Steel structures — Structural bolting — Test methods to determine loss of pretension from faying surface coatings

Background and Technological Evolution of ISO 18953:2025

With the rapid development of steel structure corrosion protection technology, various high-performance coatings (such as thermal spray coatings, composite coatings, etc.) are widely used in engineering. However, the creep and relaxation characteristics of coating materials under long-term bolt preload can lead to a significant loss of preload at connection nodes, thus affecting the safety and durability of the structure. Traditional design methods are often based on experience or simplified assumptions, lacking systematic experimental data support. Based on years of research and engineering practice, the International Organization for Standardization (ISO) Technical Committee TC 167 (Steel and Aluminum Structures) officially released ISO 18953:2025, "Steel structures—Structural bolted connections—Test methods for determining preload loss due to friction surface coatings," in 2025. This standard fills the gap in the long-term performance evaluation of coated bolted connections, providing a scientific and unified testing framework for engineering practice.

Core Concepts and Scope of Application of this Standard

The core objective of this standard is to quantitatively assess the loss of bolt preload over time due to the presence of a friction surface coating. The standard explicitly defines its scope as preloaded high-strength bolt connections used in steel structure engineering, particularly when the coating thickness is sufficient to affect preload in the short term, or when the coating material may exhibit significant deformation under sustained load (creep-prone materials). The standard specifically states that the presence of other materials with stiffness significantly lower than steel (such as insulation materials) within the bolt clamping area is not included in this test method. This clarifies the boundaries of the standard, focusing on the mechanical behavior of the coating itself.

The standard provides precise definitions for key terms, such as:

• Duplex coating: A coating composed of two different materials to provide higher corrosion resistance. It can be two electroplated metal layers or an electroplated metal layer plus an organic coating.
  Thermal spray coating: A coating formed on a surface by depositing finely subdivided metallic or non-metallic materials in a molten or semi-molten state through a high-temperature process, followed by cooling. The standard emphasizes that when the coating thickness applied to any friction surface exceeds 100 micrometers or the coating is composed of a creep-prone material, the potential preload loss of the preloaded bolted connection should be checked. This provides engineers with a clear quantitative threshold for determining when the tests specified in this standard are required. Chapter 4 of the standard explains the basic principles of the test. Coatings are applied to the surfaces of steel preloaded bolted connections for corrosion protection, appearance modification, or to improve anti-slip properties. The suitability of the selected coating or coating system for preloaded bolted connections depends on the system reserve that may arise from the tightening procedure (i.e., the actual preload level relative to the nominal minimum preload) and the preload loss over the service life of the structure. For example, if a combination method or corner method is used for preload, approximately 30% of the system reserve (approximately 1.3 times the nominal minimum preload) can be expected. This test method correlates preload loss with coating thickness and properties, but does not assess the effect of coating thickness and properties on the coefficient of friction of the anti-slip joint friction surfaces.

  Key Variables Affecting Test Results

  The applicability of the test results is limited to cases where all key variables are consistent with the specimen.

  Section 4.2 of the standard details the variables that should be considered to have a significant impact on the test results, including:
  Bolts of different performance grades have different preload levels and yield behaviors.

  Variable Category	Specific Content	Impact Description
  Coating Characteristics	Surface treatment, coating composition, treatment of each coating in a multilayer system, coating thickness, curing procedure and parameters, minimum time interval between coating and bolt preload	Directly affects the compressive stiffness, creep potential, and bonding performance with the substrate of the coating, and is a core factor determining the rate and total amount of preload loss.
  Fastener Characteristics	Size, coating characteristics, and performance grade of bolts, nuts, and washers; number and configuration of washers (if required)	Affects stress distribution, contact pressure, and overall system stiffness.
  Connection Construction	Spacing and row spacing of bolts in the connection (if applicable); preload level from initial tension to the end of the test; preload sequence (if applicable)	In multi-bolt connections, the interaction between bolts and the preload sequence affect the final preload state of individual bolts and subsequent loss patterns.

  Time Effect of Preload Loss and Evaluation Methods

  The standard introduces an evaluation framework based on time logarithm. The test method should at least establish a **preload loss-time logarithm graph**, which can be extrapolated to the expected life of the connection (see Figure 1). The minimum test duration is **14 days**, after which extrapolation to the design life of the structure is permitted. This semi-logarithmic coordinate approach, based on the principle that the creep/relaxation behavior of many materials is approximately linearly related to the logarithm of time under constant stress, makes it possible to predict long-term performance through short-term tests.

