pepe
72dd781dbc
Audit every document in the repository, convert the non-markdown ones into markdown reports, and split current documentation from outdated material. docs/ — 31 markdown documents in seven numbered sections. Twenty are new reports generated from .docx / .pdf / .xlsx / .mlx / .m sources that were previously unreadable in the browser and undiffable in git. Each report carries a provenance block (source path, format, MD5) and links back to its original; all 13 recorded checksums verify against the files on disk. Machine-extraction losses (PDF table column interleaving, Word OMML equations, embedded figures) are called out explicitly rather than silently smoothed over. superseded/ — outdated material with a documented reason per entry: two byte-identical ClickUp re-exports, an older revision of the BIDMC/UCSD energy-flow doc (the retained copy adds the SoC Violation Rate KPI), a duplicate of Shift input data.docx, the May 2026 simulation plan, the root PV+Battery.md now covered by a fuller report, GitHub's stock demo-repository template, and a zero-byte placeholder. Its README also records what was deliberately NOT retired and why — the "Old Frameworks" and "Old Simulations" folders hold unique Simulink revisions, and "Big Ugly Folder" holds the only copy of framework revision 1.3. Findings worth flagging, all documented in the reports: - Simulink lineage recovered from each .slx's internal coreProperties.xml revision counter. The current model is Current Framework/Bobert0206_Initial_Simulation_Framework.slx (rev 2.7); the top-level copy is rev 1.3, five revisions behind. - Simulations/Constants.m is a truncated byte-prefix of the Current Framework copy, silently missing H2_leak, H2_cap and E_H2_vol_h. - The PEM electrolyser and fuel cell are unmodified MathWorks Simscape examples still at vendor defaults; the "10x bigger" sizing TODO recorded in Constants.m was never carried out. - controller-claude.m does not compile — undefined P_Electro_max, outputs unassigned on several paths. - The specification set uses two incompatible variable naming conventions and disagrees on action-space size (5 vs 16). - MA_hourly_load.csv (13.7 MB) is the same 35,040 rows as 89993-0.parquet (2.4 MB). - Clinical data is the MIMIC-IV *demo* (ODbL, 100 patients), not full MIMIC-IV — redistributable, but the licence and citation are unrecorded. Housekeeping: untrack 21 Simulink build artefacts (slprj/, *.slxc) and add ignore rules for them. Root README rewritten around the new layout. Recruitment notes naming individual candidates are excluded from version control via .gitignore rather than committed; the generic question template is kept in docs/07-team-and-operations/. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
166 lines
No EOL
10 KiB
Markdown
166 lines
No EOL
10 KiB
Markdown
# Project explanation
|
||
|
||
**Vision**
|
||
A future powered by clean, intelligent, and hydrogen-enabled energy systems that eliminate
|
||
fossil fuel dependence, reduce emissions, and ensure uninterrupted power for all critical
|
||
services, driving a sustainable and resilient world.
|
||
**Mission**
|
||
To accelerate healthcare decarbonization by integrating solar PV, hydrogen storage, and
|
||
AI-driven energy management into hospital microgrids, ensuring resilience, efficiency, and
|
||
sustainability.
|
||
**Value**
|
||
● Reliability: Secure 24/7 power for critical healthcare services.
|
||
● Sustainability: Cut CO₂ emissions by replacing natural gas and diesel.
|
||
● Efficiency: Maximize renewable utilization and minimize waste.
|
||
● Innovation: Establish hospitals as leaders in the energy transition.
|
||
**What We Are Fighting Against**
|
||
● Dependence on fossil fuels (natural gas, diesel gensets).
|
||
● Unreliable grids prone to congestion or outages.
|
||
● High carbon emissions from healthcare infrastructure.
|
||
● Rising energy costs that strain hospital budgets.
|
||
**Purpose of the Project**
|
||
To design, implement, and optimize an AI-managed hospital microgrid integrating
|
||
solar PV, electrolyzer, hydrogen storage, fuel cell, and batteries, reducing emissions and costs
|
||
while guaranteeing resilience in life-critical settings.
|
||
**Objectives - Five departments**
|
||
1\. AI Control System and Forecasting: Develop ML models to predict PV
|
||
generation and hospital energy demand. Use reinforcement learning (RL) to schedule
|
||
electrolyzer, battery, and fuel cell operation optimally.
|
||
2\. Simulation & Validation: Model one year of operation and benchmark AI vs.
|
||
rule-based strategies.
|
||
3\. Energy Management: System architecture, efficiency of processes, energy
|
||
generation and usage, and grid congestion.
