Reducing manufacturing machine downtime using intelligent document processing on AWS

In the rapidly evolving automotive industry, manufacturers and suppliers must proactively address the challenges of managing vast amounts of data and documentation to streamline operations.

These challenges include paper documents, disparate document versions and folder structures, collaboration with other plants, and finding important information when needed, especially when it comes to mitigating the disruptive impact of machine breakdowns.

Imagine having all that important documentation at your fingertips on the shopfloor, making it much quicker to fix machines when they have a fault.

With this in mind, Nordcloud, an IBM company, worked alongside a big automotive customer in Germany, Austria, and Switzerland (DACH) and Amazon Web Services (AWS) to build a new solution with their UX design, application development and AWS Cloud architecture expertise. Users now have simple access to the information they need exactly when and where they need it.

The aim of this project was to build a solution that let users easily upload and retrieve documents associated with specific production machines – reducing downtime significantly because of instant access to relevant documentation which results in higher production output. However, their highest level goal was to achieve more reliability and sustainability throughout the global production processes.

Now, the key information you need to fix an issue is right in front of you. Documents are secure, accessible only to authorized users through identity and access management. Our software and infrastructure are designed to be user friendly, aiming to make life simple for users in high-pressure manufacturing situations.

Proposed solution

Nordcloud took a comprehensive approach for their digital solution by using and implementing an event-driven microservices architecture.

Let’s break down some key points and advantages of our approach:

• User-centric approach: Conducting workshops in the factory to gather daily use cases and understand user challenges demonstrates a user-centric approach. Like this, our solution aligns closely with real-world needs, increasing the likelihood of user adoption and satisfaction.
• Strategic alignment to AWS: Working alongside AWS indicates a commitment to scalability, reliability, and security for the digital solution.
• Event-driven microservices architecture: Adopting an event-driven microservices architecture offers several advantages:
  • Serverless architecture: By using serverless computing, the team can focus on developing and deploying code with less effort in managing infrastructure, leading to cost savings and increased agility. Future rollouts to additional plants can be executed seamlessly as the infrastructure grows alongside the demand.
  • Reduced integration overhead: Event-driven architecture reduces the need for point-to-point integration between services, simplifying the development and maintenance of the system.
  • Near real-time data processing: Processing data in near real time results in timely insights and actions, facilitating proactive decision-making and responsiveness to changing conditions which is essential for predictive maintenance.

This solution uses a comprehensive AWS serverless cloud architecture, and AWS services such as Amazon Simple Storage Service (Amazon S3), Amazon Textract, Amazon OpenSearch Service, AWS Lambda and AWS Step Functions as well as AWS WAF, Amazon CloudFront, AWS AppSync, Amazon Simple Queue Service (Amazon SQS), Amazon Simple Notification Service (Amazon SNS), Amazon Aurora, AWS CodeBuild, AWS CodePipeline, AWS Cloud Development Kit (AWS CDK), AWS Systems Manager Parameter Store, AWS Secrets Manager, AWS Identity and Access Management (AWS IAM) and Amazon EventBridge.

In Figure 1, you can see a visual representation of the AWS Cloud architecture Nordcloud developed with AWS for this project, along with an overview of the key components and their interactions.

Figure 1. Cloud architecture

Key services include:

• Amazon S3 is object storage built to retrieve any amount of data from anywhere. It serves as the primary object storage mechanism for all documents and offers a cost-effective solution for storing large volumes of files while providing essential features such as security and resiliency. For example, Amazon S3 gateway endpoints are used so only authorized users can access the platform and the required documents (objects).
• Amazon Textract is a machine learning service that automatically extracts text, handwriting, layout elements, and data from scanned documents. Using advanced optical character recognition (OCR) algorithms, Amazon Textract extracts specific data from documents, such as user guidelines and installation documents, which are then stored in Amazon OpenSearch Service.
• Amazon OpenSearch Service securely unlocks real-time search, monitoring, and analysis of business and operational data. It indexes the extracted text content, which facilitates rapid search functionality, and has powerful analytics capabilities.
• AWS Lambda is the backbone of our backend processing. A serverless computing platform, it performs various tasks within the application, including inserting metadata into the database, generating pre-signed URLs for Amazon S3 upload access, and processing simple events between AWS services.
• AWS Step Functions provide a serverless orchestration service for developers to design and coordinate workflows with ease. At its core, AWS Step Functions facilitate the creation of state machines, where each state represents a specific task or action within the workflow. This paradigm aligns perfectly with the modular nature of document processing pipelines. It helps ingest the data, update the indexes, and process the file at the end.

