ThermalShift

Inspiration

Cold-storage facilities need continuous cooling, but electricity prices and grid carbon intensity can vary significantly throughout the day. This creates an opportunity: when cooling demand can be shifted safely across hours, operators may be able to reduce electricity costs or the emissions associated with their energy consumption without compromising cooling constraints.

We built ThermalShift to explore this opportunity at the intersection of climate-aware optimization, energy systems, and verifiable computational results. Our goal was to turn hourly energy, tariff, and emissions data into schedules that can be compared transparently rather than relying on a single opaque recommendation.

What it does

ThermalShift is a climate-aware cooling schedule optimization and scenario comparison platform designed for cold-storage operations.

The application compares three scheduling strategies:

• Baseline: A reference schedule against which alternatives can be evaluated.
• Cost-optimized: A schedule that prioritizes lower electricity costs using hourly tariff data.
• Carbon-aware: A schedule that prioritizes lower modeled emissions using hourly grid-emissions intensity.

Users can explore the scenarios through a dashboard and import hourly data through CSV. The optimization incorporates modeled heat gains and cooling-capacity and temperature constraints. Results include comparative schedules and metrics to help users understand the trade-offs between operating cost and modeled emissions.

ThermalShift also includes report-integrity verification using cryptographic hashes. This provides a way to check whether a report matches its recorded hash. External anchoring writes are currently disabled, so the application does not claim to publish reports to a live blockchain.

How we built it

We built ThermalShift with a Python backend and a web-based interactive dashboard.

• FastAPI provides the API for optimization, CSV import, scenario comparison, and report verification.
• SciPy supports the underlying constrained optimization workflow.
• React, TypeScript, and Vite power the user interface and scenario comparison dashboard.
• CSV ingestion lets users supply hourly energy-system inputs.
• Cryptographic hashing supports report-integrity checks.
• Automated tests and API documentation help validate the implementation and make its functionality easier to inspect.

The application is deployed with a public frontend and backend API, allowing judges to explore the working application and inspect its API.

Challenges we ran into

One of the main challenges was comparing different optimization objectives without making the results misleading. A cost-minimizing schedule and a carbon-aware schedule can favor different hours, depending on the supplied electricity prices and grid-emissions profile.

We therefore implemented explicit scenario comparison rather than presenting one schedule as universally optimal. We also needed to handle infeasible scenarios carefully: when constraints cannot be satisfied, the application must not present an invalid schedule as a successful result.

Another challenge was making the system reproducible and inspectable. We added automated tests, API documentation, and report-hash verification, while keeping external anchoring writes disabled rather than implying that a blockchain integration was already operational.

Accomplishments that we're proud of

• Built and deployed a working end-to-end application with an interactive scenario comparison workflow.
• Implemented baseline, cost-optimized, and carbon-aware scheduling strategies.
• Integrated CSV import and constrained optimization into the application.
• Added handling for infeasible scenarios so invalid schedules are not reported as successful solutions.
• Implemented cryptographic report-integrity verification.
• Added automated tests and API documentation to support reproducibility and technical inspection.

What we learned

We learned that climate-aware optimization is not simply about minimizing a single number. Electricity cost, grid carbon intensity, and operational constraints represent different considerations, and their trade-offs need to be visible to the user.

We also learned the importance of distinguishing modeled results from measured real-world impact. ThermalShift's emissions estimates depend on the supplied hourly data and the assumptions of its simplified model. The current application is a prototype for evaluating scheduling strategies, not a field-validated system or a guarantee of emissions reductions.

What's next for ThermalShift

Our next steps are to improve the physical realism of the cooling model, incorporate calibrated equipment-efficiency and thermal-response data, and evaluate the optimizer against more realistic operating scenarios.

We also want to validate the approach with real cold-storage energy data, expand the analysis of cost-versus-emissions trade-offs, and implement a properly configured external anchoring workflow for report verification.

Our longer-term vision is to make climate-aware energy scheduling more transparent, testable, and practical for facilities seeking to understand how operational decisions interact with electricity costs and grid emissions.

Built With

• carbon-emissions
• climate-tech
• cryptography
• csv
• data-visualization
• energy-optimization
• fastapi
• optimization
• python
• react
• rest-api
• scipy
• sha-256
• typescript
• vite