Condition assessment of steel structures: condition categories and what changes in the calculation of an existing structure

A new structure and a working one: what differs for the calculation

A design calculation covers uncertainty with margins: steel at the minimum values from the standard, loads from the code with reliability factors, sections from the rolling tables. For a working structure measurement removes part of the uncertainty: it can be surveyed, weighed and tested. But another kind appears: physical wear, deviations from the design, an unknown loading history.

So carrying the design method over directly gives a wrong result in both directions. If the structure was built without spare margin, then, with wear taken into account, it «fails» the formulas of the new code, although it has stood for decades. If loads have grown after reconstruction, a check on the old section and old loads gives overstated reliability. A verification calculation puts the actual forces and the actual load-bearing capacity into one inequality, and takes its margins from the current codes instead of inventing its own.

There are two kinds of task. Direct: will the structure carry the given or refined loads. Inverse: what load will it carry with the defects it has, that is, what is its residual capacity. The inverse one is needed when deciding whether the structure will last until the next repair and how much to strengthen.

What is actually replaced in the calculation

The inequality itself stays the same: the forces from the loads must not exceed the capacity. What changes is the content of each part.

Loads: what is refined and by how much

The dead weight of the steel is taken from measurement, or from KMD drawings with control measurements of the sections, and where there are no drawings, from measurement of the members. The reliability factor for such a load is taken as one: the weight has been measured, not assumed. Roofs and floors are assessed by opening up the roof and finding the actual layer make-up: samples are taken from pits, not fewer than five, and the results are processed statistically.

The crane action is taken from the data sheet or from weighing the crane, taking into account where the crane really works and how close to the column row the trolley comes. The factor for the load depends on crane type, duty group and capacity. Snow and wind loads are refined from the nearest weather station in similar conditions of exposure: snow needs records for at least 20 years, wind for at least 15, and the design value depends on the period for which the decision is made.

The period also works as a correction. If the structure only has to be kept going until the next repair, for not longer than a year, the design snow, wind and ice loads may be reduced by 20 %, as for the erection period. This does not apply to a longer period.

In recalculation practice, taking the actual process into account lowers the crane action sometimes by up to 40 %, and refining snow and wind for a specific site gives 5–25 %. These are upper limits, not expected values: they cannot be assumed in advance, and each case is confirmed by process documents and observation data.

Defects: what goes into the calculation and what does not

Defects fall into those whose effect can be expressed by calculation and those that calculation does not excuse. The first group includes cut-outs, tears and burn-through holes, corrosion as section loss, general bowing and local dents, shifted member axes, joint eccentricities, fixing away from a node, and replaced sections or types of connection. They are entered either by changing the section geometry or the design scheme, or through corrections in the verification formulas.

Cut-outs and weakened sections are checked on the net area, with the eccentricity from the shift of the centroid of the weakened section. A bowed member is shown straight in the scheme but with reduced axial stiffness, and a compressed one is calculated as eccentrically compressed, replacing the bow with an equivalent eccentricity. The bow is measured on site and, where possible, brought back to the unloaded state.

The second group: cracks in the parent metal and welds, ruptures and fractures of members, weld defects (except a reduced leg), through corrosion, loose or missing bolts and rivets, bearing damage in connections, torn-off heads, dangerous types of corrosion. These are removed, not covered by calculation. How that is done is covered in the article on typical damage and repair.

Separately from the condition category, defects are classed by danger. Category A: a defect in a critical member or connection that threatens immediate failure. Category B: no direct danger now, but the defect can damage other members and in time turn into A. Category C: not linked to a threat of failure. With a category A defect the responsible persons are told at once, people and property in the zone are made safe, and then the structure is temporarily braced, unloaded or urgently repaired.

The design scheme of an existing frame

This is the step that is hardest to reduce to rules. The loading history and the causes of damage are usually unknown, so the calculation is run on several schemes to see which is worse. A typical example: a truss is taken as rigidly fixed to the column as long as the top bearing joint does not move relative to it, and a mutual movement of a few millimetres is enough to turn it into a hinge. So several variants are set for joints and supports, and the worst is taken for each member.

