Industries · Energy & Utilities

Energy and utilities AI built for ADNOC, DEWA, TAQA and GCC peers.

Refining predictive maintenance, grid forecasting, upstream optimisation and OGMP 2.0-aligned methane intelligence — integrated with your PI historian, AVEVA stack and upstream geoscience tools. Measured operations outcomes the COO and the Chief Sustainability Officer both accept.

Request a PI-native assessmentWhatsApp the teamDownload the Refining & Upstream Field Guide

Inference at Purdue L3.5. Models on your historian. No data leaves the country.

Refining unit silhouette at dusk with an RUL chip and a PI historian trend overlay

CDU furnace · pass 3 · tube 14

PI tag: 4FT-CDU-T14.SKIN.TEMP

Creep risk

RUL estimate

71 days

Next T&I

160 days

Inference at

Purdue L3.5

AVEVA PI trend rising on furnace skin temperature. Advisory raised to reliability lead 11 weeks ahead of the next planned shutdown.

11 wks
Lead time on a major furnace-creep event ahead of planned T&I
6%
NPT reduction on a 12-rig upstream programme
1.2 pp
Peak-day load forecast MAPE improvement on a UAE utility
100%
Process and upstream data inside UAE borders

The problem on the table

The USD 30 M digital-twin programme has not yet produced a quantified operations outcome.

The refinery has a fixed-interval inspection schedule that misses a furnace-tube creep pattern visible in 18 months of PI data. The peak-Ramadan-evening grid forecast is off by 6 percent so the operator spins up expensive peaking units. The methane LDAR programme is manual and the regulator's reporting cadence keeps tightening. Two of the three OEM-aligned partners on the panel cannot survive the cyber team's IEC 62443 review.

Digital-twin overpromise

The twin became a 3D model with no operating outcome.

Most digital-twin programmes stall at visualisation. Brocode connects the twin to a running PI AF model that the reliability team uses on shift.

OT cyber zoning

Anything that touches the historian needs IEC 62443.

Cloud-first vendors fail the cyber review. The Brocode pattern is Purdue Level 3.5 inference, with no internet egress from Level 2, documented before the SoW.

OEM software lock-in

Honeywell, Yokogawa and Emerson each push their own AI.

An OEM-aligned model works inside one control estate. Brocode is OEM-neutral and writes models that cross mixed control rooms.

Use-case map

Nine concrete AI use cases the COO and reliability lead recognise.

Across refining, upstream, midstream and grid. Each panel is in production with a GCC operator and the outcome is sign-off-ready.

Rotating-equipment remaining-useful-life

RUL models for furnace tubes, compressor trains and large pumps, trained on PI tag histories with asset-frame context. Catches creep and bearing-fault patterns 8-12 weeks ahead of the next planned T&I window.

Process & yield optimisation

APC-aware setpoint advisors coexisting with AspenTech RT-Opt and Honeywell Profit Suite. Energy-intensity targets per unit, with a documented operator handoff at every recommendation.

Drilling ROP and stuck-pipe risk

WITSML ingestion and Petrel/OpenWorks/Techlog/Kingdom feature stores. Real-time bit-wear and stuck-pipe risk across a rig fleet, with the rig-floor decision-support pattern reviewed by the drilling supervisor.

Short-term grid load forecasting

Demand forecasting with NWP integration from ECMWF and NCM-UAE, DER and EV-charging load awareness. Reduces peaker dispatch on Ramadan-evening peak days where every percentage of MAPE moves dispatch cost.

OGMP 2.0 methane reconciliation

Computer-vision pipelines for LDAR camera networks and drone-based optical gas imaging, fused with MethaneSAT and GHGSat satellite passes. Site-level reconciliation aligned to the OGMP 2.0 Gold Standard.

Distribution-network fault localisation

OSI Monarch and GE GridOS integration for sub-second fault localisation on the medium-voltage network. Reduces SAIDI in densely loaded UAE feeders during summer.

EV-charging load planning

Forecasting of EV-charging load at substation level for utilities and shopping-mall operators, including the interaction with rooftop solar self-consumption and time-of-use tariff design.

Refinery yield blending

AI-augmented blend optimisation for gasoline and diesel pools, feasibility-checked against tank inventories and the planning LP. Catches off-spec blend risk before the certificate of quality is issued.

Worker-safety vision at scale

PPE compliance, restricted-zone intrusion and dropped-object detection across remote upstream sites and refining gates, with edge inference inside plant networks and no internet egress from Level 2.

Integration mechanism

PI-native and AVEVA-native by default.

bespoke model we build for yous read the asset frame, write to coexistence layers and respect the OT zone the cyber team has drawn. No data warehouse migration; no upstream data leaving the country.

