HAZID Assistant

Every node, every hazard, scored before the workshop opens.

Upload a P&ID page. The system walks each node, identifies credible hazards, scores severity × likelihood on the 3×3 matrix, cross-checks against safeguards visible on the drawing, and drafts recommendations citing API / IEC / NFPA codes. Engineers walk into the workshop with a populated worksheet, not a blank one.

Book Intro Call Try the Demo

SeverityFindingRecommendationConfidence

HIGH
R-201 reactor
Reactor shell overpressure with no relief device shown
Per API 521 / ASME VIII, install a PSV on R-201 sized for fire and runaway, routed to a closed flare header.
HIGH
HIGH
R-201 catalyst bed
Thermal runaway of catalyst bed with no automated trip or quench
Per IEC 61511 / ISA-84, add a SIL-rated high-T SIS that closes feed XV and opens emergency depressurisation.
HIGH
HIGH
R-201 feed isolation
No automated emergency shutdown valve on reactor feed line
Add a fail-closed ESD valve on MX-10-20-100 interlocked with PSH/TSH from PT-201 / TT-201 / TT-202.
HIGH
HIGH
E-201 effluent / CW exchanger
Cooling-water tube rupture exposes LP CW system to ~200 bara process fluid
Per API 521, install a PSV on the CW return (or apply 10/13 rule) and add H₂ / combustibles analyzer on CWR.
HIGH
HIGH
T-201 separator
Separator overpressure with no relief device shown
Install a code-stamped PSV on T-201 sized for blocked-outlet, fire and gas-blow-by, routed to flare.
HIGH
HIGH
V-201 KO drum
KO drum overpressure with no relief device shown
Per API 521, add a PSV on V-201 sized for blocked-outlet and external-fire cases, routed to flare.
HIGH
HIGH
V-201 → C-105 recycle
Liquid carry-over to recycle compressor C-105
Add LIC + automated LV on V-201 drain plus LSHH interlock to close XV on the recycle line.
HIGH
MEDIUM
Cooling-water supply header
Loss of cooling water on header CWS-40-02-100
Add an FT with low-flow alarm interlocked to feed XV-210 (close on low CW flow) and reactor blowdown.
HIGH
MEDIUM
Reactor side connections
Ambiguous service on XV-210 / HV-209 / HV-210 — quench, sample or drain unclear
Per ISA-5.1, label the side-tap line and separate the XV/HV callouts; document XV-210 actuation.
MEDIUM

7 HIGH · 2 MEDIUM (visible)

€8.379 · 88h saved·per HAZID study

How It Works

Drawing in, populated HAZID worksheet out.

Upload

Step 01 — Input

Upload a P&ID page. PDF and PNG accepted. No legend required — the reviewer falls back to ISA-5.1 standard symbology.

Identify

Step 02 — Node walk & hazard enumeration

The reviewer walks each major node on the drawing — compressors, separators, columns, exchangers, headers — and enumerates credible hazards: loss of containment, overpressure, runaway temperature, mechanical failure, operability gaps. Every hazard is anchored to a tag drawn on the page.

Score & Recommend

Step 03 — Risk matrix + actions

Each hazard is scored on the 3×3 severity × likelihood matrix, cross-checked against safeguards visible on the sheet (PSVs, alarms, trips, check valves), and paired with a recommendation citing API / IEC / NFPA codes where applicable. Output is a tabulated worksheet with confidence scores and evidence locators on every row.

Core Capabilities

What the system does with your drawings before the workshop opens.

Node-Level Hazard Enumeration

Node Walk

Walks every drawn node, not the drawing as a whole

Each compressor train, exchanger bank, separator, column, and utility header is treated as a HAZID node. Hazards are listed against the node, never as generic checklist items.

Categories aligned with industry practice

Loss of containment, overpressure, fire/explosion, toxic release, mechanical failure, over/under-temperature, operability, control & instrumentation, human factors. Each row carries a category alongside the severity score.