|
||
4\. Economics: Quantify cost savings, CO₂ reduction, renewable utilization, hydrogen
|
||
efficiency, and blackout resilience.
|
||
5\. Business: PR, outreach efforts, internal relations.
|
||
**Benchmarking**
|
||
We have done benchmarking while coming up with the project. From our research,
|
||
what we have gathered is that Tibo offers a SaaS Energy Management System that
|
||
connects site assets and uses predictive intelligence to optimise charging,
|
||
discharging, and load shifting for cost and CO₂ savings. They position the product for
|
||
high-intensity commercial/industrial customers and emphasize market & grid
|
||
participation. Zympler focuses on optimising distributed energy assets to tackle grid
|
||
congestion, shift demand to times of abundant renewables, and reduce customer
|
||
costs.
|
||
Both projects focus on cost/emissions/grid congestion for commercial assets.
|
||
Hospitals impose different design constraints: absolute reliability for life-critical
|
||
loads, strict safety & certification requirements for hydrogen on site, clinical
|
||
workflows that cannot be interrupted, and regulatory and procurement complexity.
|
||
The H2B2–PowiDian solution addresses hardware and rule-based scheduling; the
|
||
gap is an AI control layer designed for mission-critical, hydrogen-enabled hospital
|
||
microgrids that blends optimization with safety, explainability, and resilience.
|
||
**The main differentiators for the SHIFT project are described below:**
|
||
1\. Control logic and KPIs prioritizing clinical critical loads and patient safety (
|
||
guaranteed uptime for ICUs) over pure cost optimization. This is different
|
||
from Tibo/Zympler, which optimizes for cost and grid flexibility.
|
||
2\. Direct optimization of electrolyzer/fuel-cell hydrogen flows (production,
|
||
compression/pressure strategy, state-of-charge of H₂ tanks, thermal recovery)
|
||
together with PV and batteries.
|
||
3\. Use RL to make sequential decisions that trade off immediate costs vs.
|
||
long-term hydrogen availability and safety margins. RL can adapt to changing
|
||
patient loads, seasonal trends, or PV variance, beyond the static rule-sets.
|
||
4\. Integrated intrusion detection, sensor-validation, and safe-shutdown modes
|
||
to minimise risk from deliberate attacks or data spoofing.
|
||
5\. Built-in simulation to validate RL policies under extreme scenarios (blackouts,
|
||
high patient load, extreme weather) before deployment.
|
||
**Existing Project (H2B2–PowiDian Hydrogen Optimizer at Rijnstate Hospital)**
|
||
● Focus: Hardware optimization (electrolyzer + PV + H₂ storage + fuel cell).
|
||
● Achievements: 42,000 m³ gas savings annually (~600,000 m³ over 15 years), 2
|
||
million kg CO₂ avoided.
|
||
● Method: Rule-based scheduling with fixed operational logic.
|
||
Proposed Project (AI-Driven Microgrid)
|
||
● Goes beyond electrolyzer optimization → manages entire energy system.:::
|
||
water pump, heating with burning hydrogen
|
||
● Uses reinforcement learning for adaptive decision-making instead of static logic.
|
||
● Targets not only CO₂ and costs but also resilience & patient safety continuity.
|
||
**System Construction**
|
||
● Power2Power (P2P) System: This system uses solar energy to generate hydrogen
|
||
via an electrolyzer and then uses that hydrogen in a fuel cell to produce electricity
|
||
when solar production is insufficient, according to [hydrogentechworld.com](http://hydrogentechworld.com).
|
||
● H2B2 Electrolyzer: An electrolyzer from H2B2 converts electricity from the
|
||
hospital's solar panels into hydrogen.
|
||
● PowiDian Fuel Cell & Storage: PowiDian supplies the 100-kW fuel cell and two
|
||
tanks for storing up to 200 kg of hydrogen.
|
||
● Photovoltaic (PV) Plant: A 2,300 m² PV plant on the hospital roof provides the
|
||
initial energy for the electrolyzer.
|
||
● Battery System: For fast-response, short-term balancing.
|
||
● AI Energy Management System: Forecasts, optimizes, and adapts system
|
||
scheduling.
|
||
● Grid Connection: Supports imports/exports.
|
||
**Business Case:**
|
||
The three core principles of the business case are resilience, efficiency, and
|
||
sustainability.
|
||
The plan for the construction of the business case is based in:
|
||
a) Quantifying pilot KPIs - cost savings, CO₂ avoidance, reliability metrics
|
||
b) Translating those KPIs into investor-facing revenue models and cash flows;
|
||
c) Determine the optimal commercialization route - product sale, MaaS/PPA,
|
||
licensing, or hybrid.
