ShelfSense — Predictive Shelf Life Engine for Food Exporters

Developed by S. M. Mahmud Iqbal
Food Engineering & Preservation Research | Neo-Brutalist Web Engine

---

Executive Summary

Developing nations lose 30% to 40% of packaged perishable foods at overseas ports and customs due to unexpected spoilage, microbial proliferation, and barrier failure during maritime transit. Traditional shelf-life validation requires 6 to 12 months of real-time stability testing or expensive commercial accelerated studies ($1,500 – $3,500 per SKU).

ShelfSense is an open food engineering web application that instantly simulates product degradation trajectories using validated Arrhenius kinetics, water activity ($A_w$) thresholds, equilibrium pH limits, and packaging gas transmission rates.

---

Key Capabilities

1. Arrhenius Kinetic Shelf Stability Engine (v3)

• Water Activity ($A_w$): Models exponential microbial inhibition and moisture sorption stress across critical boundaries (0.50 – 0.99).

• Acidification Thresholds: Validates against the international acidified food cutoff (pH $\le$ 4.60) to prevent Clostridium botulinum spore germination — with a visible warning, and the risk label can never read "Low Risk" above pH 4.6.

• Soluble Solids (°Brix): Live model input (osmotic / water-binding screening factor).

• Thermal Acceleration: True Arrhenius shift using activation energy $E_a$; $Q_{10}$ is derived from $E_a$, not hard-coded. Spoilage-mode presets (typical literature midpoints, editable in Advanced): microbial 85, oxidation 65, moisture 35, colour/flavour 105, balanced 55 kJ/mol.

• ASLT Anchor Mode: If you have a measured shelf life, enter it with its test temperature — the engine then only temperature-shifts that anchor.

• ASLT Planner: Suggests accelerated test temperatures, durations and sampling points for confirming the predicted life.

• Packaging Gas Barrier Matrix: Typical published OTR / WVTR values per format feed an explicit barrier factor (verify against your supplier's film spec — closures and seals vary):

  Packaging	OTR (cc/m²·day)	WVTR (g/m²·day)	Barrier
  Glass Jar (hermetic lug)	0.01	0.01	Excellent
  Lacquered Tin Can	0.01	0.01	Excellent
  Multi-layer Retort Pouch (Alu-foil)	0.1	0.1	Excellent
  MAP Tray	30	8	Moderate
  PET Bottle (standard)	75	18	Low

2. Formula A vs. Formula B Scenario Comparator

• Side-by-side formulation playground comparing baseline recipes against optimized reformulations.
• Dual-trajectory SVG decay curves plotted simultaneously.
• Dynamic Days Gained and percentage shelf-life extension callouts.

3. One-Click Export Screening Certificate (PDF Generator)

• Generates a screening certificate with a deterministic verification serial — the same inputs on the same day always produce the same serial (hash of inputs + date).
• The Hazard Evaluation Table (Clostridium botulinum, Staphylococcus aureus, lipid/Maillard kinetics) is computed from the actual inputs — a formulation at pH 5.5 / Aw 0.99 correctly FAILS, gets a "NOT EXPORT-CLEARED" watermark, and the certificate carries a screening-only disclaimer (not a lab ASLT result or a compliance guarantee).
• The ROI calculator's "Engineering Takeaway" is likewise generated from the actual numbers, including negative-ROI cases.
• Print-optimized (@media print) layout for sharing estimates with buyers and customs brokers.
• Digital sign-off block by S. M. Mahmud Iqbal.

4. Packaging Cost vs. Spoilage Savings Calculator (FCL Model)

• Full Container Load (20ft FCL) batch financial modeling.
• Quantifies packaging upgrade investment vs. cargo write-off risk elimination.
• Calculates Protected Cargo Profit and Net Return on Investment (ROI %).

