Research map / Large language models / AI and society
Automating the Expansion of Instrument Typicals in Piping and Instrumentation Diagrams (P&IDs)
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes a computer vision and domain-rule pipeline (PSENet text detection, PP-OCR recognition, Hough line detection) that automatically expands Instrument Typicals in P&IDs into full instrument lists, evaluated on McDermott EPC projects.
Problem
Manual expansion of Instrument Typical assemblies referenced in P&IDs into complete instrument item lists (Instrument Index documents) is time-consuming and error-prone in EPC projects, where P&IDs are often only available as PDFs.
Approach
Combines deep-learning text detection and recognition with Hough-transform border-line detection on legend sheets and a rule-based procedure that associates typical numbers, instrument types, and tag numbers to generate all utilized instruments, including excluding optional instruments absent from the P&ID.
Key results
- Overall expansion precision 99.24% and recall 98.25% (1574 correct, 28 missed, 12 wrong instruments).
- Legend information extraction achieved 100% accuracy on all typical groups across the three projects.
- Estimated potential yearly savings of 1,500,000 USD for a single McDermott office under stated assumptions.
Limitations
- Legend extraction customized to McDermott's legend format; other EPC companies' formats may need adjustment.
- Errors from text detection misses and Euclidean-distance association of typical numbers with wrong instruments.
- Evaluation limited to 55 P&IDs from 3 projects by the industry authors.
Experimental setup
Datasets: 315 McDermott industrial P&IDs (PSENet training) McDermott legend sheets
Benchmarks: 55 P&IDs and 4 legend sheets from 3 McDermott EPC projects
Metrics: precision recall legend group extraction accuracy
Builds on: PSENet PP-OCR Canny edge detection Probabilistic Hough Transform
Closest papers by meaning
- ArchCAD-400K: A Large-Scale CAD drawings Dataset and New Baseline for Panoptic Symbol Spotting (NeurIPS 2025 · similarity 0.77)
- Text2CAD: Generating Sequential CAD Designs from Beginner-to-Expert Level Text Prompts (NeurIPS 2024 · similarity 0.77)
- DiagnosticIQ: A Benchmark for LLM-Based Industrial Maintenance Action Recommendation from Symbolic Rules (NeurIPS 2026 · similarity 0.76)
- IndustryCode: A Benchmark for Industry Code Generation (NeurIPS 2026 · similarity 0.76)
- Automated Defect Report Generation for Enhanced Industrial Quality Control (AAAI 2024 · similarity 0.76)
- Structured Document Generation for Industrial Equipment (AAAI 2025 · similarity 0.75)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.