Two different studies share the abbreviation PHA, and a project team that is asked for one has to find out which is meant. A preliminary hazard analysis is an early-design review that lists a concept's hazards, makes an initial risk assessment and proposes mitigation. A process hazard analysis is the formal hazard evaluation that United States process safety rules require for a covered process. The same name is used in defence and aerospace system safety, so this guide states where that meaning ends and the process-industry meaning begins. It draws on the US Department of Defense (DoD) standard MIL-STD-882E and on Center for Chemical Process Safety (CCPS) guidance. The regulated meaning comes from rules of the US Occupational Safety and Health Administration (OSHA) and Environmental Protection Agency (EPA). The approach reflects the HAZOP and HAZID studies that Global Formulation supports for plant owners.
The two names look interchangeable, and they are not. A preliminary hazard analysis belongs to early design, when only a concept exists. A process hazard analysis belongs to a defined process with drawings, and in the United States it is a legal requirement for covered processes. The table compares the two on the points that decide which one a project needs.
| Criterion | Preliminary hazard analysis | Process hazard analysis |
|---|---|---|
| Where the term is used | System-safety standards such as MIL-STD-882E, and the CCPS hazard evaluation guidelines | United States rules: OSHA 29 CFR 1910.119(e) and EPA 40 CFR 68.67 |
| Project stage | Concept and early design, before detailed drawings exist | Detailed design before start-up, then revalidation during operation |
| Question it answers | What hazards does this concept create, and what should the design do about them? | Are the hazards of this process identified, evaluated and controlled? |
| Typical inputs | The concept description, the materials, the energy sources, the site concept and experience from similar plants | Process safety information, including piping and instrumentation diagrams (P&IDs), relief design basis and thermal and chemical stability data |
| Method | A non-scenario-based review of hazards, causes, effects and mitigation, in the CCPS classification | One or more listed methods, such as What-If, Checklist or hazard and operability study (HAZOP), or an appropriate equivalent |
| Status in the US rules | Not a listed method; OSHA's non-mandatory Appendix C suggests it as an aid to prioritising which processes to analyse first | Required for covered processes under OSHA process safety management and EPA Program 3 |
| Typical output | A hazard register with initial risk assessment and proposed mitigation | A record of the analysis, recommendations, and the documented resolution of each recommendation |
Both studies are forms of hazard identification and risk assessment, the generic activity, but neither name is a synonym for it. The sections below take each study in turn, place them in a project and show how to choose between them.
A preliminary hazard analysis asks what can go wrong with a design that is still a concept. It is run early because the cheapest hazard to remove is one that has not been drawn yet. The study is broad and shallow by design, since the detail it would need does not exist. Its value is to shape the design and to decide what the later, deeper studies must examine.
In system safety the term comes from MIL-STD-882E, the US Department of Defense standard practice for system safety. Task 201 compiles a Preliminary Hazard List of the potential hazards of a concept early in development. Task 202 then performs a preliminary hazard analysis to identify hazards, assess the initial risks and identify potential mitigation measures. The standard applies when a solicitation or contract requires it, and the edition read for this guide is MIL-STD-882E with Change 1, dated 27 September 2023.
For each hazard, Task 202 requires an initial risk assessment using the standard's severity and probability definitions, unless a tailored scheme is formally approved. It also requires mitigation measures chosen by the system safety design order of precedence, which ranks them by decreasing effectiveness and starts with eliminating the hazard through design selection. The standard lists the sources of hazard to consider, including energy sources, hazardous materials, interfaces, operating environment, procedures, health hazards and human factors. Its tables and matrix are written for defense systems, so this guide does not reproduce them or apply them to a process plant.
The CCPS book Guidelines for Hazard Evaluation Procedures (3rd edition, 2008) lists preliminary hazard analysis as a procedure in its own right. It sits in the chapter on non-scenario-based procedures, beside safety review, relative ranking and checklist analysis. The scenario-based chapter holds What-If analysis, HAZOP, failure mode and effects analysis (FMEA), fault tree analysis and bow-tie analysis. The book also includes a worked example of a preliminary hazard analysis at conceptual design.
