Thomas Smoker

Much textual engineering knowledge is captured in tables, particularly in spreadsheets and in documents such as equipment manuals. To leverage the benefits of artificial intelligence, industry must find ways to extract the data and relationships captured in these tables. This paper demonstrates the application of an ontological approach to make the classes and relations held in spreadsheet tables explicit. Ontologies offer a pathway because they use formal descriptions to define… Show more

Much textual engineering knowledge is captured in tables, particularly in spreadsheets and in documents such as equipment manuals. To leverage the benefits of artificial intelligence, industry must find ways to extract the data and relationships captured in these tables. This paper demonstrates the application of an ontological approach to make the classes and relations held in spreadsheet tables explicit. Ontologies offer a pathway because they use formal descriptions to define machine-interpretable definitions of shared concepts and relations between concepts. We illustrate this with two case studies on a failure modes and effects analysis (FMEA) table. Our examples demonstrate how the relationship between rows and columns in a table can be represented in logic for FMEA entries, thereby allowing the same ontology to ingest instance data from the IEC 60812:2006 FMEA Standard and a real industrial FMEA. We give relationships in the FMEA and asset hierarchy spreadsheets an explicit representation, so that OWL-DL reasoning can infer final failure effects at the system level from component failures. The prototype ontologies described in this paper are modular and aligned to a top level ontology, and hence can be applied to other use cases. Our contribution is showing that engineers can make data captured in commonly used spreadsheet tables machine readable using a FMEA ontology. Show less

Much textual engineering knowledge is captured in tables, particularly in spreadsheets and in documents such as equipment manuals. To leverage the benefits of artificial intelligence requires industry to find ways of extracting the data and relationships captured in these tables. This paper demonstrates the application of an ontological approach to make explicit the classes and relations held in spreadsheet tables. Ontologies offer a pathway, as they are used to define machine-interpretable… Show more

Much textual engineering knowledge is captured in tables, particularly in spreadsheets and in documents such as equipment manuals. To leverage the benefits of artificial intelligence requires industry to find ways of extracting the data and relationships captured in these tables. This paper demonstrates the application of an ontological approach to make explicit the classes and relations held in spreadsheet tables. Ontologies offer a pathway, as they are used to define machine-interpretable definitions of shared concepts, and relations between concepts, using formal descriptions. We illustrate this with a case study on a Failure Modes and Effects Analysis (FMEA) table, using an example from the IEC 60812 FMEA Standard. Our example demonstrates how the relationship between rows and columns in a table can be represented in logic. Further, we give relationships in the FMEA and asset hierarchy spreadsheets an explicit representation, so that OWL-DL reasoning can infer final failure effects at the system level from component failures. The prototype ontologies described in this paper are modular and aligned to a Top Level Ontology, and hence can be applied to other use cases. Our contribution is to show engineers needing to make data captured in spreadsheet tables machine readable, how ontologies can be applied using a real example. Show less

Maintenance has largely remained a human-knowledge centered activity, with the primary records of activity being textbased maintenance work orders (MWOs). However, the bulk of maintenance research does not currently attempt to quantify human knowledge, though this knowledge can be rich with useful contextual and system-level information. The underlying quality of data in MWOs often suffers from misspellings, domain-specific (or even workforce specific) jargon, and abbreviations, that prevent… Show more

Maintenance has largely remained a human-knowledge centered activity, with the primary records of activity being textbased maintenance work orders (MWOs). However, the bulk of maintenance research does not currently attempt to quantify human knowledge, though this knowledge can be rich with useful contextual and system-level information. The underlying quality of data in MWOs often suffers from misspellings, domain-specific (or even workforce specific) jargon, and abbreviations, that prevent its immediate use in computer analyses. Therefore, approaches to making this data computable must translate unstructured text into a formal schema or system; ie, perform a mapping from informal technical language to some computable format. Keyword spotting (or, extraction) has proven a valuable tool in reducing manual efforts while structuring data, by providing a systematic methodology to create computable knowledge. This technique searches for known vocabulary in a corpus and maps them to designed higher level concepts, shifting the primary effort away from structuring the MWOs themselves, toward creating a dictionary of domain specific terms and the knowledge that they represent. The presented work compares rules-based keyword extraction to data-driven tagging assistance, through quantitative and qualitative discussion of the key advantages and disadvantages. This will enable maintenance practitioners to select an appropriate approach to information encoding that provides needed functionality at minimal cost and effort. Show less

Companies are investing heavily in predictive maintenance algorithms without considering how the predictions will be validated. When components are removed, the observations of the maintenance technicians about their state (failed or not) and the failure mode are crucial to this validation process and to developing accurate component reliability distributions. Despite years of effort to get maintenance technicians to collect data that is usable and useful to engineers, either by trying to… Show more

Companies are investing heavily in predictive maintenance algorithms without considering how the predictions will be validated. When components are removed, the observations of the maintenance technicians about their state (failed or not) and the failure mode are crucial to this validation process and to developing accurate component reliability distributions. Despite years of effort to get maintenance technicians to collect data that is usable and useful to engineers, either by trying to enforce the use of codes or apply management controls, little progress has been made. Advances in cognitive computing processes such as text mining, natural language processing and knowledge representation hold the key to solving this problem. The purpose of this paper is to explain key concepts in text mining, knowledge representation and ontology development in a way that is accessible to reliability and maintenance engineers. We illustrate, using a conveyor system as an example, how these concepts can be applied. Our aim is to convince the reader of the value of investing time to understand and develop cognitive computing methods. Sooner, rather than later, these concepts will be used to translate manually entered maintenance work order data to support validation of condition based maintenance predictions and generation of near real time reliability distributions. Show less