Digitize-PID: Automatic Digitization of Piping and Instrumentation Diagrams

Paliwal, Shubham; Jain, Arushi; Sharma, Monika; Vig, Lovekesh

doi:10.1007/978-3-030-75015-2_17

Cited by 20 publications

(3 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One approach is to use deep learning to recognize symbols, as presented in an article about capturing information from piping and instrumentation diagrams (PI&Ds) presented by Rahul et al (2019). Paliwal et al (2021) also used deep learning as the chosen methodology for symbol classification and recognition. Cordella and Vento (2000) stated that template matching in the classification phase of symbol recognition is the preferred technique 46% of the time.…”

Section: Template Matchingmentioning

confidence: 99%

Knowledge-based engineering and computer vision for configuration-based substation design

2023

View full text Add to dashboard Cite

Introduction: As the increase in electrification poses new demands on power delivery, the quality of the distribution system is paramount. Substations are a critical part of power grids that allow for control and service of the electrical distribution system. Substations are currently developed in a project-based and manually intensive manner, with a high degree of manual work and lengthy lead times. Substations are primarily sold through tenders that are accompanied by an inherent need for engineering-to-order activities. Although necessary, these activities present a paradox as tender processes must be agile and fast. To remedy this shortcoming, this article outlines a knowledge capture and reuse methodology to standardize and automate the product development processes of substation design.Methods: A novel framework for substation design is presented that implements knowledge-based engineering (KBE) and artificial intelligence methods in computer vision to capture knowledge. In addition, a product configuration system is presented, utilizing high-level CAD templates. The development has followed the KBE methodology MOKA.Results: The proposed framework has been implemented on several company cases where three (simplified) are presented in this paper. The framework decreased the time to create a 3D model from a basic electric single line diagram by performing the identification and design tasks in an automated fashion.Discussion: Ultimately, the framework will allow substation design companies to increase competitiveness through automation and knowledge management and enable more tenders to be answered without losing engineering quality.

show abstract

Section: Template Matchingmentioning

confidence: 99%

Knowledge-based engineering and computer vision for configuration-based substation design

2023

View full text Add to dashboard Cite

show abstract

“…We use Tesseract 4.0.0 (Smith, 2007) to extract text from the detected regions and tag them to the corresponding roles. The results of the chart element extractions and text role region extractions are described in Appendix F. We define rules to identify (i) the origin, axes and chart region from the detected chart lines by a line detection algorithm (Paliwal et al, 2021), (ii) location of the legend previews and their styles (color and pattern) using the detected legend-labels, and (iii) chart elements (bars, dots, lines) which are regions matching with each legend preview style. We extract a schema (table) from the above available chart information, by filtering noise (Appendix C) and extracting the data series elements (Appendix E).…”

Section: Chart Schema Extractionmentioning

confidence: 99%

Program Synthesis for Complex QA on Charts via Probabilistic Grammar Based Filtered Iterative Back-Translation

Bhaisaheb,

Paliwal,

Patil

et al. 2023

Findings of the Association for Computational Linguistics: EACL 2023

View full text Add to dashboard Cite

Answering complex reasoning questions from chart images is a challenging problem requiring a combination of natural language understanding, fine-grained perception, and analytical reasoning. Current chart based Question Answering (QA) approaches largely address structural, visual or simple data retrieval type questions with fixed-vocabulary answers and perform poorly on reasoning queries. We focus on answering realistic, complex, reasoningbased questions where the answer needs to be computed and not selected from a fixed set of choices. Our approach employs a neural semantic parser to transform Natural Language (NL) questions into SQL programs and execute them on a standardized schema populated from the extracted chart contents. In the absence of program annotations, i.e., in a weak supervision setting, we obtain initial SQL predictions from a pre-trained CodeT5 semantic parser and employ Filtered Iterative Back-Translation (FIBT) for iteratively augmenting our NL-SQL training set. The forward (neural semantic parser) and backward (language model) models are initially trained with an external NL-SQL bootstrapping data. We iteratively move towards the required NL query distribution by generating NL questions from the synthesized SQL programs using a Probabilistic Context-Free Grammar (PCFG) where the production rule probabilities are induced to be inversely proportional to the probabilities in the training data. We filter out the generated NL queries with mismatched structure and compositions. Our FIBT approach achieves State-of-the-Art (SOTA) results on reasoning-based queries in the PlotQA dataset yielding a test accuracy of 60.44%, superseding the previous baselines by a large margin.

show abstract

“…In these steps, input data from various sources will be used to create a graph model of the process system. Because graph modeling is both easy and flexible, it is commonly used to describe extracted models from P&ID files [52,53]. In addition, under the umbrella of graph theory, there are numerous available algorithms, theories, and tools that can be utilized in the development, study and evaluation of the model.…”

Section: E Graph Processing (B4 C2)mentioning

confidence: 99%

A Methodology for Generating a Digital Twin for Process Industry: A Case Study of a Fiber Processing Pilot Plant

et al. 2022

View full text Add to dashboard Cite

Digital twins are now one of the top trends in Industry 4.0, and many companies are using them to increase their level of digitalization, and, as a result, their productivity and reliability. However, the development of digital twins is difficult, expensive, and time consuming. This article proposes a semiautomated methodology to generate digital twins for process plants by extracting process data from engineering documents using text and image processing techniques. The extracted information is used to build an intermediate graph model, which serves as a starting point for generating a model in a simulation software. The translation of a graph-based model into a simulation software environment necessitates the use of simulator-specific mapping rules. This paper describes an approach for generating a digital twin based on a steady state simulation model, using a Piping and Instrumentation Diagram (P&ID) as the main source of information. The steady state modeling paradigm is especially suitable for use cases involving retrofits for an operational process plant, also known as a brownfield plant. A methodology and toolchain is proposed, consisting of manual, semi-automated and fully automated steps. A pilot scale brownfield fiber processing plant was used as a case study to demonstrate our proposed methodology and toolchain, and to identify and address issues that may not occur in laboratory scale case studies. The article concludes with an evaluation of unresolved concerns and future research topics for the automated development of a digital twin for a brownfield process system.

show abstract

Digitize-PID: Automatic Digitization of Piping and Instrumentation Diagrams

Cited by 20 publications

References 12 publications

Knowledge-based engineering and computer vision for configuration-based substation design

Knowledge-based engineering and computer vision for configuration-based substation design

Program Synthesis for Complex QA on Charts via Probabilistic Grammar Based Filtered Iterative Back-Translation

A Methodology for Generating a Digital Twin for Process Industry: A Case Study of a Fiber Processing Pilot Plant

Contact Info

Product

Resources

About