In a process project the study is run at conceptual design, when the process concept and the main materials are known but detailed drawings are not. The team reviews the materials and their properties, the reactions, the energy sources, the inventories, the site concept and the lessons from similar plants. These prompts echo the sources of hazard in MIL-STD-882E, translated to a chemical plant. The output is a hazard register with initial mitigation proposals and a list of items for the later studies.
A preliminary hazard analysis records each hazard in a short worksheet. The columns are the hazard, its cause, its effect, the controls proposed at concept stage and the follow-up. The row below is invented and generic, and it carries no risk rating. System-safety versions add an initial risk assessment from the project's own severity and probability definitions.
| Hazard | Cause | Effect | Controls proposed at concept | Follow-up |
|---|---|---|---|---|
| Flammable solvent held in an intermediate tank | Overfilling, or a leak near an ignition source | Fire or release of flammable vapour | Reduce the inventory, and place the tank clear of ignition sources and occupied buildings, with distances taken from the applicable code | Carry forward to the detailed-design hazard study and the layout review |
At detailed design the same hazard is examined again, node by node. The transfer line that fills the tank is tested against deviations such as more flow, and the safeguards actually drawn on the P&ID are recorded. That second examination is the work of the process hazard analysis.
In the process industries the term process hazard analysis comes from United States regulation. The OSHA process safety management standard requires one for covered processes, and the EPA Risk Management Program (RMP) rule requires one for Program 3 processes. The requirement is for a hazard evaluation that is appropriate to the complexity of the process. It must identify, evaluate and control the hazards involved in the process.
Paragraph (e) of 29 CFR 1910.119 sets the requirement for a covered process. The employer performs an initial analysis, using one or more of the listed methodologies, and revalidates it at least every five years. Before any analysis, the employer compiles process safety information under paragraph (d), including P&IDs, the relief system design and design basis, and thermal and chemical stability data. The bullets below set out what an owner can check against a report.
Paragraph (e)(1) also requires the employer to determine and document the order in which covered processes are analysed. The rationale considers the extent of the hazards, the number of potentially affected employees, the age of the process and its operating history. OSHA's non-mandatory Appendix C suggests a preliminary or gross hazard analysis as one aid to that ranking. Paragraph (c) adds that employers consult employees and their representatives on the conduct of the analyses and give them access to the results.
The standard's coverage, set out in paragraph (a), depends on the chemicals and quantities in the process. Our guide to the HAZOP study sets out the coverage test and the deliverables when HAZOP is the chosen method.
The EPA Risk Management Program rule has its own process hazard analysis requirement in section 68.67, which applies to Program 3 processes. Program 3 includes covered processes that are subject to the OSHA standard. The section lists the same methodologies as OSHA and states that analyses completed to comply with 29 CFR 1910.119(e) are acceptable as initial analyses. It adds topics such as natural hazards, siting with respect to public receptors and, for certain processes, safer technology analysis.
Program 2 processes carry a hazard review under section 68.50 instead. That review identifies the hazards, the opportunities for equipment malfunction or human error, the safeguards, release detection, natural hazards and siting. It is updated at least every five years. The section allows checklists developed by knowledgeable persons as a guide.
The rule is under revision. The eCFR text read on 28 October 2026 includes the 2024 amendments. EPA published a proposed revision on 24 February 2026 (91 Federal Register 8970) and accepted comments until 11 May 2026. Check the current text on the EPA RMP page before relying on any provision.
Other jurisdictions run their own major-accident regimes. Examples are the Seveso III Directive (2012/18/EU) in the European Union and the Control of Major Accident Hazards Regulations 2015 in Great Britain. This guide names them only and does not describe their clauses. Confirm the regime that applies to the site before relying on a United States description.
The two studies are stages of one hazard-management sequence, not alternatives. An early study shapes a design that does not yet have drawings. The regulated analysis then examines the controls in the design that was drawn, and it is repeated as the plant ages. Figure 1 shows the sequence for a new process under United States rules.
