What ASME B31.1 Actually Requires You to Examine, Heat Treat and Prove

Scope, and the piping that quietly belongs to another code

B31.1 applies to piping in electric power generating stations, in industrial and institutional plants, in geothermal heating systems and in central and district heating and cooling systems. That sounds broad until you stand in a combined heat and power plant and try to draw the line. The steam and feedwater are B31.1. The fuel gas train may be B31.3 depending on the specification. The building heating loop is B31.9. Nothing in the pipe itself tells you which, and the welder cannot tell by looking. The code of construction is a specification decision, made early, and it is binding.

The exclusions are equally important. Components covered by the Boiler and Pressure Vessel Code, including boilers, economizers and pressure vessels, sit outside B31.1, as does piping internal to those components. What remains is the connecting piping, and the boundary is drawn at a defined joint rather than at a convenient flange. Where the boundary is ambiguous on the isometric it becomes ambiguous in the records, and that ambiguity is what an auditor pulls on. Fixing it costs an hour at the specification stage and weeks after fabrication.

The financial consequence of the code choice is large and frequently overlooked. B31.3 Normal Fluid Service asks for 5% random radiography of girth welds under paragraph 341.4.1. B31.1 Table 136.4 drives full volumetric examination on the heavier, hotter, higher pressure welds. Move a line between codes and the examination budget moves by an order of magnitude. If your project has not had that conversation, an ASNT Level III consulting review of the line list against the code of construction usually pays for itself before fabrication starts.

Boiler external piping: inside B31.1, administered by Section I

Boiler external piping is the piping between the boiler proper and the valves required by ASME Section I. It is the single most misunderstood jurisdiction in power piping. The technical rules, the design equations, the welding and the examination come from B31.1. The administration does not. Certification, the involvement of an Authorized Inspector and the data report forms come from Section I, which means a Certificate of Authorization holder, an Authorized Inspection Agency, and a signed data report before the work is accepted.

Non-boiler external piping carries none of that. There is no code stamp for it, no Authorized Inspector and no data report. The owner's inspector of paragraph 136.1 is the acceptance authority. The recurring failure is a fabricator who treats the whole scope as non-boiler external piping, welds it, tests it, and then tries to obtain the Section I data report retrospectively. You cannot retro-certify work the Authorized Inspector never saw. The remedy at that point is cut-out and re-do, or an owner-accepted deviation that will follow the plant for its life.

The control is simple and almost never done properly. Mark the boiler external piping boundary on the process and instrumentation diagram and on every affected isometric, name the Section I stamp holder on the drawing, and state on the weld map which welds fall inside the Section I boundary. Then the welder, the examination technician and the document controller are all working from the same line. Boundary drawings agreed before the first cut are worth more than any amount of argument after hydrotest.

Pressure design: the equation, and the allowances that get lost

The minimum wall for straight pipe under internal pressure comes from paragraph 104.1.2: design pressure multiplied by outside diameter, divided by twice the sum of the allowable stress times the joint efficiency plus the pressure times the Y coefficient, with an additive allowance A. The allowance A covers threading and grooving depth, corrosion and erosion, and any mechanical strength requirement. It is not a fudge factor, and it is not optional where the service demands it.

The number that disappears most often is not in the equation at all. Seamless pipe carries a mill tolerance of 12.5% on wall thickness, so a nominal wall is not a guaranteed wall. If the calculated minimum equals the nominal ordered wall, the pipe delivered can be thin by an eighth of its thickness on arrival, before a single hour of service. Ordering practice has to add the tolerance on top of the calculated minimum plus allowances. This is a design office discipline, and it shows up years later as a thickness reading that looks like corrosion but is original.

Allowable stresses come from the B31.1 mandatory appendix, keyed to material specification and temperature. They are derived on the same philosophy as the boiler code values but they are not automatically identical to the numbers a designer remembers from another section, and they move between editions. Anyone re-rating existing piping should verify which edition the original calculation used. Where a plant carries that calculation history in a records system rather than in a spreadsheet, re-rating becomes an afternoon rather than a project, which is one practical reason plants move examination and design records into inspection management software.

Table 136.4 is a floor, not a ceiling

Paragraph 136.4 states that piping shall be examined to the extent specified in the code or to any greater extent specified in the engineering design. That single clause resolves most disputes on site. When the project specification requires more radiography than the table, the specification governs, and no appeal to the code will reduce it. When the specification is silent, the table governs. Reading the table alone, without the specification, is how a fabricator ends up with an unexamined weld that the client will not accept.

The examination required is keyed to the weld type, the size and wall thickness, and the service. Volumetric examination concentrates on girth and longitudinal groove welds in the heavier, higher temperature, higher pressure piping, with full coverage where boiler external piping brings Section I with it. Fillet, socket and seal welds are surface examined by magnetic particle or liquid penetrant rather than radiographed, because the geometry defeats a useful radiograph. Branch connections and their reinforcement need attention that the table addresses but that estimators habitually leave out of the take-off.

Paragraph 136.4.2 offers in-process examination, a structured sequence of visual checks during welding by qualified personnel, as an alternative to radiography where the engineering design permits it. It is a legitimate route and a good one on socket-welded small bore where radiography is close to useless. The audit finding is never about the technique. It is that the written approval permitting the substitution does not exist, so the shop has simply not radiographed welds the code required. Where in-process examination is used, keep the approval, the technique record and the examiner qualifications together, and put the resulting records through independent review of inspection reports before they enter the turnover package.

