Open-pit mine with haul trucks at dawn

Predictive maintenance · Sub-Saharan mining

An unplanned haul truck
stoppage costs  R90,000–R180,000
per hour.

TradeCore builds predictive maintenance capability on the sensor data your operation already records. It runs on-premise, on the SCADA and historian infrastructure you have now.

2,000–20,000 tpd
Operations we build for
25–45%
Of operating cost is maintenance
0
New sensors required
On-premise
Deployment model
01 / The Position

The data is already
being collected.
Nothing is reading it.

Instrumented, not analysed

Most mid-cap operations already stream equipment sensor data through SCADA and a data historian: vibration, temperature, pressure, load, run hours. Usually that archive only gets opened after an incident to work out what went wrong, and nothing reads it continuously to see the next failure coming.

Maintenance carries the cost base

Maintenance typically eats 25–45% of a mine's operating cost. Plenty of that goes on components changed too early, or on recovering from failures that showed up in the data long before they stopped the fleet.

Prediction on existing infrastructure

TradeCore connects to the data infrastructure you already run and predicts developing failures before they reach shutdown. You don't install new sensors or send data to a cloud, and your control systems and maintenance planning stay as they are.

Maintenance engineer reviewing sensor data at an industrial control terminal

Condition data from fixed and mobile plant, read where it already lands.

02 / Engagement Model

Three stages, scoped in that order

01

Diagnostic Audit

A scoped assessment of your existing data infrastructure, tag coverage and equipment fleet. We establish where historian data is complete enough to support failure prediction, and where it is not. The output is a written readiness position and an expected-value estimate per asset class.

02

Implementation

We build and deploy the monitoring layer on your existing systems, on-premise, behind your firewall. Models are fitted to your own failure history and operating context rather than a generic fleet baseline. Existing control and maintenance workflows stay as they are.

03

Ongoing Support

A retainer relationship covering continuous monitoring, model refinement as conditions and duty cycles change, and reporting into your planning cycle. Your reliability and maintenance engineers hold the decisions; we keep the layer accurate and accountable.

03 / Operating Constraints

Built for sites where the cloud is not an option

Data sovereignty

Operational data stays on site and under your control. Nothing leaves the perimeter.

No cloud dependency

The layer runs on-premise and continues working through link outages.

No new sensors

We work with the tags your historian already records.

No control-system change

Read-only integration. Existing SCADA and planning workflows are untouched.

04 / Services

Three tiers, each scoped to the site.

Starting prices are entry points rather than fixed quotes. Scope depends on fleet size, how mature your data is and which asset classes you want covered first.

01

Diagnostic Audit

from R100,000

Typically 2–4 weeks

A data and equipment readiness assessment against your existing SCADA, historian, and fleet. We map tag coverage, data quality, failure history availability, and which asset classes can support prediction without new sensors. The report tells you what's ready, what isn't, and what implementation is likely to be worth to you.

Deliverable · Written readiness report + expected-value position by asset class

02

Implementation Sprint

from R150,000

Typically 6–12 weeks

We build and deploy the predictive monitoring system behind your firewall, on systems you already operate. Models are fitted to your operating context and failure history. Control systems and maintenance planning workflows stay intact; we add the layer that reads the data continuously and flags developing failures before they force a shutdown.

Deliverable · On-premise predictive monitoring layer on client infrastructure

03

Monthly Retainer

from R100,000 / month

Ongoing · monthly cycle

Ongoing monitoring support, model refinement as duty cycles and conditions change, and structured reporting for reliability and maintenance leadership. Your engineers keep decision authority; we keep the prediction layer accurate and in step with how the site plans its work.

Deliverable · Monitoring, model updates, and reporting into your planning cycle

Larger fleets · multi-site

For larger fleets or multi-site operations we scope each engagement on its own. Get in touch and we'll put a quote together.

Request a tailored scope
05 / About

Built for mining maintenance in Sub-Saharan Africa, and nothing else.

TradeCore is a 100% Black- and youth-owned AI company building Sentinel, a predictive maintenance platform for Sub-Saharan mining that reads the sensor data your SCADA and historian already record.

TradeCore Solutions is founder-led and based in Durban, South Africa. Sentinel is the product, and the diagnostic audit, implementation sprint and retainer are how it gets onto a site. The founders work across mining operations, industrial data systems (OT, SCADA, OPC-UA) and applied AI, and deal directly with the people who own downtime cost: chief engineers, reliability managers and plant leadership.

We keep the company narrow on purpose. TradeCore doesn't sell AI workshops or cross-industry analytics; it does predictive maintenance on the plant data infrastructure that mid-cap mines in Sub-Saharan Africa already run. When a crusher or haul fleet is down and the historian held the warning all along, knowing mining matters far more than knowing a bit of everything.

Engagements stay hands-on. From the diagnostic audit through deployment and retainer support, the founders are your technical and commercial contact, so there are no account managers between your site and the people building the models.

Operating posture

  • Founder-led · direct engagement
  • Product: Sentinel predictive maintenance
  • Ownership: 100% Black-owned · youth-owned
  • Base: Durban, South Africa
  • Theatre: Sub-Saharan mining only
  • Stack: OT / SCADA / historian + on-premise AI
Why we're different

Mining technical teams usually weigh two defaults before a specialist like TradeCore enters the conversation. Both have structural limits for a mid-cap, mixed-fleet operation.

Alternative A

Large global enterprise AI / analytics vendors

Enterprise platforms are built for multi-year programmes with heavy professional services behind them. For a mid-cap site that usually means long procurement cycles, a high minimum commitment, and deployment timelines that outrun the maintenance season you're trying to protect. TradeCore scopes smaller, deploys on your existing infrastructure and prices for a single operation rather than a corporate transformation programme.

Alternative B

Doing nothing / OEM tools alone

Equipment OEM condition tools are useful inside a single brand envelope. They rarely give a unified view across a mixed fleet where each OEM brings its own data path and its own alert logic. Relying on them alone leaves cross-fleet patterns and plant-level risk unjoined. TradeCore reads the historian and SCADA layer you already run, so prediction is not locked to one supplier's portal.

The third path is a narrow specialist that already speaks OT and failure economics, and won't water that down into a generalist AI catalogue.
06 / Insights

Notes for mining engineering leadership.

Short pieces on downtime, data readiness, deployment architecture, and maintenance economics. Full articles are on the way; for now there are titles and summaries.

Reliability
Forthcoming

Reducing haul truck downtime without new sensors

How mid-cap open-pit fleets can use existing vibration, load, and run-hour streams to catch developing failures before a truck leaves the circuit.

Data readiness
Forthcoming

Data readiness for AI in mining: what chief engineers should check first

A practical checklist for tag coverage, historian quality, and failure-history gaps before any predictive model is worth commissioning.

Architecture
Forthcoming

On-premise vs cloud AI for industrial data

Why OT networks, data sovereignty, and latency constraints push predictive maintenance onto the plant side of the firewall for many Sub-Saharan sites.

Cost
Forthcoming

The economics of predictive vs reactive maintenance

Framing unplanned downtime, early component change-out, and maintenance OpEx share so technical leadership can defend investment in prediction.

07 / Contact

Request a Diagnostic Audit conversation.

Describe the operation and the equipment classes that drive downtime cost. We reply with next steps or a clear decline if the fit is wrong.

Direct

Reach the founders

Primary purpose of this form: open a Diagnostic Audit discussion. Implementation and retainer conversations follow only after readiness is established.

The founders read and answer every enquiry themselves.

Form