AIE aims to address a mountain of challenges faced by operators, who often have large volumes of historical asset data sitting in unusable formats.
Over the years, P&ID and Isometric diagrams, as well as maintenance and inspection data have been manually generated, scanned, and stored as image/PDF files. Asset data digitilisation is essential for before it can be usefully analysed. Despite the obvious advantages of having accessible and readable data, many companies do not embrace this, due to the huge amount of manual processing which they believe is necessary, along with the associated manpower and costs. In many cases, they fail to convince themselves of the trade-off and undervalue the power of their historical asset data and persist with error-prone manipulation of results.
INTRODUCING ARTIFICIAL INTELLIGENCE TECHNIQUES
Artificial Intelligence (AI) can however minimise the pain that engineers face when working with such information and can maximize efficiency and realise significant savings, both from the data digitalisation transition process, and the efficiencies that can be seen from digital data driven decision making.
Using Artificial Intelligence, AIE can accomplish tasks within minutes, that would otherwise need several manhours to address. We are working on the capability to automatically transform unstructured engineering diagrams and inspection data into digital information, using Artificial Intelligence. This significantly reduces the manual work necessary for the digitalisation of big data and understanding them. All the tasks will be automated, and corresponding load sheets will be prepared as per the format required to create asset registers and data upload for a number of IT platforms, including AIE’s proprietary Veracity software.
Despite the recent significant advances in image processing, and deep neural networks, automatic analysis and processing of these engineering drawings is still far from being complete. To the best of our knowledge, AIE will be one of the first to propose a novel end-to-end automatic data extraction system from P&ID, Inspection / Maintenance Reports, and Isometric diagrams.
AIE UTILISES MACHINE LEARNING
AIE applies machine learning technology, which is a popular AI method used in data analysis, to improve processing speed and accuracy. This digital solution enhances the performance of the recognition automation process by decreasing the rates of both misrecognition and nonrecognition. With the intelligent digitized drawings generated by the automatic conversion system, the consistency of the original design is maintained, and the problems experienced with the traditional and manual input method by engineering companies are solved through finetuning and validation processes.
The sample diagram below shows how asset tags could be extracted automatically using AI. Similarly, AIE leverages its machine learning and AI technology to collect and manipulate inspection data, digitize P&IDs, Isometrics and Inspection Reports to be uploaded in its Veracity software for further data analytics.
Figure 1: Preview from veracity inspection
Figure 2: The output using artificial intelligence.