TLDR: This research paper presents a method to transform traditional, document-based engineering design standards into modular, reusable, machine-interpretable ontologies. By aligning these ontologies with the Industrial Data Ontology (IDO) and using semantic reasoning, the approach enables automated quality assurance for equipment selection, specifically demonstrated through a valve specification use case. This significantly improves efficiency and ensures compliance in engineering design processes, moving towards a future of ‘SMART standards’ that can be interpreted by machines without human intervention.
Engineering design is a complex field that relies heavily on technical specifications and adherence to numerous standards. Traditionally, these standards, product specifications, and data sheets have been primarily document-centric, often existing as paper or PDF files. This document-based approach presents significant challenges in an increasingly digital industrial landscape, hindering efficient decision-making and the development of competitive products and services.
A recent research paper, Machine-interpretable Engineering Design Standards for Valve Specification, addresses this challenge by demonstrating a novel approach to transform information from engineering design standards into modular, reusable, and machine-interpretable ontologies. These ontologies can then be used to automate quality assurance in plant design and equipment selection processes, specifically focusing on valve specification.
The core idea involves using modeling patterns to create modular ontologies from the knowledge embedded in the text and tables of international standards for piping, material, and valve design. These modules are designed to be exchangeable, stored in a W3C compliant format, and interoperable, as they are aligned with the top-level Industrial Data Ontology (IDO), ISO DIS 23726-3.
The Challenge of Traditional Standards
Engineering standards, developed over decades by various organizations like API, ASME, and ASTM, contain a vast body of rules ensuring safe and economical design. However, most are only available in paper or PDF formats, making their content inaccessible to machines. This means engineers must manually search, extract, and interpret information from thousands of complex documents, a process that is time-consuming, prone to error, and costly for organizations.
The paper highlights that a single valve specification can refer to as many as 20 standards and data sheets, with even more references when considering cross-referenced documents. This complexity underscores the immense opportunity for machine-interpretable standards to improve efficiency and quality in engineering design.
Towards SMART Standards
The International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC) have recognized this need, proposing the SMART concept for standards that are machine-applicable, readable, and transferable. This vision includes moving beyond simple digital PDFs to machine-readable XML content, executable content, and ultimately, machine-interpretable content based on information modeling. This research contributes directly to achieving Level 4 SMART standards, where content and relations between elements are expressed through information models.
A Modular Ontology Approach
The authors developed a hierarchical modular ontology structure, with IDO serving as the top-level ontology. This framework allows for the reuse of existing domain-independent and technical ontologies, which are then extended with domain-specific ontologies (e.g., for piping and valves). Crucially, ontologies based on the contents of engineering standards are created as specializations of these domain ontologies.
To ensure consistency and interoperability, the project utilizes modeling patterns, specifically the OTTR (Ontology Templates and Terminology for Reasoning) framework. OTTR provides documented, quality-controlled building blocks for software engineers, enabling the construction and use of large-scale ontologies and knowledge graphs.
The Valve Use Case
To validate their approach, the researchers tested these ontologies on a real-world valve selection process. They instantiated valves in semantic asset models, along with a semantic representation of environmental conditions. This enabled automated validation that a specific Valve Data Sheet (VDS) complies with relevant industry standards and that a particular product type meets the valve specification.
The paper presents a use case involving two valve design specifications (from Equinor and Aker BP) and a vendor product. Using a semantic reasoner, they demonstrated the ability to answer critical questions, such as whether a proposed VDS meets the requirements for a specific line location and if a proposed valve product complies with the VDS. This automated reasoning mimics the rule-checking process typically performed manually by experienced piping engineers.
Also Read:
- Digital Town for Chemical Research: AI Agents Collaborate for Discovery
- The Sandbox Configurator: A Framework for Standardized AI Assessment in Regulatory Environments
Conclusion and Future Outlook
This work provides tangible evidence that machine-interpretable standards are a viable option for Standards Development Organizations (SDOs) and demonstrates how OWL/RDF and alignment to the IDO ontology can realize the vision of machine-interpretable standards. It shows how information in engineering standards can be made machine-interpretable and directly usable by engineers, automating compliance checks and improving efficiency.
The authors recommend that SDOs accelerate their journey towards digitization, making data tables available as ontology modules. They also advocate for the adoption of IDO as a top-level ontology for the industrial standards community to facilitate sharing of semantic product specifications and asset information. This approach promises significant benefits for high-hazard, safety-critical process industries where standards compliance and knowledge sharing are paramount.