A preliminary study is done before the controls are designed, so it proposes controls but cannot test the ones in the final design. Paragraph (e)(3) requires the regulated analysis to address the controls and the consequences of their failure. Whether a given study satisfies a legal requirement is a question for the regulator and a qualified assessor. For a new facility, paragraph (i) requires the pre-startup safety review to confirm that a process hazard analysis has been performed and its recommendations resolved or implemented before start-up.
Chemistry data feeds both studies. OSHA's process safety information includes thermal and chemical stability data, and a reaction's heat release and decomposition behaviour are measured, not assumed. Our guides to reaction calorimetry and to adiabatic temperature rise and MTSR explain that data. Moving a route from laboratory to plant puts the process on equipment that nobody has yet operated, so scale-up support is a natural point for the early hazard review.
The right study follows from the stage of the project, the rule that applies and the question the owner has to answer. A contract that asks for a PHA without more detail needs a written clarification before a quotation is accepted. The table matches common situations to studies. Where a regulation applies, its own list of methods decides which methods qualify.
| Your situation | Study | Reason |
|---|---|---|
| A concept or route selection, with no P&IDs yet | Preliminary hazard analysis, or a hazard identification (HAZID) study | Hazards can still be designed out at the cost of a drawing change |
| Detailed design is complete and the process is covered by OSHA 29 CFR 1910.119 | Process hazard analysis using a listed method, before start-up | Paragraph (i) asks for it before highly hazardous chemicals are introduced to a new facility |
| A United States process is in EPA RMP Program 3 | Process hazard analysis under section 68.67 | OSHA analyses are accepted as initial analyses, and the section adds its own topics |
| A United States process is in EPA RMP Program 2 | Hazard review under section 68.50 | The section sets its own list of items and a five-year update |
| Five years have passed since the last analysis of a covered process | Revalidation by a team that meets the team requirements | Paragraph (e)(6) keeps the analysis consistent with the current process |
| A modification, new product or higher capacity at an existing plant | The study that the site's change-control process calls for | Change control decides how much of the plant to re-examine |
| Several covered processes need analysis and the order is unclear | A preliminary or gross hazard analysis to rank them | Paragraph (e)(1) requires a documented priority order, and Appendix C suggests this aid |
| A lender, insurer or contract names a PHA without saying which | Ask in writing which study and which standard | The abbreviation fits both studies |
| The system is a defense or aerospace system | System safety under the standard the contract names, such as MIL-STD-882E | This guide covers process plants only |
A staged approach pairs the two studies: a preliminary hazard analysis shapes the concept, and the regulated analysis tests the design that results. Each study answers a different question, so neither replaces the other.
A hazard study goes wrong in a few recognisable ways. It is requested by an ambiguous name, run at the wrong time, run on the wrong drawings or left open. Each carries a plant consequence that appears after the money is spent. The table lists the failures and the step that prevents each one.
| What goes wrong | Consequence | How to prevent it |
|---|---|---|
| The contract says only "PHA" | Each party prices a different study, and the missing one is discovered late | Name the study, the rule or standard and the deliverable in the scope document |
| A preliminary study is run once at concept and never revisited | Design changes made after the concept go unexamined | Carry its findings into the detailed-design study and review again at design freeze |
| An early-design study is presented as the regulated analysis | No examination of the designed controls exists when the pre-startup review asks for the analysis | State which requirement each study answers, and confirm it with the regulator or a qualified assessor |
| The regulated analysis is run on outdated P&IDs | The team studies a plant that does not exist, so real hazards go unexamined | Compile the process safety information first and freeze a verified drawing revision |
| Recommendations are not tracked to closure | The same findings return at revalidation, and the record fails an audit | Keep a register with an owner, a date and a closure note for each action |
| Revalidation passes the five-year date | The analysis no longer matches the process, and the next change builds on it | Record the due date in the report and in the site calendar |
| A scoring scheme built for another domain is copied across | Ratings that do not map to the owner's own criteria or to process-plant consequences | Agree the owner's criteria with a qualified facilitator before the first session |
| Thermal data is missing for a reactive process | A runaway is named as a consequence but its speed is unknown | Measure the heat release and decomposition behaviour before the study |
The first three rows are prevented by a clear brief, written before any session is held. The rest are prevented by preparation and follow-through. A study of the wrong drawing or the wrong chemistry produces confident worksheets about a plant that does not behave that way.