3×3 Severity × Likelihood Matrix

Risk Scoring

Standard 3×3 — high / medium / low

Each hazard is scored on the project's risk matrix. Severity reflects credible consequence; likelihood reflects how often the initiating event can occur given visible safeguards.

Risk computed from the matrix, not freehand

Risk = f(severity, likelihood). The reviewer never inflates the risk box without a corresponding severity/likelihood bump — the matrix is the contract.

Safeguard Cross-Check & Code-Cited Recommendations

Verified Output

Existing safeguards are tag-anchored

Safeguards listed on each row reference a tag drawn on the P&ID — PT-101 with H/L alarms, XV-210 fail-closed, LIC-303. Where no safeguard is shown on the sheet, the row reads 'None visible on this sheet.' No hallucinated PSVs, trips, or interlocks.

Recommendations cite codes

API 521 for relief sizing, API 618 for reciprocating compressors, ASME B31.3 for piping, IEC 61511 / ISA-84 for SIS, NFPA 30 for flammables, ASME Section VIII for vessels. Engineers get an actionable worksheet, not a list of platitudes.

We will populate your first HAZID for free

Send one P&ID page. Within 24 business hours we email you a sample HAZID worksheet on your drawing. Manual processing, no upsell.

Try with your drawing

The HAZID bottleneck

From Blank Worksheet to Pre-Populated Worksheet

Workshop From Scratch

6–8 engineers · 2–4 days per study

Workshop room with 6–8 engineers

expensive

Enumerate hazards from blank worksheet

slow

Score severity / likelihood by hand

repetitive

Look up code references mid-discussion

interrupts flow

Blank worksheet, every nodeDays / Weeks

With HAZID Assistant

Pre-Populated Worksheet

Same scope, fraction of the room time

Upload P&ID page

minutes

Walk every node, list every credible hazard

per page

Score 3×3 matrix, cross-check safeguards

instant

Tabulated worksheet, code-cited recommendations

immediate

Same matrix, every drawingMinutes

The Engineer Still Decides

Hazard enumeration, severity scoring, and safeguard cross-check are repeatable. The tool does those. The workshop handles what requires experience and consensus.

ROI Estimate

Estimated Workshop Hours Saved

Based on typical HAZID workshop effort: drawings × workshop days × people in the room × hours per day, reduced by the share of node enumeration / severity scoring / safeguard cross-check that the tool can pre-populate before the workshop starts.

Your Inputs

Drawings per study

12 drawings

Workshop days per drawing

0.5 days

People in the workshop room

7 people

Workshop hours per day

6 hrs

Engineer hourly rate

95 €/hr

Workshop time reduction

35 %

Live estimate

Estimated Savings

manhours saved — per HAZID study

€8.379

estimated cost savings

Baseline workshop effort252 manhours

Reduction rate35%

Hours saved88 manhours

At €95/hr€8.379

Indicative figures — actual values are determined per project during scoping.

You define safe. We map to it.

P&IDs are trade secrets. Before we talk controls, tell us what your IT, legal, and IP teams need to see. Common asks we've met: per-upload isolation, no model training, files deleted on delivery, EU-hosted processing, self-hosted deployment. The evidence below is what we can already point at.

No Model Training

Zero Retention

Uploaded drawings are never used to train models. Period.

Isolated Sessions

Per-Upload Isolation

Each upload runs in its own session. No data crosses between projects, clients, or users.

Deleted After Processing

Automatic Cleanup

Files are removed from processing servers once the worksheet is delivered. Retention window configurable.

Encrypted & Authenticated

TLS/SSL + Your SSO

TLS/SSL in transit. Authentication via your existing sign-in. Every action logged for audit.

Full security posture, NIS2 control mapping, and subprocessors at /trust.

What Changes for Your Team

Concrete differences you will notice in the first HAZID cycle.

Workshop Starts at Row 14, Not Row 1

Speed

The blank-worksheet phase of every node disappears. The team walks in to a populated draft and spends the day on judgment, not enumeration.