  ---

  Test Equipment, Preparation, and Data Acquisition Specifications

  Test Instrument Selection: Bolt Strain Gauges and Circular Load Cells

  Chapter 5 of the standard specifies two mainstream methods for long-term bolt tension measurement: bolt strain gauges and circular load cells (also known as load washers or through-hole load cells). When using other methods, comparison and verification with one of these two methods are necessary.

  Before each test series, the correspondence between the measured force and the applied force should be checked. The standard states that other load indication technologies, such as ultrasonic sensors, are more suitable for measuring bolt forces within the elastic range and are not practical for fully pre-tightened bolts that have exceeded the yield point. Test Plate Preparation and Test Execution All surfaces of the test plate should be free of burrs, grooves, substrate defects, and excessively thick coating deposits. Burrs should be sanded smooth before coating. The surface treatment and coating to be tested should be applied to the friction surfaces of the specimen in a manner consistent with the intended structural application. Surface roughness and dry film thickness should be recorded according to ISO 8503-1, ASTM B659, or SSPC-PA 2. The curing procedure should also be recorded in detail. After coating curing, any localized excessive coating deposits exceeding the coating manufacturer's process allowance should be sanded smooth, taking care not to excessively remove coating or damage adjacent coated surfaces. During test execution, the plate should be aligned with the edges. Install bolts, nuts, and washers of the selected specifications. Preload the bolt assembly using the selected preload procedure according to one of the methods described in Annex I, J, K, L, or M of ISO 17607-6:2023. The data acquisition system should be started at the beginning of the preload operation and run continuously until the end of the test (not less than 14 days). The standard specifically notes that after bolt preload, the measured preload curve will typically drop significantly from its peak in the first few seconds (approximately 3 seconds) after tightening. This drop is not entirely related to component loosening, but also to nut rotation and bolt thread elastic recovery upon wrench removal, and therefore is generally **not considered** as a preload loss.

  Data Acquisition Frequency Requirements

  The rate of preload loss is high during the first 24 hours. Therefore, increasing the data acquisition frequency is crucial in the first few minutes and hours after preload. The standard specifies: For the first 30 minutes of testing, the sampling frequency should not be lower than 1 Hz. After 30 minutes, the sampling frequency can be reduced. At least once every 30 minutes during the first 24 hours, and twice daily for the remainder of the test.

  ---

  Detailed Explanation of Two Standard Test Procedures

  Test A: Standardized Single Bolt Test

  Purpose: Focusing on single bolt connections, directly assessing the impact of coating stiffness, thickness, and long-term performance on bolt preload loss over time. Used to study the influence of coating variables themselves.

  Specimen Requirements: A sample consists of 5 identical specimens (sheet metal and fasteners from the same production batch). The specimen consists of two square plates with a standard-sized hole in the center.

  Each side of the square plate is coated according to the required process. Plate dimensions and hole dimensions are selected according to the standard tables based on bolt diameters (see Tables 1 and 2). Results Processing: The average preload loss of the five specimens at the design life (or predetermined time point) is the reported result of this test. If the preload loss of one test is significantly lower than the average of the other four, it can be determined whether it is an outlier according to the statistical criteria given in the standard (using a threshold of 1.71). If there is only one outlier, it can be removed and the average of the remaining four can be calculated; if there are more than one outlier, a new set of five tests should be conducted. Test B: Representative Multi-Bolt Construction Test Purpose: To determine the preload loss due to compaction, creep, and relaxation on specimens simulating steel structure connection constructions. The preload loss of bolt groups can be evaluated, including the preload sequence and the effects of multilayer plates, and considering potential interactions within the bolt group.

  Specimen Requirements: A single specimen must contain at least 4 bolts. The number of plate layers, plate thickness, surface treatment and coating, bolt diameter, length, grade, spacing and row spacing, preload method, and level in the clamping should all be representative of an actual steel structure connection. Figure 3 shows an example of an eight-bolt connection specimen.

  Preload Sequence: Bolt preload should systematically begin from the stiffest part of the specimen to minimize the slack of adjacent bolts. Typically, it begins from the middle of the connection and proceeds towards the free edge. The data acquisition system should remain active during the preload operation to monitor changes in bolt tension.