|
||
The primary focus of our business and economics subteam in the early stages will be
|
||
the pilot KPIs, ensuring the validity of the product. With real pilot data, we will be
|
||
able to model revenues-one-time system sales and maintenance, recurring SaaS/PPA
|
||
fees or shared savings, and license/royalty streams, and show which path yields the
|
||
best risk/reward trade-off. This technology is what Tibo Energy has developed and
|
||
thus, will not be the focus of the project. The B&E team will run scenario analyses
|
||
and recommend the commercial model that balances impact, scale, and investor
|
||
requirements; however, the focus will be on resilience and energy security.
|
||
We envision that Team SHIFT and EIRES will be strategic partners in the
|
||
development and validation of the technology. Upon full development, Team SHIFT
|
||
will maintain commercial ownership of the AI algorithms, software, and integration
|
||
platform to enable future commercialization and scalability. At the same time, we
|
||
acknowledge the crucial contribution that EIRES brings in terms of research and
|
||
institutional credibility. In this regard, we would expect EIRES to assume a
|
||
non-commercial research license and commercial dimensions for this innovation.
|
||
**We are exploring three main routes to market:**
|
||
a) Hospital Ownership Model: The AI tool is delivered as a deployable
|
||
system for hospitals by Team SHIFT and EIRES jointly; thus, the
|
||
hospital will be able to operate its own microgrid autonomously using
|
||
our platform.
|
||
b) Service Model-Microgrid-as-a-Service (SaaS): It proposes that team
|
||
SHIFT, with EIRES, would provide system optimization on a
|
||
continuous basis as a service, with the minimum upfront investment
|
||
from the hospitals.
|
||
c) Licensing Model: Partner with existing energy solution providers,
|
||
where in Team Shift and EIRES co-develop the AI technology and
|
||
license it while scaling up.
|
||
Our immediate goal is to launch a pilot project in partnership with EIRES, using this
|
||
as a proof of concept for both the technical and economic benefits of the system.
|
||
|
||
**How AI can help:**
|
||
● Process optimization: Operating parameters are optimized by AI to boost hydrogen
|
||
generation. Inefficiencies are found by analysing electrolysis and steam methane
|
||
reforming data. ML models can optimize hydrogen yield in real-time.
|
||
● Predictive maintenance optimization: Monitoring equipment performance using AI
|
||
enhances predictive maintenance. It forecasts failures using past data. A proactive
|
||
strategy lowers downtime and maintenance expenses.
|
||
● Development catalysts: AI assists in developing hydrogen catalysts. Simulations
|
||
identify compounds that boost reaction rates. The result speeds electrolysis and
|
||
reforming catalyst development.
|
||
Hydrogen storage and distribution
|
||
● Improving storage conditions: AI optimizes hydrogen storage by analyzing pressure
|
||
and temperature. It indicates how to store compressed gas or liquid hydrogen.
|
||
● Logistics and supply chains: Logistics are improved by AI's route and storage
|
||
optimization. This finds cost-effective hydrogen distribution routes, decreasing
|
||
carbon emissions.
|
||
● Forecasting demand for hydrogen: AI analyses prior hydrogen consumption to
|
||
estimate demand. The result lets producers match production schedules to market
|
||
needs, reducing waste.
|
||
Hydrogen fuel cell and utilization
|
||
● Performance optimization: Optimization of operating parameters by AI boosts
|
||
hydrogen fuel cell performance. This analyzes data to determine optimal power
|
||
output conditions.
|
||
● Intelligent management of energy: AI analyses real-time data to regulate energy in
|
||
hydrogen vehicles. It promotes sustainable energy consumption and efficient
|
||
operation.
|
||
● Hydrogen vehicle: Hydrogen fuel cells offer a sustainable substitute for fossil fuels in
|
||
automobiles. The energy transfer between the fuel cell, battery, and motor is
|
||
regulated by AI. This enhances the efficacy and broadens the operating range of
|
||
hydrogen vehicles.
|
||
Hydrogen protection and risk prevention
|
||
● Hydrogen leak recognition: AI monitors manufacturing and storage facilities for
|
||
leakage. It detects irregularities in sensor data for prompt correction.
|
||
● Emergency response: Hydrogen operations risk assessment frameworks were
|
||
generated by AI. This identifies risks using previous data. Real-time AI insights boost
|
||
emergency response planning.
|
||
Hydrogen market analysis
|
||
● Hydrogen market forecast: Trends in hydrogen demand and pricing are predicted by
|
||
AI. This assists stakeholders in making well-informed investment decisions. |