Reliadrive.ai, solve what matters
Oil · Gas · LNG · Mining · Power

How to use AI to prevent unplanned downtime in your plant

See how it works

Footage: Wikimedia Commons (CC BY / CC BY-SA) and Mixkit. Credits

Field-proven production risk management. Built by engineers. Sharpened with AI.

  • OilMMboe (million barrels of oil equivalent)
  • GasMMscf/d (million standard cubic feet per day)
  • Upstream gasTJ/d (terajoules per day)
  • LNGMtpa (million tonnes per annum)
  • LNGkt/d (kilotonnes per day)
  • PowerMW (megawatts)
  • Iron orekt/d (kilotonnes per day)
  • Bauxitekt/d (kilotonnes per day)
  • Aluminiumt/d (tonnes per day)
  • Copper orekt/d (kilotonnes per day)
  • Lithiumt/d LCE (tonnes per day, lithium carbonate equivalent)
  • Zinc concentratet/d (tonnes per day)
  • Goldoz/d (troy ounces per day)
  • Manganese orekt/d (kilotonnes per day)
  • Coalkt/d (kilotonnes per day)

Introducing Reliadrive.ai

Before: stressed operators in a plant control room
Before
After: smiling operators reviewing production risks
After

The pain

  • Downtime can cost millions of dollars
  • Restart of plant can cause other things to break costing you more
  • Teams act in silos, important priorities get lost in emails, Excel sheets, meeting minutes etc.

Introducing PaR Production at Risk Panel

  • Increase profits by reducing downtime with a globally proven production risk management system. Sharpened with AI
  • Open Reliadrive and see the production risks that matter most—in 5 seconds flat
  • Prioritisation and action engine across departments and disciplines

The pain

  • Repeat failures keep disrupting your plant
  • Costly actions do not address the root cause
  • Past lessons are not applied, forgotten and fragmented across the organisation

Introducing RCA Sherlock

  • Utilise Best Practice RCA digitised to help teams investigate equipment failures with proven frameworks, to resolve the underlying cause.
  • Platform ICE – Impact Confidence and Ease. Assess with AI whether solutions and actions will be effective.
  • Improve institutional memory: RCA are structured and stored for easy retrieval. Past failures can be prompted with AI.

The pain

  • Reactive maintenance drowning out high value PMs
  • Not knowing where to reduce cost, what maintenance works and what doesn’t
  • Recurring failures hidden in noisy history

Introducing Maintenance Intelligence

  • Target pain points in your corrective maintenance running proven frameworks like FMEA and RCM – up to 90% faster with AI
  • Cost saving quick wins – find out what maintenance is effective and what isn’t
  • Quickly gain insights into your bad actors, easily perform multivariate analysis with AI based on what’s important to you – frequency of failure, cost, downtime etc.

See how it works

Why it matters

The failure you are paying for was on a list somewhere.

You have a maintenance budget, a CMMS, inspection plans and turnaround scopes. And still, every so often, something fails that somebody already knew about. Official investigations into four costly incidents show how much is at stake.

Fire-damaged air intake and filter housing of gas turbine generator 4 at Hammerfest LNG, seen from outside
Damaged air intake of turbine generator 4 after the fire. Photo: PSA/AHS, Havtil investigation report, reused with credit under Havtil terms

Hammerfest LNG, 2020

Offline
612 days1
Capacity
6.5 bcm/yr of gas2

The regulator found the turbine intake filters that caught fire had last been changed in May 2015. A 2019 recommendation to change them was logged. “That was not done.” The plant was shut until June 2022.

Sources: Havtil investigation report (fire 28 Sep 2020); Equinor, production start-up 2 Jun 2022 (restart and capacity).

The ruptured 4-sidecut pipe on the Richmond refinery crude unit after the incident
The ruptured 4-sidecut pipe. Photo: US Chemical Safety Board, public domain

Chevron Richmond, 2012

Crude unit offline
263 days1
Capacity
245,000 bbl/d crude unit1

Over ten years, specialists recommended inspecting or upgrading a corroding line. The CSB found the replacement requests were denied at the 2007 and 2011 turnarounds. The crude unit was down for more than eight months.

Sources: US Chemical Safety Board final report (incident 6 Aug 2012, photo); Hydrocarbon Processing, 26 Apr 2013 (Chevron restart statement and crude unit capacity).

Aerial view of the mudflow path below the failed Dam I at the Córrego do Feijão iron ore mine near Brumadinho
The mudflow path below Dam I. Photo: Ibama, via Wikimedia Commons, CC BY-SA 2.0

Brumadinho iron ore mine, 2019

Lives lost
2701
Vale iron ore output
302 Mt in 2019, down 21.5%2

Tailings Dam I at Vale’s Córrego do Feijão mine failed on 25 January 2019. In 2017 Vale’s international panel of experts recommended a minimum safety factor of 1.3, meaning the dam should be 30% stronger than the forces acting on it. At 1.0 a dam is on the edge of failure. A 2018 review calculated 1.09, and the dam was still certified stable against a 1.05 threshold.

Sources: Vale, Brumadinho (lives lost); Expert Panel report, 12 Dec 2019 (failure mode); Vale Independent Advisory Committee report, Feb 2020 (safety factors); Vale, production and sales in 2019 (output).

Fire boats spraying water on the burning Deepwater Horizon drilling rig in the Gulf of Mexico
Fire boats battle the Deepwater Horizon fire. Photo: U.S. Coast Guard, public domain, via Wikimedia Commons

Deepwater Horizon, 2010

Workers killed
111
Well flowing
87 days1

The National Commission found the crew treated a negative-pressure test as a success and missed critical signs that a kick was under way. More than 4 million barrels of oil were released before the well was capped.

Sources: National Commission on the BP Deepwater Horizon Oil Spill, report to the President, Jan 2011 (blowout 20 Apr 2010, capped 15 Jul 2010).

Days offline are counted between the incident date and the restart date in the linked sources. We do not claim Reliadrive would have prevented any incident described here. Each had several causes, set out in the linked official reports.

About Us

Reliadrive celebrates the people who solve what matters most—and make the fix last.

  • The Operator who sees what others miss.
  • The Technician who solves the underlying problem.
  • The Engineer who asks the uncomfortable question.
  • The Leader who clears the way for work that matters most.

Solve What Matters.

See how it works