The work starts with a simpler plane scheme with rigid or hinged joints and goes from the upper structures downward, passing on the support reactions. A spatial scheme is used when by the simple scheme the building should already have collapsed, yet it stands: the spatial scheme takes in the support from neighbouring structures. The state of the supports is checked on site: anchor bolts, shifted base plates, eccentric bearing. If a joint is encased in grout and cannot be seen, its properties are taken from indirect signs, for example the integrity of the grouting.

Linear elastic analysis remains the main one. Non-linear elastic analysis is needed to assess stability. Plastic analysis is usually not used for existing structures: the loading history is unknown, and it gives a higher estimate of capacity, that is, it removes the margin of a working structure.

A refined scheme can reveal reserves: the rotation of foundations, the true stiffness of the truss-to-column fixing, the action of the roof diaphragm, longitudinal bracing and end wall framing. But it works the other way too: an error in the joint model shifts the column moments by tens of per cent, up or down. The roof diaphragm is taken as rigid only when the survey has confirmed that the roofing, its fixing to the purlins and the joints are undamaged.

Two other routes to an assessment: service experience and load testing

Calculation is not the only way. A steel structure can be accepted as serviceable on service experience alone, but only when two conditions hold at once: it has worked for at least 25 years and no change of duty is expected. The period can be shortened, but not below ten years, if the share of snow load in the design combination is under a quarter and the decision lasts no longer than a third of the service already completed. Even then there must be no category A or B defects, otherwise a calculated or experimental assessment is needed.

A load test is an experiment, and it has a programme: the aim, which members are loaded, how the load is applied, the steps, what is measured and where. The load is applied in steps, not fewer than three up to the first control value, with a hold; the load is controlled to ±5 %. Loading stops at once if a member fails, local buckling occurs, cracks open wider than 0.2 mm or grow, or bolts or rivets shear. People are forbidden to be under the loaded structure.

The structure is accepted as serviceable if there was no failure, the movement at the first control value is acceptable and leaves no permanent deformation, and the ratio of residual to total movement at the second control value does not exceed the limit. The limit is stricter for welded structures than for riveted and bolted ones. When different methods give different results for one member, the most cautious assessment is taken.

Condition categories and what to do about them

The four condition categories (sound, serviceable, serviceable with limits, emergency) and the I–IV scale are described in the article Diagnostics and Assessment of the Technical Condition. What matters more here is the decision behind each category.

How the decision is made: keep, strengthen or replace

The decision is always tied to the forecast period: to the end of the standard life of the structure, to the next major repair, or to a repeat assessment, when it is checked whether operation can be extended. The same survey gives different conclusions for the three horizons, and the report must name the horizon.

The choice between options rests first of all on economics, but factors that cannot be turned into money are also counted, such as safety or historic value. Repair or strengthening should bring the structure back to a serviceable state. Replacing a single member is permitted when it cannot be strengthened or when that does not pay. In the end it is the designer of the project who decides on strengthening damaged structures or replacing them. Methods of strengthening are covered in the article Strengthening steel structures.

The assessments of single structures add up to an assessment of the frame: the condition of the building is set by the worst of its load-bearing structures. If even one is in an emergency state, the building is treated as such too, and in justified cases parts singled out by function or construction are assessed separately.

What is needed for a verification calculation

The input data are usually with the owner: KM and KMD drawings, the object passport, as-built documents, records of repairs and changes. Missing sections and joint sizes are taken by the survey. The order of collecting data and the documents are covered in the articles Survey procedure and passport and Supervision and monitoring.

POLYGONAL carries out technical inspection, prepares KM and KMD drawings for repair and strengthening (design) and fabricates strengthening and replacement members to drawings (fabrication). Erection on site is done by our partner construction company. The general principles of checking beams, columns and trusses under the current codes are given in the calculation articles: beam, column, truss.

Standards and validity

The condition of structures is assessed under the current procedure for the technical inspection of buildings and structures and the current Methodology for conducting an inspection and recording its results (approved by the order of 06.08.2022 No. 144); the methodological basis for the assessment is also given by DSTU 9273:2024. The verification calculation is made under DBN V.2.6-198:2014 «Steel structures. Design standards» and DBN V.1.2-14:2018 «General principles of reliability and structural safety of buildings and structures» (both with Amendment No. 1).

The techniques described above (assessment on service experience, load testing, reducing loads for a short period, limits from recalculation practice) are methodological in nature. In each case their use is agreed with the current codes and justified in the report. The validity and current edition of each document named need to be checked before they are used in a project.