Project Governance, Technology and Future Practice

Introduction

Structural fire safety is no longer managed only through calculations made by a structural engineer and checked against a code table. Contemporary projects involve architects, fire engineers, structural engineers, contractors, authorities, facility managers, insurers, occupants and technology providers. Their decisions are connected through digital models, approval records, construction information, inspection data and operational feedback.

Project governance provides the system by which these decisions are authorised, coordinated, recorded and reviewed. Technology can improve that system, but it can also create new risks: incorrect model data, unverified artificial-intelligence outputs, cyber-security failures, excessive confidence in sensors and unclear responsibility for decisions. A competent fire-safety professional must therefore understand both the opportunities and the limits of digital practice.

This chapter explains how structural fire safety can be integrated into building information modelling (BIM), monitoring and artificial-intelligence-assisted workflows. It also examines competence, quality assurance, ethics, records, stakeholder decisions, resilience, climate risks and emerging materials.

Learning Objectives

By the end of this chapter, the learner should be able to:

1. Explain governance principles for structural fire-safety projects.
2. Identify the information required to represent fire resistance in BIM.
3. Describe suitable uses and limitations of monitoring systems and AI-assisted tools.
4. Apply basic quality-assurance, competence, ethical and record-management principles.
5. Evaluate structural fire safety in relation to resilience, climate change and emerging materials.
6. Communicate safety-critical decisions to project stakeholders.

1. Project Governance as a Fire-Safety Function

Governance is the framework through which an organisation directs a project, assigns authority, manages risk and demonstrates accountability. It is different from project management. Project management concerns planning and controlling work; governance determines who has the authority to make decisions, what evidence is required and how competing objectives are resolved.

For structural fire safety, governance should answer five questions:

• Who is responsible for each safety-critical decision?
• What performance is required, and how is it demonstrated?
• Which assumptions and interfaces could invalidate the design?
• How are changes reviewed and approved?
• What information must remain available throughout the building life cycle?

A useful governance principle is that responsibility follows decision authority. A software operator may produce a model output, but the person who accepts that output for a safety decision remains accountable for verifying its suitability.

1.1 Safety objectives and performance criteria

A project should establish a fire-safety basis of design at an early stage. It normally includes:

• evacuation and life-safety objectives;
• prevention of disproportionate collapse;
• fire resistance periods and fire scenarios;
• compartmentation and fire-separation assumptions;
• structural materials and protection systems;
• access for firefighting and emergency operations;
• inspection, maintenance and alteration requirements;
• applicable legislation, codes and approval conditions.

Prescriptive requirements, such as a specified fire-resistance period, are often appropriate for ordinary buildings. Performance-based design may instead demonstrate that structural capacity, stability and fire spread remain within acceptable limits for defined scenarios. In either approach, the assumptions must be explicit. A fire-resistance rating is not a universal property of a member; it depends on loading, restraint, section size, protection, connections, fire exposure and acceptance criteria.

1.2 Governance roles

A project governance plan should distinguish the following roles:

No role eliminates the need for coordination. For example, a structural engineer may design a protected steel column, but the protection thickness, fixing method, joint treatment and inspection regime may depend on information from a specialist contractor.

2. Information Management and BIM

BIM is not merely a three-dimensional drawing. It is a managed process for creating, exchanging and maintaining information about a built asset. Structural fire safety benefits when geometry, material properties, fire strategy, protection systems, inspection requirements and operational records are connected in a controlled information environment.

2.1 Fire-safety information in a model

A useful structural fire-safety model may contain:

• member identification and location;
• material grade, section properties and connection type;
• fire compartment and fire-cell references;
• required fire-resistance performance;
• applied protection type, thickness and installation specification;
• design fire scenario and exposure assumptions;
• load combinations or utilisation information;
• penetrations, openings and interfaces with fire stopping;
• inspection and maintenance requirements;
• product certificates, test reports and approval limitations;
• construction status and as-built verification;
• change history and responsible approver.

The model should not imply a precision that the evidence does not support. If a protection product has only been tested within a limited range of section factors, temperatures, orientations or connection conditions, those limitations should be visible in the project information.

