ES
← Back to Portfolio
3D & Visualization August 2025 Proprietary

Crusher Liner Wear Management System

A full-stack platform for tracking and forecasting crusher liner wear from raw 3D laser scan data through to production deployment. Replaced manual caliper measurements with automated 3D point cloud analysis.

Measurement
3D point cloud vs. manual calipers
Prediction
Remaining useful life forecasting
Deployment
Desktop (offline) + Web (centralized)
Coverage
Concave and mantle wear profiles
Crusher Liner Wear Management System — Architecture
#point-cloud#3d-analysis#wear-forecasting#fastapi#nextjs#python

Business Context

Crusher liner replacement is one of the highest-cost maintenance activities in mineral processing. A single gyratory crusher liner set costs hundreds of thousands of dollars, and each change requires days of production downtime. The stakes are asymmetric — replacing too early wastes material and production time, but replacing too late risks catastrophic failure that can shut down the entire crushing circuit for weeks. The previous approach relied on manual caliper measurements taken by personnel inside the crusher during maintenance windows: slow, imprecise, dangerous, and fundamentally limited to a few cross-section points on a surface that degrades non-uniformly.

Strategic Value

The platform replaces manual caliper measurements with automated 3D point cloud analysis from laser scans (millions of points per scan), transforming wear assessment from imprecise periodic snapshots into continuous data-driven forecasting. Cylindrical coordinate transformation, campaign-based survey management, and per-zone regression models generate remaining useful life projections with confidence bounds — turning a high-stakes judgment call into a quantitative planning activity. Dual deployment (desktop via PyInstaller for offline sites, web via Next.js/FastAPI/PostgreSQL for centralized management) ensures adoption across the full spectrum of mine site connectivity.

The Challenge

Crusher liner replacement is one of the highest-cost maintenance activities. Replacing liners too early wastes material; too late risks catastrophic failure. Traditional wear estimation relied on manual calipers — slow, imprecise, and dangerous in the crushing environment.

Our Approach

Five-stage architecture: (1) Point cloud ingestion from DXF/PTS to cylindrical coordinates, (2) Campaign and survey management for wear progression tracking, (3) Wear trend modeling with remaining useful life forecasting, (4) Dual deployment — desktop for offline mine sites, web platform for centralized management, (5) Docker Swarm orchestration with Traefik and Ansible automation.

Key Performance Indicators

KPIBaselineResultImpact
Processing TimeHours (manual caliper measurements)~80% reduction (automated 3D scan)Expert time redirected to analysis
Profile CoverageSpecific cross-section cutsFull point cloud: max/mean/min profilesComplete wear characterization
Wear ProjectionExperience-based estimatesData-driven remaining-life forecastingOptimal replacement timing

Proprietary — source code not publicly available

Architecture

crusher wear system

crusher wear system

The Cost of Getting It Wrong

A single gyratory crusher liner set costs hundreds of thousands of dollars. Replacement requires days of downtime and a full maintenance mobilization. Replace too early, and you’ve wasted material and production time. Replace too late, and you risk catastrophic failure — a damaged crusher can shut down the entire processing circuit for weeks.

The previous approach: send personnel into the crusher during maintenance windows to take manual caliper measurements at specific cross-sections. Slow. Imprecise. Dangerous. And fundamentally limited — calipers measure a few points on a surface that degrades non-uniformly across millions of square millimeters.

This platform replaces that entire process with automated 3D point cloud analysis.

From Point Cloud to Prediction

Ingestion

Raw 3D laser scan files arrive as DXF or PTS — typically millions of points per scan, an unstructured cloud of spatial data. The first processing step transforms these into cylindrical coordinates (r, θ, z) aligned to the crusher’s rotational axis via least-squares fitting. Points are binned by angular sector and axial elevation, collapsing the dense cloud into interpretable radial-axial wear profiles that capture the geometry of concave and mantle surfaces.

Campaign Management

Each liner installation defines a campaign. Multiple surveys (scans) are registered within a campaign, building a time-series of wear progression. The system manages alignment corrections between scans (the scanner position isn’t identical each time), reference geometries (the original liner profile), and metadata for traceability.

Forecasting

Wear rates are computed per profile zone using regression models. The system projects remaining useful life with confidence bounds, generating recommended change dates that maintenance planners can use for scheduling — transforming a judgment call into a data-backed planning activity.

Deployment

The platform operates in two modes to cover the full spectrum of mining site connectivity:

A desktop application (Streamlit/Dash packaged with PyInstaller) serves offline mine sites where internet is unreliable or nonexistent — the geologist processes scans on a local laptop.

A web platform (Next.js frontend, FastAPI backend, PostgreSQL, Redis) provides centralized management for multi-site operations, with Docker Swarm orchestration, Traefik load balancing, and Ansible-automated provisioning.

Technology Stack

FastAPIPostgreSQLRedisCeleryNext.jsReactTypeScriptStreamlitDashPyInstallerNumPySciPyOpen3DDocker SwarmTraefikAnsible

Visual assets for this project are not publicly available.

This is a proprietary project. Source code and external resources are not publicly available.