Preheat, PWHT and hardness: where P-Number discipline is enforced

Table 131 sets minimum preheat by P-Number, thickness and material tensile strength. Table 132 sets post-weld heat treatment holding temperature and holding time by P-Number, with thickness based exemptions for carbon steel. These are not interchangeable with the equivalent tables in other codes, and a shop working several codes at once will eventually apply the wrong one. The most consequential material is Grade 91, a creep strength enhanced ferritic steel with a narrow usable holding band in the region of 1,350 to 1,425 degrees Fahrenheit, bounded above by the lower critical temperature of the actual heat.

Take Grade 91 above that lower critical temperature and the weld partially re-austenitises. Cooling from there produces untempered martensite that no amount of subsequent soaking at the original temperature will repair. The joint looks finished, passes radiography, and carries a fraction of the creep life the design assumed. The only field indications are the chart and the hardness reading. That is why paragraph 132.1.1 exists and why its sampling rule is written the way it is.

That rule is at least 10% of welds, hot bends and hot formed components in each furnace batch, and 100% of welds given local post-weld heat treatment, with the Brinell result inside the Table 132.1 limit for the P-Number. Shops generally handle the furnace sampling correctly because the furnace produces a tidy batch record. Field local treatment, done with resistance pads on a scaffold at night, is where the 100% requirement is quietly dropped. Ask for the hardness record for every locally treated weld, cross-check chart identification against the weld number and the thermocouple positions, and expect gaps.

Leak testing, and the things that void the test you just did

Paragraph 137.4.5 requires hydrostatic test pressure of not less than 1.5 times design pressure at every point in the system, limited so as not to exceed 1.5 times the maximum allowable pressure of any component within it. That upper limit is the one that catches people: a single low-rated valve or gauge in the circuit caps the test for the entire boundary, and either it comes out or the test pressure comes down. The pressure is held for at least 10 minutes and may then be reduced to design pressure for the examination itself.

Test temperature is a safety matter as much as a code matter. Heavy-wall low-alloy piping tested cold sits in a regime where the critical flaw size is small, and a hydrostatic test stores a large amount of energy in a stiff system. Establish a minimum metal temperature for the test and record it. Where a pneumatic test is unavoidable, paragraph 137.5 imposes additional conditions and requires the owner's agreement, not the contractor's convenience.

The things that invalidate an otherwise sound test are almost all mechanical and almost all preventable: spring hangers not pinned so the system settles under water weight, expansion joints unrestrained, relief valves left in circuit, vents not proven open so a trapped air pocket cushions the system, and temporary blinds of unverified rating. The record fails just as often as the test. If the gauge range, gauge serial number and calibration date are not on the test report, the pressure you achieved cannot be evidenced later, and an auditor will treat the test as unproven regardless of what the witnesses remember.

Who is allowed to accept the work

Paragraph 136.1 places the inspection duty on the owner. The owner's inspector must not be an employee of the manufacturer, fabricator or erector unless the owner is that party, and must hold real experience in power piping design, fabrication or inspection. The intent is independence of judgement at the point of acceptance, and the code says so plainly. It is not a paperwork role and it is not delegable to the contractor's quality department by private agreement.

The finding that recurs is structural rather than technical. The nominally independent inspector is engaged and paid by the erector, reports to the erection superintendent, and is asked to accept the erector's own work. Nothing about that arrangement satisfies paragraph 136.1, and it is visible on a contract chart in about ninety seconds. Owners intending to rely on an external inspection resource should engage it directly and have the scope written so the inspector reports to the owner. A short consultation at the contracting stage is far cheaper than restructuring the arrangement mid-project.

Do not conflate the owner's inspector with the Authorized Inspector. The Authorized Inspector is a Section I role and appears only where boiler external piping brings Section I jurisdiction. Projects that confuse the two either pay for an Authorized Inspector on scope that does not need one, or discover at data report time that nobody holding the necessary commission ever attended. Both errors are avoidable by writing the boundary and the inspection authority into the quality plan on day one.

Personnel, procedures and the review of the data

Paragraph 136.3 requires nondestructive examination personnel to be qualified and certified under the employer's written practice, based on ASNT SNT-TC-1A or ANSI/ASNT CP-189, with the edition fixed by the referencing documents. The practice has to reflect what the employer actually does, including training hours, examination content, vision testing and the recertification interval. Auditors do not compare your practice to an ideal. They compare your records to your own practice, and the practice is usually more demanding than the records can support.

Procedures are the other half. Section V requires written procedures and, for several methods, demonstration that the procedure achieves what it claims on representative geometry. Radiographic technique sheets have to match the shot actually taken, ultrasonic instrument linearity records have to identify the instrument used by serial number, and penetrant work has to record temperature and dwell times. Companies whose technicians come through structured NDT training to SNT-TC-1A tend to produce these records without prompting, because the requirement was taught alongside the technique.

Finally, someone competent has to read the data with no commercial interest in the answer. Radiographs interpreted by the technician who shot them, on a schedule set by the erector, produce a predictable bias toward acceptance. Independent review by a Level III who did not perform the examination is the cheapest quality control available on a piping project, and it is the control that most often finds the missing coverage, the mismatched technique sheet and the repair that was never re-examined to its full extent.

Request a consultation