2.2 Common data environments and information states

A controlled common data environment (CDE) should distinguish information states such as work in progress, shared for coordination, approved for construction and archived. File naming, revision codes, permissions and approval status should be defined before substantial design work begins.

A practical information workflow is:

1. Create: the originator develops information using approved methods.
2. Check: another competent person checks geometry, data and assumptions.
3. Coordinate: interfaces with architecture, services and fire strategy are reviewed.
4. Approve: the authorised person accepts the information for its intended use.
5. Issue: the approved version is transmitted to the relevant parties.
6. Record: the final and superseded versions are retained with an audit trail.

A model element should be regarded as safety-critical when an error in its geometry, material, protection, location or status could affect life safety or structural stability. Safety-critical elements deserve stronger validation and change control than purely graphical objects.

2.3 Interoperability and model limitations

Different software platforms may interpret classifications, property fields and object relationships differently. Exporting a model can remove information, alter units or break links. Before relying on an exchange, the project team should conduct a round-trip test: export the information, import it into the receiving system and compare key properties against the source.

BIM also has limitations. A model may be geometrically accurate but operationally wrong. It may show a fire door or fire stop as present when no installation evidence exists. It may contain duplicate members or outdated revisions. Digital confidence must therefore be supported by physical inspection, traceable evidence and professional judgement.

3. Monitoring, Sensing and Digital Twins

Monitoring may be used during construction, commissioning, occupation or post-fire investigation. Sensors can measure temperature, deformation, vibration, strain, smoke conditions, compartment pressure or environmental variables. A digital twin is a maintained digital representation linked to information from the physical asset; it is more than a static BIM model.

3.1 Monitoring objectives

Monitoring should begin with a decision question. Examples include:

• Is a newly installed fire-protection system curing or drying as specified?
• Has a structure experienced abnormal thermal or deformation effects?
• Is a post-fire member safe for temporary stabilisation or further assessment?
• Are environmental conditions increasing corrosion or degradation risk?
• Is a performance-based design assumption supported by operational evidence?

A sensor without a decision rule can generate data without improving safety. The monitoring plan should state the variable measured, location, sampling frequency, accuracy, calibration interval, alarm threshold, responsible person and required response.

3.2 Measurement uncertainty

A measured value is not exact. If a sensor reports temperature T_m, a simple representation is:

where T_true is the unknown true value, T_m is the indicated value and U is the stated measurement uncertainty. The relationship is meaningful only when the uncertainty is defined consistently, including calibration, resolution, environmental effects and the confidence convention used by the supplier.

Alarm thresholds should account for uncertainty and consequence. A reading just below a threshold should not automatically be treated as safe if the uncertainty range crosses that threshold. Sensor failure, drift, loss of power and communication failure must also be considered.

3.3 Data governance

Monitoring data should have an owner and a retention period. Data records should identify the sensor, location, time standard, calibration status, software version and any missing or corrected values. Access should be controlled because building data can reveal security-sensitive information.

Monitoring supports engineering judgement; it does not replace inspection or analysis. A sensor located away from a connection, void or unprotected interface may fail to detect the most important defect.

4. Artificial Intelligence and Responsible Digital Assistance

AI-assisted tools can classify images, detect anomalies, search technical documents, estimate patterns and help organise inspection records. They may assist with option generation or prioritisation, but safety-critical acceptance requires human review by a suitably competent professional.

4.1 Appropriate and inappropriate uses

Appropriate uses include:

• identifying possible missing fire-stopping images for human review;
• prioritising inspections using documented risk indicators;
• comparing revisions and highlighting changed objects;
• extracting requirements from a large technical document for checking;
• supporting scenario analysis where inputs and limitations are understood.

Inappropriate uses include:

• accepting an AI-generated structural design without verification;
• treating a language model's citation as evidence without checking the source;
• allowing an opaque classification system to decide that a defect is safe;
• using training data from one building type without establishing applicability;
• concealing uncertainty or human intervention from the client or authority.

4.2 AI assurance questions

Before deploying an AI tool, ask:

1. What decision is the tool supporting?
2. What data trained or configured it, and are those data relevant?
3. What is the measured error rate for the intended application?
4. How are false negatives treated, particularly when missed defects could cause fatalities?
5. Can results be explained, reproduced and audited?
6. Who approves the output and who remains accountable?
7. What happens when the system is unavailable or produces an uncertain result?

