Portfolio / Project 02
Predictive Maintenance & Lifecycle System
A maintenance scheduling and health monitoring platform for Shoft Shipyard's autonomous vessel fleet — combining real-time sensor telemetry with mechanical engineering maintenance schedules to predict failures before they happen.
Client
Shoft Shipyard
Role
Lead Product Designer
Timeline
4 months
Platform
Desktop + Tablet
Domain
Maritime Engineering
Year
2021
The Problem
As Shoft Shipyard scaled its autonomous vessel fleet from 3 to 7 active ASVs, the maintenance team was drowning. Each vessel has 400+ monitored components across propulsion, hull integrity, electrical, and sensor systems. Maintenance schedules were tracked in spreadsheets. Parts inventory was managed by memory. Classification society compliance deadlines were discovered by accident.
The cost of this chaos was real: ASV-03 missed an Indian Register of Shipping annual survey, grounding it for 6 weeks. ASV-05 suffered an in-mission propulsion failure that could have been predicted from sensor trends visible weeks earlier. The maintenance team was reactive, not predictive.
I had already designed Shoft's inventory management system (which reduced lead times by 15%). Now the team needed me to extend that thinking to the entire vessel lifecycle — from sensor telemetry to dry-dock scheduling to classification society compliance.
Research and Discovery
Stakeholder Interviews
I conducted 14 interviews across 4 user groups over 3 weeks. Conversations were held on-site at Shoft's yard in Goa and aboard vessels during port maintenance.
Maintenance Chiefs4
"I keep it all in my head. If I get hit by a bus, nobody knows what's due."
Engineering Officers3
"Sensor data exists but nobody looks at trends — only alarms."
Logistics Coordinators4
"We order parts when something breaks, not before."
Navy Liaison Officers3
"Classification lapses ground the entire program. Non-negotiable."
Pain Point Frequency
Mapped pain points by severity and frequency across all 14 interviews.
No predictive failure visibility13/14
Compliance deadlines tracked manually12/14
Parts ordered reactively11/14
Sensor trend data unused10/14
Dry-dock scheduling conflicts9/14
No cross-vessel parts view7/14
Maintenance history not centralized7/14
Mission vs. maintenance conflicts6/14
Current Workflow Audit
Step 1
Sensor Alert
Email
Often missed
Step 2
Diagnosis
Walk to vessel
No remote data
Step 3
Parts Check
Call warehouse
No real-time stock
Step 4
Order Parts
Email to supplier
No lead times
Step 5
Schedule Repair
Whiteboard
Mission conflicts
Step 6
Compliance Log
Paper forms
Often forgotten
Design Approach
01
Predict, Don't React
Every component has a failure probability curve based on operating hours, environmental exposure, and sensor trend data. The system surfaces problems 2-4 weeks before they become critical.
02
Mission-Aware Scheduling
The system never recommends pulling a vessel during active operations without commander approval. Maintenance windows are proposed around mission schedules.
03
Supply Chain Integration
Parts inventory connects directly to supplier lead times and vessel compatibility matrices. When a component degrades, the system checks stock and initiates procurement.
04
Compliance First-Class
Classification society requirements (IRS, Lloyd's, DNV) are primary scheduling constraints. Certificate expiry drives timelines, and inspection readiness is tracked continuously.
Engineering Background
My mechanical engineering degree isn't just a line on a resume — it's the reason this system works. I understand that propulsion degradation follows a bathtub curve, that hull anode depletion accelerates in warm tropical waters, and that classification societies don't care about your mission schedule when a survey is overdue.
I built the failure probability models with Shoft's chief engineer, calibrating curves against 3 years of actual maintenance records. The component taxonomy mirrors real vessel systems — not a generic asset management hierarchy.
This project extended the inventory management system I had already designed at Shoft — the one that reduced procurement lead times by 15%.
Interactive Prototype
Explore the System
Six screens covering the complete maintenance lifecycle — from fleet-wide health monitoring to compliance tracking.
Key Design Decisions
Failure Probability Curves
Dynamic failure probability curves based on operating hours, environmental data, and sensor trends. Each component shows projected degradation over 30/60/90 days.
Impact
Predicted 4 of 5 component failures — 3 weeks average advance warning
Mission-Aware Scheduling
Unified calendar showing mission blocks, maintenance windows, and compliance deadlines. System proposes optimal slots avoiding active operations.
Impact
Zero unplanned mission interruptions in 14 months
Automated Compliance Tracking
Every IRS, Lloyd's, and DNV requirement mapped with automatic countdown alerts at 90, 60, 30, and 7 days. Checklists generated automatically.
Impact
100% classification compliance — zero missed surveys
Results
73%
reduction in unplanned downtime
15%
further lead time improvement
100%
classification compliance
3 wks
average advance warning
Designed by Rebecka Raj at Shoft Shipyard, 2021
This system remains in active use across Shoft's autonomous vessel fleet and maintenance operations.
Back to Projects