Same Matrix, Every Drawing

Consistency

The 3×3 risk matrix is applied identically across every page and every reviewer. No variation between facilitators or sessions.

Tag-Anchored, Not Generic

Evidence on Every Row

Every hazard names the tags involved and quotes a row/column drawing locator. Reviewers can verify the row by glancing at the P&ID, not by interpretation.

See three more▾

Nothing Skipped

Full Coverage

Workshop fatigue and time pressure skip nodes. The tool does not. Every node, every credible hazard, every visible safeguard.

Code-Cited Recommendations

Actionable Output

API 521, API 618, ASME B31.3, IEC 61511, NFPA 30 — recommendations are anchored to codes, not platitudes. Engineers get something they can put in front of a vendor.

No Fabricated Safeguards

Visibility-Bound

If a safeguard is not drawn on the sheet, the row reads 'None visible on this sheet.' No hallucinated PSVs, no invented trips. The worksheet is honest about what's on the drawing and what isn't.

What HAZID Assistant won’t do

Where the tool stops and the workshop takes over. On purpose.

Does not chair the workshopFacilitation stays human: The tool drafts the worksheet. Facilitation, consensus, and sign-off stay in the room.
No invented safeguardsDrawing-grounded only: If a PSV, trip, or alarm is not drawn on the sheet, the row reads ‘None visible on this sheet.’ No fabricated controls.
Not a HAZOP or LOPA replacementPre-population only: It pre-populates the HAZID worksheet; deeper HAZOP node-by-node deviation analysis and quantitative LOPA stay in their own studies.

Mission

The Tool Drafts. The Workshop Decides.

Most of a HAZID workshop's first day is spent enumerating credible hazards from a blank worksheet. Walking each node, listing what could go wrong, scoring severity and likelihood, looking up which API or IEC code the recommendation should cite. Work that is repeatable and verifiable. HAZID Assistant does that part: walk every drawn node, list every credible hazard, score the matrix, anchor the safeguards to tags, draft the recommendations.

What it does not do: chair the workshop, interpret operating context, weigh trade-offs, or sign off on the risk register. That stays with the team in the room.

Nusret Furkan
Co-Founder, ProcessEngineer.io

Frequently Asked Questions

PDF and PNG. One P&ID page per upload. The reviewer renders the page at high resolution and runs node-by-node hazard enumeration in a single call.

No. The reviewer falls back to ISA-5.1 standard symbology when no legend is provided. Per-row confidence drops one notch on rows that would normally have been confirmed by the legend, and a top-level analyzer note records that the run was legend-free.

The prompt mandates that every listed safeguard is anchored to a tag drawn on the page (e.g., 'PT-101..PT-107 with H/L alarms'). When no safeguard is visible on the sheet, the row reads 'None visible on this sheet.' Tag-anchored safeguards are cheap to verify because the tag must exist in the extracted instruments / equipment list.

Demo runs a standard 3×3: severity (high/medium/low) on one axis, likelihood (high/medium/low) on the other. Risk = f(severity, likelihood). For your study, you upload your corporate matrix and severity / likelihood scoring adapts to it.

Loss of containment, mechanical failure, fire/explosion, toxic release, overpressure, over/under-temperature, operability, control/instrumentation, human factors, external/environmental, other. Aligned with common HAZID guideword sets without forcing one specific framework.

API 521 (relief sizing), API 618 (reciprocating compressors), API 510 / ASME Section VIII (pressure vessels), ASME B31.3 (process piping), IEC 61511 / ISA-84 (SIS), NFPA 30 (flammable liquids), API 2000 (tank venting). Engineers get a recommendation they can hand to a vendor or specifier.

Isolated session per upload. Drawings are not used for model training. Files are deleted after the worksheet is delivered. No third-party access.

See the Worksheet on Your Own Drawings

Two ways in. Send 1 P&ID page and we email you a sample HAZID worksheet within 24h, no call required. Or book a 30-minute call and we run it through the system live.

Book intro call Try with your drawing, 24h turnaround

Or watch the interactive demo →