Historian & APC
OSIsoft / AVEVA PI via PI Web API and PI AF templates; coexistence with AVEVA APC, AspenTech RT-Opt and Honeywell Profit Suite. Feature stores keyed to the asset frame.
Upstream geoscience
Petrel, OpenWorks, Techlog and Kingdom integration; WITSML ingestion for daily drilling reports.
Process simulation
AVEVA Process Simulation, AVEVA APM and Honeywell UniSim hooks for the engineering-grade model alongside the data-driven one.
Grid platforms
OSI Monarch and GE GridOS for distribution; NWP ingestion from ECMWF and NCM-UAE; DER and EV load signals as first-class features.
Methane / emissions
LDAR camera networks, drone OGI passes, MethaneSAT and GHGSat satellite ingestion. Reporting layer aligned to OGMP 2.0 Gold Standard.
Sovereign deployment
G42 Cloud or an on-premise plant appliance. Purdue L3.5 inference pattern. No upstream or process data leaves the country.

Standards & posture

Anchored to the cyber and emissions standards your second meeting will ask about.

IEC 62443 (OT cyber)

Control mapping document shared under NDA. Purdue L3.5 inference pattern, no internet egress from Level 2.

ADNOC HSE & cyber standards

Brocode delivery aligned to ADNOC group cyber and digital standards; familiar with the procurement and security review path.

OGMP 2.0 Gold Standard

Methane reporting reconciliation across satellite, drone and ground LDAR with a documented evidence trail for the OGMP submission.

UAE Net Zero 2050 / NCEMA

Sovereign deployment, no upstream or process data leaving the country; outcomes mapped to the group decarbonisation reporting cycle.

Reference engagements

Four anonymised cases across refining, upstream, grid and methane.

All numbers are signed off by the engagement lead and verified against the operator's own reporting before publish.

“Furnace-tube RUL programme caught a creep event 11 weeks ahead of the next planned T&I, protected the CDU run length and gave operations a documented OEE saving the board accepted.”
GCC refining operator — rotating-equipment PdM
“ROP and stuck-pipe risk model deployed across 12 rigs cut non-productive time by 6 percent. The drilling supervisor kept the decision; the model surfaced the early warning into the daily report.”
NOC upstream — drilling optimisation
“Short-term load forecast MAPE improved by 1.2 percentage points on peak day. Peaker dispatch fell measurably across the Ramadan-evening window without a public reliability incident.”
UAE utility — grid forecasting
“Site-level methane emissions reconciled across LDAR cameras, drone OGI and satellite passes. OGMP 2.0 Gold Standard reporting accepted on the first cycle, with the evidence pack ready for audit.”
Midstream operator — methane intelligence

Vs the alternatives on your shortlist

SLB, Halliburton, hyperscaler practices, OEM digital services — where Brocode sits.

The CDO has usually been offered all four. The comparison below mirrors the question set we hear on first call.

Capability	Brocode	SLB / Halliburton / BHGE	OEM digital (ABB/Honeywell)	Hyperscaler energy
PI System / AVEVA PI native integration (AF templates)			Within own platform	Within own platform
OEM-neutral across Honeywell, Yokogawa, Emerson, ABB		Upstream-focused	Own-DCS bias	Hyperscaler-agnostic but tooling-led
Purdue L3.5 inference, no internet egress from Level 2		Cloud-first by default	Mixed	Cloud-first by default
IEC 62443 control mapping shared at first review		Available on request	Available on request	Limited
OGMP 2.0-aligned methane reconciliation				Tooling, not reporting
UAE-resident training and inference	G42 Cloud or on-prem appliance	Regional, not always UAE	Customer-defined	Public-cloud regions
Coexistence with AspenTech RT-Opt / Profit Suite			OEM-aligned	Limited

Compared with SLB DELFI, Halliburton DecisionSpace, Baker Hughes Leucipa; ABB / Honeywell Process Solutions / Emerson Impact Partners; Microsoft Energy & Resources, AWS Energy, Google Cloud Energy & Materials.

Free download

Refining & Upstream AI Field Guide

34-page practitioner's reference for ADNOC and GCC operators. Covers the PI / AVEVA APC integration pattern that survives an OT cyber review, an OGMP 2.0-aligned methane reporting blueprint, the eight failure modes of digital-twin programmes in refining and a worked rotating-equipment RUL example.

OSIsoft PI / AVEVA PI Web API and AF integration pattern
AVEVA APC and AspenTech RT-Opt coexistence
Petrel / OpenWorks / Techlog upstream workflow
Methane reconciliation across LDAR, drone OGI and satellite
Refinery yield optimisation worked example
Sample IEC 62443-aligned field deployment runbook

FAQ

The questions a Director of Digital asks before the cyber review.

If something is missing here, message us on WhatsApp or send it via the form below.

We are not replacing DELFI, DecisionSpace or Leucipa. We integrate with Petrel, OpenWorks, Techlog and Kingdom and write operator-aligned models that sit alongside the OEM stack. The contracting pattern is documented: Brocode is engaged by the operator, not by the OEM, and the IP on the models is yours under the SoW.

Talk to the engagement lead

Tell us about your historian estate and OT cyber posture.

Six short questions, then a senior energy engineer responds within one business day with a scoped opportunity map and an IEC 62443-aligned delivery sketch.

Prefer to chat?

Many operators ping the team on WhatsApp first. The message is pre-filled with the use case so you do not type the brief twice.

Continue on WhatsApp

Continue exploring