For image detection, false negatives are often more dangerous than false positives. However, reducing false negatives may create many false alarms and overwhelm inspectors. The acceptable balance depends on the consequence, inspection capacity and the possibility of independent verification.

5. Competence, Quality Assurance and Change Control

Competence combines knowledge, practical skill, experience, judgement and the ability to recognise personal limitations. A person may be competent to operate BIM software but not competent to approve a fire-resistance assessment. Competence should therefore be defined against the task, not merely against a job title.

5.1 Competence management

A project competence matrix can record required capabilities, named personnel, evidence of competence, supervision arrangements and gaps requiring training or specialist appointment. Relevant capabilities may include:

• structural behaviour at elevated temperature;
• fire dynamics and scenario definition;
• fire-protection products and installation;
• BIM information management;
• inspection and non-destructive testing;
• software verification and validation;
• regulatory interpretation;
• communication of residual risk.

Where a task exceeds a person's competence, escalation is a professional responsibility, not a failure.

5.2 Quality-assurance principles

Quality assurance is planned prevention; quality control is checking the completed or developing work. A proportionate quality plan should identify hold points, witness points, checking levels, acceptance criteria and records.

Typical checks include:

• independent review of critical calculations;
• verification of units, load paths and boundary conditions;
• comparison of model geometry with drawings and site conditions;
• confirmation that protection products match the tested or assessed system;
• inspection of penetrations, joints and interfaces;
• review of deviations and non-conformances;
• confirmation that as-built information reflects installed work.

The level of independent checking should be related to consequence and complexity. Checking is not simply repeating a calculation; it should challenge assumptions, applicability and interfaces.

5.3 Change control

Changes can arise from value engineering, late services coordination, material substitution, construction sequencing or refurbishment. A change-control request should describe the proposed change, reason, affected elements, new hazards, evidence, reviewers, decision and residual risk.

A change that appears minor geometrically may be major in fire performance. For example, a new opening through a compartment wall can affect fire spread, smoke movement and structural protection. No change should be approved solely because it fits the model.

6. Ethics, Transparency and Records

Fire-safety decisions involve public safety, professional trust and sometimes uncertainty. Ethical practice requires honesty about assumptions, limitations and residual risk. It also requires resistance to commercial pressure when a proposed shortcut could reduce safety without adequate evidence.

Important ethical duties include:

• prioritising life safety and legal compliance;
• presenting evidence accurately and without selective omission;
• declaring conflicts of interest;
• protecting confidential and personal information;
• avoiding unsupported claims of compliance;
• ensuring that automated tools do not discriminate or conceal uncertainty;
• retaining records sufficient for future investigation.

6.1 The safety case and decision record

For complex or performance-based projects, a safety case can provide a structured argument supported by evidence. A simplified argument chain is:

Claim: the structure will maintain required stability for the defined fire scenarios.

Argument: the structural system, protection, compartmentation, inspection and emergency assumptions work together to support the claim.

Evidence: calculations, test results, assessments, drawings, inspection records, commissioning data and approved deviations.

A decision record should state the decision, options considered, assumptions, evidence, participants, date, approval authority and conditions. Records are essential because buildings outlive project teams.

7. Resilience and Climate Risk

Traditional fire design often focuses on a design fire and a defined resistance period. Resilience broadens the question: can the building anticipate, withstand, recover from and adapt to disruptive events?

Climate change may affect structural fire safety through:

• hotter and drier conditions that increase wildfire exposure;
• more extreme wind affecting ember transport and façade ignition;
• flooding that damages electrical, fire-alarm and access systems;
• heat and humidity accelerating material degradation;
• power interruptions affecting detection, suppression and smoke control;
• compound events, such as wildfire followed by loss of utilities.

A risk assessment should consider hazard, exposure, vulnerability and consequence. A simple qualitative model is:

This is a conceptual risk relationship, not a universal numerical law. It is useful for ranking risks when the scales are defined consistently, but it does not capture every dependency or uncertainty. A low-likelihood event with catastrophic consequence may require controls even when its product appears modest.

