Enhancing MRO Documentation Through Automated Translation of Non-Standard English to Simplified Technical English Using Offline LLMS

Year : 2025 | Volume : 15 | Issue : 02 | Page : 9-15
    By

    Sundaresan Poovalingam,

  • Janani Varun,

  • Hemanth R. Herle,

  • Anshu Anand,

  1. Distinguished Technologist, Strategic Technology Group, Infosys Limited, Hosur Road, Bangalore, Karnataka, India
  2. Distinguished Technologist, Strategic Technology Group, Infosys Limited, Hosur Road, Bangalore, Karnataka, India
  3. Distinguished Technologist, Strategic Technology Group, Infosys Limited, Hosur Road, Bangalore, Karnataka, India
  4. Distinguished Technologist, Strategic Technology Group, Infosys Limited, Hosur Road, Bangalore, Karnataka, India

Abstract

document.addEventListener(‘DOMContentLoaded’,function(){frmFrontForm.scrollToID(‘frm_container_abs_193426’);});Edit Abstract & Keyword

Maintenance, Repair, and Overhaul (MRO) operations rely heavily on accurate and consistent documentation to ensure operational safety and compliance. However, the presence of non-standard English in technical documents often leads to ambiguity, misinterpretation, and inefficiencies in maintenance processes. This study presents an innovative solution that leverages an offline Large Language Model (LLM) to automatically translate non-standard English in MRO documents into standardized and technically precise language. By integrating predefined linguistic and technical rules, the system ensures clarity, consistency, and adherence to industry standards. The proposed solution operates by processing technical documents containing non-standard terminology alongside a comprehensive set of rule-based transformations. These rules guide the LLM in rephrasing content without altering its technical context, thereby improving readability and compliance. The use of an offline LLM guarantees data security, prevents potential data leaks, and aligns with stringent industry confidentiality requirements. Performance evaluation is conducted by comparing the rephrased output with standardized benchmarks, measuring linguistic accuracy and contextual relevance. The solution utilizes a custom-trained offline LLM for technical language processing, integrated with a rule-based engine to guide translation based on predefined linguistic and technical standards. Natural Language Processing (NLP) tools are employed for syntactic and semantic analysis, while evaluation metrics such as BLEU scores and semantic similarity measures assess accuracy and performance. Additionally, offline deployment ensures data privacy and integrity through a robust security framework.

Keywords: Large language model (LLM), technical documentation standardization, rule-based language transformation, natural language processing, offline LLM deployment

[This article belongs to Journal of Aerospace Engineering & Technology ]

How to cite this article:
Sundaresan Poovalingam, Janani Varun, Hemanth R. Herle, Anshu Anand. Enhancing MRO Documentation Through Automated Translation of Non-Standard English to Simplified Technical English Using Offline LLMS. Journal of Aerospace Engineering & Technology. 2025; 15(02):9-15.
How to cite this URL:
Sundaresan Poovalingam, Janani Varun, Hemanth R. Herle, Anshu Anand. Enhancing MRO Documentation Through Automated Translation of Non-Standard English to Simplified Technical English Using Offline LLMS. Journal of Aerospace Engineering & Technology. 2025; 15(02):9-15. Available from: https://journals.stmjournals.com/joaet/article=2025/view=0


document.addEventListener(‘DOMContentLoaded’,function(){frmFrontForm.scrollToID(‘frm_container_ref_193426’);});Edit

References

1. Knezevic J. Improving quality of maintenance through Simplified Technical English. J Qual Mainten Eng. 2015 Aug 10; 21(3): 250–7.
2. Maruyama T, Yamamoto K. Extremely low-resource text simplification with pre-trained transformer language model. Int J Asian Lang Process. 2020 Mar 30; 30(01): 2050001.
3. Zheng Y, Chen Y, Qian B, Shi X, Shu Y, Chen J. A review on edge large language models: Design, execution, and applications. ACM Comput Surv. 2025 Mar 23; 57(8): 1–35.
4. Kuhn T. A survey and classification of controlled natural languages. Comput Linguist. 2014 Mar 1; 40(1): 121–70.
5. Zhang T, Kishore V, Wu F, Weinberger KQ, Artzi Y. Bertscore: Evaluating text generation with bert. arXiv preprint arXiv:1904.09675. 2019 Apr 21.
6. Wu R. The strategy and realization of enterprise integration. Australia: University of Technology Sydney; 2007.
7. Birru H. Exploring The Use of LLMs In Agile Technical Documentation Writing. Thesis. Sweden:M ̈alardalen University; 2024 Sep 25.
8. Akcali Z, Cubuk HS, Oguz A, Kocak M, Farzaliyeva A, Guven F, Ramazanoglu MN, Hasdemir E, Altundag O, Agildere AM. Automated Extraction of Key Entities from Non-English Mammography Reports Using Named Entity Recognition with Prompt Engineering. Bioengineering. 2025 Feb 10; 12(2): 168.
9. Yeom J, Lee H, Byun H, Kim Y, Byun J, Choi Y, Kim S, Song K. Tc-llama 2: fine-tuning LLM for technology and commercialization applications. J Big Data. 2024 Aug 2; 11(1): 100.
10. Tomenchuk M, Popovych K. Large Language Models and Machine Translation. Věda a Perspektivy. 2024; 11(42): 422–431.

Regular Issue Subscription Review Article
Volume 15
Issue 02
Received 23/04/2025
Accepted 28/04/2025
Published 22/05/2025
Publication Time 29 Days
53

[last_name]

My IP

PlumX Metrics