An owner who briefs the study precisely gets a deliverable that can be checked against a rule. The brief is short, and each line closes a question that would otherwise surface in the middle of the project. Work through the eight items below before inviting quotations. The list applies to either study.
On delivery the owner should receive the following documents:
To check the work, compare the drawing revisions in the report with those you issued, and confirm that the report names the study and the requirement it answers. Then sample a handful of actions at random and confirm that each has an owner, a date and a status.
Owners who need hazard studies scoped and facilitated can start from our process safety and thermal hazard assessment service. It covers HAZOP and HAZID studies, thermal hazard assessment and pre-startup safety review. Where the studies sit in a delivery schedule, see our EPC and EPCM project management service. The checklist above is the brief to bring to that conversation.
A preliminary hazard analysis is an early-design review that lists the hazards of a concept, makes an initial risk assessment and proposes mitigation. In MIL-STD-882E it is Task 202, performed to identify hazards, assess the initial risks and identify potential mitigation measures.
Process-industry guidance, such as the CCPS hazard evaluation guidelines, lists it as a non-scenario-based procedure.
A process hazard analysis is the hazard evaluation that OSHA 29 CFR 1910.119(e) requires for a covered process. It must be appropriate to the complexity of the process and must identify, evaluate and control the hazards involved.
The employer uses one or more listed methods, such as What-If, Checklist or HAZOP, or an appropriate equivalent.
PHA stands for preliminary hazard analysis in system safety and in MIL-STD-882E. In process safety it stands for process hazard analysis, the term used in the OSHA and EPA rules.
The CCPS guidelines abbreviate the early study as PreHA, so a contract that says only PHA needs the study named in full.
A preliminary hazard analysis is done at concept stage, before detailed drawings exist, to shape the design. A process hazard analysis is done on a defined process and covers the controls and the consequences of their failure.
In the United States it is a legal requirement for covered processes. The two are stages of one sequence, and each answers a different question.
Not as such. It is not one of the methodologies listed in paragraph (e)(2) of 29 CFR 1910.119, and no paragraph requires it.
OSHA's non-mandatory Appendix C notes that a preliminary or gross hazard analysis is an aid to prioritising which covered processes are analysed first.
In the United States, OSHA 29 CFR 1910.119 requires one for the processes it covers, and EPA 40 CFR 68.67 requires one for Program 3 processes. For a new facility the pre-startup safety review confirms that the analysis has been performed and its recommendations resolved or implemented before start-up.
Other countries have their own major-accident regimes, so confirm the one that applies to the site.
In the United States, OSHA paragraph (e)(6) and EPA section 68.67 both require the analysis to be updated and revalidated at least every five years. The updating team must meet the same requirements as the original team.
Other jurisdictions set their own intervals, so check the regime that applies.
Process hazard analysis is the requirement, and HAZOP is one of the methods that can be used to meet it. The other methods named in the OSHA text are What-If, Checklist, What-If/Checklist, Failure Mode and Effects Analysis and Fault Tree Analysis.
A preliminary hazard analysis is an earlier study with a different purpose.
A typical worksheet has columns for the hazard, its cause, its effect, the controls proposed and the follow-up. MIL-STD-882E adds an initial risk assessment for each hazard, using the standard's severity and probability definitions.
The worksheet records hazards at concept stage and carries them forward to the detailed-design study.
Global Formulation supports HAZOP and HAZID studies, thermal hazard assessment and pre-startup safety review, starting from the drawings and process data your plant already has.
Discuss a Safety Assessment