Resilient design may include protected utilities, alternative access routes, passive fire protection that remains effective during power loss, water-storage arrangements, landscape management, compartmentation and recovery plans. Adaptability should be considered: future changes in occupancy, energy systems, façade materials or internal layouts may alter fire risk.

8. Emerging Materials and Systems

New materials can reduce embodied carbon or improve construction efficiency, but their fire behaviour must be evaluated as systems rather than by marketing category. Relevant issues include ignition, heat release, smoke and toxicity, charring or melting, connection behaviour, delamination, moisture effects and inspection difficulty.

8.1 Mass timber and bio-based materials

Mass-timber structures may develop protective char layers, but charring rates, adhesive behaviour, exposed surfaces, connections and concealed cavities require project-specific assessment. The presence of timber in a protected assembly does not automatically establish acceptable performance. Design must address structural capacity, fire spread and the possibility of continued burning after partial protection loss.

8.2 Composite and recycled materials

Recycled polymers, insulation products, fibre-reinforced composites and novel claddings may have limited fire-test evidence or unusual degradation mechanisms. Test results must be checked for scale, orientation, joints, substrates, ventilation and end-use configuration. A product classification cannot automatically be transferred to a different assembly.

8.3 Batteries and electrification

Battery energy-storage systems introduce hazards associated with electrical faults, thermal runaway, gas release, re-ignition and emergency response. Structural design should coordinate enclosure integrity, separation, ventilation, detection, suppression, access and replacement strategy. The structural engineer should not treat the battery hazard as solely an electrical issue, because heat and explosion effects can affect compartments and supporting structures.

9. Integrated Professional Workflow

An effective workflow links governance, design, construction and operation:

1. Establish objectives, hazards, stakeholders and regulatory requirements.
2. Define the fire-safety basis of design and information requirements.
3. Develop coordinated structural, architectural and fire models.
4. Identify safety-critical assumptions and interfaces.
5. Verify calculations, products, model data and construction details.
6. Control substitutions, deviations and temporary conditions.
7. Inspect and record installed protection and structural elements.
8. Commission systems and transfer reliable information to the operator.
9. Monitor relevant conditions and review data against decision rules.
10. Update the asset record after alterations, incidents and lessons learned.

Technology is valuable when it strengthens this workflow. It is harmful when it creates a false impression that coordination, inspection or professional judgement is unnecessary.

Applied Case: Illustrative Riverside Learning Centre

Background

A four-storey learning centre is designed with a reinforced-concrete frame, a steel roof, photovoltaic equipment and a limited area of exposed mass timber. The project uses BIM and plans to retain a digital asset record for facilities management. The building is located in a region experiencing hotter summers and occasional wildfire smoke.

Problem

During coordination, a mechanical-services route is proposed through a compartment wall near a steel beam. The route requires a larger opening than shown in the approved model. At the same time, an AI image tool identifies several apparently incomplete fire-stopping installations from site photographs. A sensor network reports elevated roof temperatures, but two sensors have uncertain calibration status.

Analysis

The opening could compromise compartmentation and may require additional protection to the beam. The AI results are treated as inspection leads, not as proof of non-compliance. The roof data are reviewed with calibration records, weather data and sensor locations. The team recognises that a high temperature at the roof surface does not directly establish the temperature of the structural steel.

The governance team records the issue as a safety-critical change. The structural engineer checks the beam's protection and restraint conditions; the fire engineer reviews smoke and fire-spread implications; the services designer revises the route; and the contractor provides installation details and photographs. An independent reviewer checks the revised design.

Approach and solution

The services route is moved where practicable. The remaining penetration receives a tested fire-stopping system with documented compatibility. The steel protection detail is revised only after confirming the product's applicability to the section and connection arrangement. Inspectors revisit all AI-flagged locations and record accepted, repaired or unresolved conditions. Sensors with uncertain calibration are recalibrated or replaced, and the monitoring plan defines alarm and response procedures.

The CDE is updated with the approved revision, inspection evidence, product information, decision record and residual-risk statement. Facilities staff receive training on the locations and limitations of the monitoring system.

Outcome

The project retains a traceable basis for accepting the compartment, structure and monitoring arrangements. The AI tool reduces inspection search time but does not replace competent inspection. The case also demonstrates that a digital model is useful only when it is connected to verified physical evidence.

