images/              ← README screenshots
        └── versions/            ← EST_R2021a/b, R2022a

The inner EST-model-main/ is the pristine upstream copy. It holds the LICENSE, the real README, the model, and the version ladder. The outer est-model-main/ is SHIFT's layer — five new .m files sitting beside a GitLab-template README that nobody edited.

The outer README.md is entirely GitLab's "To make it easy for you to get started with GitLab, here's a list of recommended next steps" boilerplate — checkbox lists about setting up CI/CD and inviting collaborators. It contains zero project information. Do not read it looking for EST documentation; read EST-model-main/README.md.

Upstream requirements and usage

From the real README:

• Requires MATLAB R2022b or newer with the Simulink toolbox.
• Install by cloning/downloading the entire repository and opening EST.slx. Downloading only EST.slx does not work.
• Older, untested versions are in versions/ — copy the file matching your release into the main directory and restart both MATLAB and Simulink before running.
• preprocessing.m runs automatically before the model (initFcn callback) to define parameters and read supply/demand data; postprocessing.m runs after to plot results.
• The MATLAB working directory must be the folder containing EST.slx.

Version ladder

File	MATLAB release
EST.slx	R2022b+ — current
versions/EST_R2022a.slx	R2022a
versions/EST_R2021b.slx	R2021b
versions/EST_R2021a.slx	R2021a

The versions/ files are upstream's deliberate backwards-compatibility exports, not stale drafts. They stay.

Default parameters

From preprocessing.m:

timeUnit   = 's';
supplyFile = "SolarExample_supply.csv";   supplyUnit = "kW";
demandFile = "SolarExample_demand.csv";   demandUnit = "kW";

deltat = 5*unit("min");
stopt  = min([Supply.Timeinfo.End, Demand.Timeinfo.End]);

aSupplyTransport = 0.01;   % Dissipation coefficient

Note the mismatch: preprocessing.m references SolarExample_supply.csv / SolarExample_demand.csv, but the files actually present in data/ are Team03_demand.csv and Team03_supply.csv. Running the model as shipped will fail on the missing example files unless preprocessing.m is edited or the data files renamed. This is an upstream/vendoring seam, not a SHIFT bug.

SHIFT's modifications — gravity storage, not hydrogen

The five .m files in the outer layer model a gravity storage system: electric motors lifting mass blocks up towers to store energy, releasing them to generate. This is a storage technology SHIFT evaluated alongside hydrogen; it does not appear in the current system architecture.

injection.m — the core function

function [PfromInjection, DInjection, nUse, aInjection] = injection( ...
        Psupply, massBlock, v, rho, Cd, areaBlock, n, mu, eta_motor)
    g = 9.81;                                          % Gravity [m/s²]

    P_lift = massBlock * g * v;                        % Mechanical power to lift [W]

    P_drag = 0.5 * rho * Cd * areaBlock * abs(v)^3;    % Air drag [W]
    P_fric = mu * massBlock * g * abs(v);              % Friction [W]
    Ploss  = P_drag + P_fric;

    P_required = (P_lift + Ploss) / eta_motor;         % Electrical input [W]

    aInjection     = Ploss / P_required;               % Dynamic loss coefficient
    DInjection     = aInjection * P_required;          % Power lost [W]
    PfromInjection = P_required - DInjection;          % Net stored [W]

    if Psupply == 0
        nUse = 0;
    elseif Psupply <= n * P_required
        nUse = ceil(Psupply / P_required);             % Towers needed
    else
        nUse = n;                                      % All towers active
    end
end

The interesting idea here is aInjection — replacing EST's fixed dissipation coefficient with one computed from physics at the current operating point. That is exactly the move the hydrogen models need too, and it is the reusable insight from this branch of work.

Parameter sets

Parameter	run2.m	Injection_run.m	Unit
massBlock	10,000	10,000	kg
v	2	2	m/s
rho	1.225	1.293	kg/m³
Cd	1.05	1.05	—
areaBlock	100	100	m²
n	10	10	towers
mu	0.05	0.05	—
eta_motor	0.85	0.85	—
Psupply	1e6 (1 MW)	1e6	W

run2.m is annotated "Constants from poster". The two scripts differ only in air density — 1.225 kg/m³ (ISA sea level, 15 °C) versus 1.293 kg/m³ (0 °C). Injection_run.m also notes v = 2 is "slower than extraction to reduce losses".

The other three files

File	Content	State
Add_Mu_and_Gen_eff.m	Sets mu = 0.1, eff_gen = 0.9, then calls dissipationBeta(...)	Broken — passes eta_gen (undefined; the variable set is eff_gen) and calls dissipationBeta, which is not defined anywhere in the repository
Injection_run.m	Parameter block, then calls injectionBeta(...)	Broken — injectionBeta is not defined anywhere in the repository; the defined function is injection
Control_to_sell_and_buy.m	211 bytes	Stub

Three of the five SHIFT files call functions (dissipationBeta, injectionBeta) that do not exist in this repository. Only run2.m → injection.m is a runnable pair. The missing *Beta functions were presumably a later iteration that never got committed here.

Duplicate data files

Team03_demand.csv and Team03_supply.csv each exist twice, byte-identical:

Path	MD5
EST-model-main/Team03_demand.csv	26553f17
EST-model-main/data/Team03_demand.csv	26553f17
EST-model-main/Team03_supply.csv	7ce0330e
EST-model-main/data/Team03_supply.csv	7ce0330e

0.65 MB each, so ~1.3 MB duplicated. preprocessing.m reads from the data/ directory, so the root-level copies are the redundant ones — but they are inside the vendored subtree and are left untouched. Removing them would create a divergence from upstream for a 1.3 MB saving.

resources/project/ — machine-generated

Roughly 100 XML files with opaque hash names (8d6963sHFjo3KDafHny-kBqH4Ew/HzZAGbjCGV27z4vxUS7B-4SgGnAd.xml and similar), organised into hash-named directories with paired …d.xml / …p.xml files.

This is MATLAB Project bookkeeping — the metadata store behind ModelDevelopment.prj, tracking file labels, dependencies, and project structure. It is machine-generated, unreadable, and carries no documentation value. It is not documentation and is excluded from all documentation indexes.

ModelDevelopment.prj itself is essentially empty:

<?xml version="1.0" encoding="UTF-8"?>
<MATLABProject xmlns="http://www.mathworks.com/MATLABProjectFile"
               xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" version="1.0"/>

Disposition

Everything in this subtree is kept in place:

Item	Why kept
Inner EST-model-main/	Pristine upstream, MIT-licensed — keeping it intact preserves attribution and the ability to diff against upstream
versions/EST_R202*.slx	Deliberate upstream compatibility exports, not stale drafts
Duplicate Team03_*.csv	Inside the vendored tree; divergence not worth 1.3 MB
resources/project/**	Machine-generated, but deleting it breaks the MATLAB Project
Outer stock README.md	Boilerplate, but it is upstream's file — flagged here instead of deleted
SHIFT's five .m files	Contain the reusable physics-based dissipation-coefficient idea

The one thing worth adding is a short NOTE.md in Exisiting EST Model/ recording the upstream URL, the MIT licence, and the fact that three of the five SHIFT scripts reference undefined functions — so the next person does not spend an afternoon discovering it.

Related

• Simulink Model Inventory — SHIFT's own models
• PV + Battery Simulink Model
• Work Packages — WP3, simulation environment