TCR Engineering Lab using Artificial Intelligence

TCR Engineering uses artificial intelligence at its Mahape laboratory in Navi Mumbai, NABL accredited to ISO/IEC 17025 (NABLT0726MH18640), for the records around each test, not the test itself. AGAAS, the group's AI agent and automation architecture, triages samples, maps them to the NABL scope and flags proficiency-testing variance. None of it performs a test or signs a result.

• Sample receipt: classification, scope mapping and NABL deviation flagging (LIMS triage)
• Compliance: regulatory tracking for NABL, BIS, IBR and AERB, with document control
• Quality: Z-score variance flagging in proficiency testing and inter-laboratory comparison, and calibration alerts
• Radiography: film-to-digital archival with weld-joint metadata
• The architecture is described on the AGAAS page; the rest of the post describes where AI can help a testing laboratory more broadly

TCR Engineering Services' materials testing labs are critical to the safety, quality, and performance of materials used in various industries, from aerospace to petrochemicals. In our labs, laboratory technicians perform a wide range of tests to evaluate material properties, often in compliance with standards like ASTM and NACE. While the tasks require precision and expertise, they can also be time-consuming and prone to human error. Enter Artificial Intelligence (AI), a technology that is changing the way materials testing labs operate.

AI is not replacing laboratory technicians; it supports them. By automating routine tasks, assisting result interpretation and providing analytics, AI can take clerical load off a laboratory so that technicians spend their time on the tests. With AI as an aid, technicians can meet the growing demands of modern industries while working to ASTM and NACE standards.

Here is how AI is empowering laboratory technicians and improving the evaluation of test results:

1. Automation of Routine Tasks

AI-powered systems can automate routine processes, such as sample preparation, data logging, and initial analysis. Automation reduces the workload on technicians, enabling them to focus on more complex tasks. For example:

• Automated Image Analysis: AI can analyse images from Scanning Electron Microscopes (SEMs) or Optical Microscopes to identify microstructural features like grain boundaries, inclusions, or cracks.
• Machine Learning for Pattern Recognition: AI models trained on historical test data can identify patterns in test results, flagging anomalies that require attention.

2. Checking Test Results Against Standards

AI can check test results against established standards, consistently:

• ASTM Standards Compliance: AI algorithms can cross-reference test results against ASTM standards to check compliance. For instance, in mechanical testing, AI can analyse stress-strain curves to automatically calculate parameters like tensile strength, yield strength, and elongation within prescribed tolerances.
• Standards for Corrosion Testing: In corrosion testing, AI can evaluate data from Electrochemical Impedance Spectroscopy (EIS) or salt spray tests to ASTM B117 and ISO 9227, flagging deviations in corrosion rates and other critical metrics against the method's limits.

3. Predictive Analytics for Material Behaviour

AI can predict material performance based on test results and historical data, offering valuable insights for technicians:

• Failure Prediction: AI models can identify early signs of material degradation, such as High-Temperature Hydrogen Attack (HTHA) or Stress Corrosion Cracking (SCC), using data from non-destructive testing methods like Phased Array Ultrasonic Testing (PAUT).
• Real-Time Corrosion Rate Analysis: AI can process corrosion mapping data in real-time, enabling technicians to forecast material lifespan and recommend proactive maintenance.

4. Streamlined Data Management and Reporting

Laboratories generate vast amounts of data during testing. AI can organise and manage this data:

• Centralised Data Systems: AI-driven Laboratory Information Management Systems (LIMS) can store, retrieve, and analyse test data, supporting traceability and compliance.
• Automated Reporting: AI can generate detailed test reports, complete with graphs, statistical analyses, and conformity checks against ASTM or NACE standards, taking manual work off technicians.

5. AI-Assisted Quality Control

AI enhances quality control by identifying subtle defects or irregularities that might escape the human eye:

• Defect Detection in Welds: AI-driven image analysis can detect porosity, slag inclusions, or cracks in weld radiographs, for review against the acceptance standard by a qualified interpreter.
• Surface Analysis: AI algorithms can analyse surface roughness data from profilometers, for comparison with the specified surface finish requirement.

6. Training and Decision Support

AI acts as a virtual assistant, providing real-time guidance to technicians:

• Interactive Training Tools: AI can simulate test scenarios, helping technicians gain hands-on experience in analysing complex data sets.
• Decision Support Systems: AI offers actionable insights, recommending corrective actions or alternative test methods when discrepancies arise.

7. Accelerating Research and Development

AI can support R&D by analysing vast datasets to uncover trends and correlations:

• Material Property Optimisation: AI can suggest optimal compositions or heat treatment processes to achieve desired material properties.
• Standards Updates: AI systems can monitor changes in ASTM or NACE standards, so that lab practices can be kept current.

Real-World Application Example

In a materials testing lab evaluating pipeline materials, AI might assist technicians by:

• Analysing ultrasonic data for detecting internal flaws, checked against ASTM E213 (Standard Practice for Ultrasonic Testing of Metal Pipe and Tubing).
• Processing corrosion coupon data in line with NACE SP0775 (Preparation, Installation, Analysis, and Interpretation of Corrosion Coupons in Oilfield Operations).

By automating these steps, AI can reduce transcription and calculation errors; the result is still reviewed and signed by the laboratory's authorised staff.

FAQ

This article was written and edited by the TCR Newsroom, the editorial desk of the TCR group, covering materials testing, inspection and asset integrity work across TCR’s laboratories in India and Saudi Arabia and its alliances in Qatar, Japan and beyond.