---

Design System: Neo-Brutalism

ShelfSense is built with a high-contrast Neo-Brutalist aesthetic:

• Palette:
  • Primary: #ffe17c (Yellow) with a 32px x 32px radial dot grid (10% opacity)
  • Background: #171e19 (Charcoal)
  • Accent: #b7c6c2 (Sage)
  • UI Surfaces: #ffffff (White)
  • Text & Borders: #000000 (Pitch Black, 2px solid #000000)
• Typography:
  • Headings: Cabinet Grotesk (Extrabold 800, tight tracking -0.05em)
  • Body Text: Satoshi (Medium 500 & Bold 700)
• Hard Shadows:
  • Standard elements: box-shadow: 4px 4px 0px 0px #000000;
  • Large containers: box-shadow: 8px 8px 0px 0px #000000;
• Micro-Interactions:
  • Buttons feature physical press physics: transform: translate(4px, 4px); box-shadow: 0px 0px 0px 0px #000000; on hover.

---

Mathematical and Food Engineering Formulation

The core simulation couples the Arrhenius rate equation with moisture sorption kinetics:

$$k = A \cdot \exp\left(-\frac{E_a}{R \cdot T}\right)$$

Temperature enters through the exact Arrhenius ratio for the selected $E_a$ (with $Q_{10}$ derived as $Q_{10} = \exp\left[\frac{E_a}{R}\cdot\frac{10}{T(T+10)}\right]$):

$$\text{Shelf Life} = \text{BaseDays}_{\text{pkg}} \times f_{\text{barrier}} \times f(A_w) \times f(\text{pH}) \times f(°\text{Brix}) \times \exp\left[\frac{E_a}{R}\left(\frac{1}{T} - \frac{1}{298.15}\right)\right] \times \text{Buffer}_{\text{market}}$$

Where $f_{\text{barrier}}$ comes from the pack's OTR/WVTR relative to standard PET, weighted by the selected spoilage mode, and $f(°\text{Brix})$ is a clamped osmotic screening factor, $1 + (°\text{Brix} - 30) \times 0.003$ within $[0.90, 1.18]$.

Where:

• $f(A_w) = 0.25 + 0.75 \cdot \left[1 - \left(\frac{A_w - 0.50}{0.50}\right)^{1.85}\right]$
• $f(\text{pH}) = 1.25 - (\text{pH} - 2.0) \cdot 0.05$ (for $\text{pH} < 4.60$)
• $f(\text{pH}) = 1.00 - (\text{pH} - 4.60) \cdot 0.16$ (for $\text{pH} \ge 4.60$)

---

Getting Started

Local Setup

No build tools or heavy dependencies required. Just clone and open in any modern browser:

git clone https://github.com/SMMahmudIqbal/shelfsense.git
cd shelfsense
# Open index.html directly
start index.html # On Windows

Live Deployment

Deploy with zero configuration to Vercel:

npm install -g vercel
vercel --prod

---

Author and Attribution

Developed by S. M. Mahmud Iqbal

• GitHub: @SMMahmudIqbal
• Focus: Food Process Engineering, Preservation Science & Applied AI Systems

---

License

This project is licensed under the MIT License — see the LICENSE file for details.

---

Changelog

v3.0.0 — 15 Oct 2026

• Certificate hazards are now evaluated from the actual inputs (previously hard-coded PASS); deterministic serial; "NOT EXPORT-CLEARED" watermark on failures; screening-only disclaimer and honest stamp wording.
• °Brix is a live model input (it previously changed nothing).
• pH > 4.6 warning + risk-label cap; dynamic ROI takeaway (including negative ROI).
• Ea-based Arrhenius with spoilage-mode presets and derived Q10; real OTR/WVTR packaging data with an explicit barrier factor; per-market screening notes.
• ASLT planner (collapsed, optional) and ASLT anchor mode for measured shelf life.
• Honest hero stats (the previous "2,847 batches / 94.2% lab correlation" figures had no source and were removed).
• Meta description, Open Graph card (og-image.png) and favicon; visible keyboard focus; Escape/backdrop closes the certificate modal.
• UX kept deliberately simple: the default form gained exactly one plain-language field ("What spoils first?"); everything advanced lives behind one collapsed panel.

v2.4.1 and earlier

• Original single-file estimator: fixed Q10 = 2.0, packaging base-days lookup, static certificate table.