Learning points

• A model change can be a fire-safety change even when it appears to concern services only.
• AI should identify possibilities, not make unreviewed safety decisions.
• Sensor values require calibration, context and a defined response.
• Approval records should explain why a decision was made, not merely show that a file was issued.
• Resilience requires consideration of climate, utilities, maintenance and future change.

Challenges and Limitations

Digital and governance systems have practical limitations. Small projects may lack resources for sophisticated platforms, and excessive documentation can distract from physical safety. Interoperability remains imperfect. Monitoring produces incomplete data and may fail during the very event being studied. AI tools can reproduce bias, become obsolete when materials or regulations change, and encourage automation bias. Performance-based design can be difficult to communicate to non-specialists. Emerging materials may lack long-term evidence, while climate projections contain uncertainty.

These limitations support a proportional approach: use stronger assurance where consequence, novelty, uncertainty and interdependence are high. Simpler projects still require clear responsibility, competent design, inspection and records.

Best-Practice Checklist

Before project completion, confirm that:

• the fire-safety basis of design is approved and current;
• safety-critical model objects have verified properties and status;
• assumptions and interfaces are recorded;
• changes and substitutions have been assessed;
• protection and fire stopping have physical inspection evidence;
• AI-assisted outputs have human review and an audit trail;
• sensors are calibrated and linked to response procedures;
• residual risks are communicated to the operator;
• records are secure, accessible and retained;
• future alterations and climate-related hazards are addressed.

Chapter Summary

Project governance establishes authority, accountability, evidence and change control for structural fire safety. BIM can connect structural, architectural, fire and operational information, but model data must be validated against installed conditions. Monitoring and digital twins are useful when they begin with a decision question, define uncertainty and provide a responsible response pathway. AI can support inspection, coordination and information analysis, but it cannot remove professional accountability or replace competent judgement.

Competence management, independent checking, ethical transparency, secure records and controlled change are essential to reliable fire-safety outcomes. Resilience requires consideration of climate hazards, power loss, recovery and future adaptation. Emerging materials and energy systems must be assessed as complete assemblies with appropriate evidence. The central professional principle is simple: technology should strengthen traceability and judgement, not conceal uncertainty or transfer responsibility to an algorithm.

Key Terms

Governance: The framework for directing, authorising and assuring project decisions.

Accountability: The obligation to answer for a decision or action.

BIM: A managed process for creating and maintaining information about a built asset.

Common data environment: A controlled system for storing, sharing and managing project information.

Information requirement: Defined information needed for a particular project purpose or decision.

Safety-critical element: An element whose failure or incorrect information could significantly affect life safety or structural stability.

Digital twin: A maintained digital representation connected to information from a physical asset.

Monitoring: Systematic observation or measurement of conditions, behaviour or performance.

Measurement uncertainty: The quantified doubt associated with a measured value.

Artificial intelligence: Computer-based methods that perform tasks involving pattern recognition, prediction or decision support.

Automation bias: The tendency to trust an automated output excessively.

Competence: The combination of knowledge, skill, experience, judgement and recognised limits.

Quality assurance: Planned actions intended to prevent defects and ensure reliable processes.

Quality control: Inspection or testing used to identify whether requirements have been met.

Change control: A formal process for assessing, approving and recording changes.

Safety case: A structured claim, argument and evidence demonstrating that risks are controlled.

Resilience: The capacity to anticipate, withstand, recover from and adapt to disruptive events.

Residual risk: Risk remaining after planned controls have been applied.

Thermal runaway: Rapid self-heating in a system, especially a battery, that can lead to fire or hazardous gas release.

Review Questions

Short Answer Questions

1. What is the difference between project governance and project management?
2. List four types of information that should be associated with a structural fire-safety BIM object.
3. Why should an AI system normally be treated as decision support rather than as the final decision-maker?
4. What is measurement uncertainty, and why is it important for alarm thresholds?
5. State three ethical duties relevant to structural fire-safety practice.