  Instrument Type	Working Principle and Installation	Advantages	Disadvantages and Precautions	Calibration Requirements
  Bolt Strain Gauge	Drill a concentric hole along the bolt axis in the bolt head, insert the strain gauge, and fix it with epoxy resin. Measure the elastic strain of the bolt shank to calculate the axial force.	The sensor cost is relatively low. Directly measures the bolt shank strain, with less susceptibility to external interference.	1. The installation process is complex, requiring drilling, which may weaken the bolt cross-section. 2. Usually for single use (cannot be reused after pre-tightening). 3. During long-term testing, the creep of the adhesive may affect the reading. 4. Not suitable for small-diameter bolts (drilling has a significant impact). Each instrumented bolt needs to be individually calibrated on a universal testing machine to establish a force-voltage coefficient. The calibration force range should be lower than the nominal yield strength of the bolt.
  Ring Load Sensor	In series with the bolt (under the bolt head or between the nut/washer and the substrate), it has at least four built-in strain gauges forming a full-bridge circuit to measure axial loads.	1. Reusable, requiring only periodic calibration. 2. Easy to install, does not damage the bolt. 3. Direct measurement, unaffected by bolt plastic deformation.	1. High initial purchase cost. 2. The range selection must match the expected force value; otherwise, the measurement accuracy is poor at low force values. 3. Increasing the clamping length will change the axial stiffness of the bolt, potentially affecting the measurement value.	Requires periodic calibration, with intervals not exceeding 12 months.

  Application Scenarios

  Comparison Dimensions	Experiment A (Single Bolt)	Experiment B (Multiple Bolts)
  Core Objectives	Isolate and quantify the impact of coating material properties (thickness, stiffness, creep) on preload loss.	Evaluate the preload loss in actual connection constructions under combined factors (including bolt interaction, preload sequence, multi-layer plates).
  Number and Composition of Specimens	Five identical single-bolt double-plate specimens constitute the statistical sample.	One representative structural specimen containing at least four bolts.
  Performance comparison and screening of coating materials, establishing a database of the relationship between coating parameters and losses.	Design verification and performance evaluation of specific engineering connections, more closely reflecting actual working conditions.
  Result Representativeness	Reflects the influence of the coating's inherent properties, and the results can be generalized to other connection forms using the same coating.	Reflects the comprehensive performance under specific structures, with more targeted results that directly serve specific designs.
  Cost and Complexity	Relatively low cost, simple specimens, but requires a certain number of replicates.	High cost, complex specimens, simulating real structures, but only requires a single (or a few) specimens.

  ---

  Test Results, Reports, and Standard Implementation Recommendations

  Results Analysis and Report Preparation

  Based on the collected data, a **preload loss-time logarithmic relationship graph** should be plotted. Linear extrapolation from the measurement point on day 14 to the structural design life is permitted to estimate the preload loss at the design life.

  The test report should include at least the following items: sample identification, standard number, method used (Test A or B), results (including a list of key variables, specimen geometry, actual preload of all bolts in each specimen at the end of the preload process, measurement data and time intervals during the test, average preload loss assessment at the structural design life, applicable coating thickness range), any deviations from the procedure, any observed anomalies, and the test date. Engineering Implementation Recommendations 1. **Test Necessity Judgment:** When the coating thickness on the friction surface is greater than 100 micrometers or when using known creep-prone coating materials (such as certain organic thick coatings or soft metal coatings), it is strongly recommended to conduct a preload loss assessment according to this standard. 2. **Test Method Selection:** - **New Material R&D and Certification:** Test A should be prioritized to systematically study the long-term performance of coatings under different thicknesses and curing processes, and to establish a design database.
  - **Verification of Key Nodes in Major Projects:** **Test B** should be used to simulate actual connection details (plate thickness, number of layers, bolt layout, preload sequence) to obtain design parameters that are closest to reality.

  3. **Instrument Selection Recommendations:**
  - For long-term, large-scale comparative testing of coating performance, considering cost, **bolt strain gauges** can be used, but drilling and bonding processes must be strictly controlled, and their single-use nature must be taken into account.
  - For applications requiring high precision, reusability, and relatively fixed bolt specifications, investing in **ring load cells** is a better choice. It is essential to select an appropriate range based on the bolt preload magnitude and establish a regular calibration system.

  4. **Data Application and Design Compensation:** Combine the percentage of preload loss obtained from the test with the design preload level to calculate the minimum possible effective preload of the structure during its service life.

  In anti-slip connection design, this effective preload is used to calculate the anti-slip bearing capacity. If necessary, measures such as **over-tensioning** (adding a certain percentage to the standard preload to compensate for expected losses) or **periodic retightening** (for maintainable parts) can be considered to ensure the long-term performance of the connection. 5. **Quality Control Correlation**: Correlate the test results of this standard with coating application specifications (such as coating thickness control range and curing conditions). Ensure that the coating quality (especially thickness and curing degree) applied on-site is not inferior to the coating quality of the test specimens; otherwise, the test results will be unsafe. The publication of ISO 18953:2025 marks a significant step forward in the design of bolted connections for steel structures, moving from static strength control to long-term performance control that incorporates time factors. Through scientific and systematic testing, the preload loss caused by the coating is quantified, providing crucial technical support for ensuring the safety and durability of steel structures in complex corrosive environments.