Long Answer Questions

1. Explain how a common data environment can improve the reliability and traceability of structural fire-safety information.
2. Discuss the competence and quality-assurance arrangements needed for a performance-based structural fire design.
3. Describe how climate change can alter structural fire-safety risks and propose resilience measures.
4. Compare the benefits and limitations of BIM, monitoring systems and AI-assisted tools.
5. Explain why emerging materials must be assessed as systems rather than solely by material name or product classification.

Analytical / Application Questions

1. A contractor proposes replacing a specified fire-protection product with a cheaper alternative that has a similar advertised rating. Identify the governance steps, technical evidence and approval records required before acceptance.
2. A monitoring sensor reports a temperature of 78 °C with an uncertainty of ±5 °C, while the project review threshold is 80 °C. Explain why the reading should not automatically be classified as safe and state an appropriate response.
3. During refurbishment, a new service penetration is proposed through a compartment wall supporting a protected beam. Analyse the structural, fire, construction and information-management issues that should be resolved before the change is approved.

Multiple Choice Questions with Answers

Question 1

Which statement best describes project governance in structural fire safety?

A. It is limited to producing a three-dimensional model.

B. It determines authority, accountability, evidence and change-control arrangements.

C. It is the same as calculating the capacity of a structural member.

D. It removes the need for professional judgement.

Correct Answer: B. It determines authority, accountability, evidence and change-control arrangements.

Question 2

Which item is most appropriate for a safety-critical BIM object?

A. Only its colour and viewing scale

B. Only the name of the software used to create it

C. Its required fire resistance, protection information and approval status

D. A decorative rendering of the completed building

Correct Answer: C. Its required fire resistance, protection information and approval status

Question 3

A sensor indicates a temperature of 78 °C with an uncertainty of ±5 °C. What is the main governance implication when the threshold is 80 °C?

A. The value is certainly below the threshold.

B. The uncertainty range crosses the threshold, so the reading requires review.

C. The sensor should be ignored because all sensors are inaccurate.

D. The threshold should automatically be raised to 85 °C.

Correct Answer: B. The uncertainty range crosses the threshold, so the reading requires review.

Question 4

What is the most responsible use of an AI image-recognition tool on a construction site?

A. Using its result as final proof that all fire stopping is compliant

B. Using it to identify possible defects for competent human inspection

C. Allowing it to approve material substitutions automatically

D. Using it to replace all inspection records

Correct Answer: B. Using it to identify possible defects for competent human inspection

Question 5

Which action is an example of quality assurance?

A. Planning checking levels and hold points before work begins

B. Deleting superseded design records

C. Accepting an unverified product because it is cheaper

D. Changing a model without recording the change

Correct Answer: A. Planning checking levels and hold points before work begins

Question 6

What should a change-control request for a new compartment penetration primarily include?

A. Only the contractor's preferred installation date

B. The proposed change, affected hazards, evidence, reviewers and decision

C. Only a revised architectural colour scheme

D. A statement that all penetrations are minor

Correct Answer: B. The proposed change, affected hazards, evidence, reviewers and decision

Question 7

Which statement about a digital twin is correct?

A. It is always identical to a static design drawing.

B. It is a maintained digital representation linked to information from the physical asset.

C. It eliminates the need for physical inspection.

D. It can be used without data ownership or access control.

Correct Answer: B. It is a maintained digital representation linked to information from the physical asset.

Question 8

The conceptual relationship Risk = Likelihood × Consequence is useful mainly for:

A. Ranking risks when scales and assumptions are defined consistently

B. Proving that low-likelihood catastrophic events can be ignored

C. Replacing all engineering analysis

D. Calculating exact structural temperatures in every fire

Correct Answer: A. Ranking risks when scales and assumptions are defined consistently

Question 9

Which issue is particularly relevant to mass-timber fire design?

A. Only the colour of the timber surface

B. Charring, exposed surfaces, connections and possible continued burning

C. The assumption that all timber assemblies have identical performance

D. The elimination of the need for compartmentation

Correct Answer: B. Charring, exposed surfaces, connections and possible continued burning

Question 10

Who remains accountable when an AI tool produces an output used in a safety-critical decision?

A. The algorithm alone

B. The internet service provider in every case

C. The competent professional or authorised decision-maker who accepts and uses the output

D. No one, because automated results are objective

Correct Answer: C. The competent professional or authorised decision-maker who accepts and uses the output