Now let’s review the upload flow within our AWS system:

1. The frontend starts the process by requesting a pre-signed URL, granting temporary access to upload documents directly to an Amazon S3 bucket.
2. Upon receiving the request, the backend service verifies the user’s permissions for document upload and responds accordingly.
3. Armed with the pre-signed URL, the frontend proceeds to upload the documents directly to the designated Amazon S3 bucket.
4. Upon completion of the upload, the frontend receives a response detailing the progress of the upload.
5. Subsequent processing occurs asynchronously. Following the upload, a notification is initiated within Amazon S3 to prompt the Amazon Textract service to commence text extraction from the uploaded document.
6. Amazon Textract reads the content of the Amazon S3 bucket and extracts the texts within the file. This operation unfolds asynchronously, requiring some time to complete.
7. After extracting the texts from the document, the content insertion service updates metadata in the Aurora database. In addition, the extracted content is inserted into the Amazon OpenSearch Service database.

The following visual illustrates the seamless orchestration of a document uploading and processing within our AWS environment, encompassing frontend request handling, Amazon S3 upload, Amazon Textract extraction, and database insertion. It offers an overview of our streamlined workflow driving efficient document management.

Figure 2. Streamlined workflow

Architecture summary

This architecture adopts a modular approach, using a combination of AWS services to streamline document digitization, standardization, storage, indexing, and retrieval processes.

Amazon S3 efficiently stores documents, Amazon Textract extracts text from files, and Amazon OpenSearch Service indexes the extracted content for rapid search functionality.

AWS Step Functions orchestrate backend processing tasks, resulting in seamless interaction between different components of the architecture.

Benefits

The customer received the following benefits from this solution:

• Cost-effectiveness: By using Amazon S3 for storage and AWS Lambda for serverless computing, Nordcloud is optimizing resource use and minimize operational costs.
• Efficiency: The advanced OCR algorithms in Amazon Textract facilitate accurate extraction of text from files, enhancing the overall efficiency of document processing.
• Scalability: The modular design of our architecture results in seamless scalability to accommodate growing data volumes and user demands.
• Performance: Amazon OpenSearch Service provides rapid search functionality and robust analytics capabilities, enhancing the user experience and facilitating informed decision-making.
• User experience: In some cases, the solution provides access to relevant documents more than 60 times faster than legacy practices, which is particularly crucial in stressful machine downtime situations, contributing to an optimal end-user experience.
• Increased production output: The improvement in production performance is further emphasized by the reduction of mean time to repair (MTTR), resulting in increased operational efficiency and therefore higher productivity.

Conclusion

The combination of industry expertise, a deep understanding of user needs, and the strategic integration of AWS technologies presents an exciting opportunity to revolutionize manufacturing processes. By prioritizing a good user experience, driving production process improvements, and enhancing production efficiency, we pave the way for increased productivity and revenue generation.

But this is just the beginning. To delve deeper into how our solution can transform your manufacturing operations, and to explore the full potential of our innovative approach, we invite you to reach out and discover more.

Ready to embark on a journey toward enhanced efficiency and profitability? Contact Nordcloud today to learn more about its solution and take the first step toward unlocking your manufacturing excellence. Check out the AWS Solutions Library for vetted solutions and guidance and for more Intelligent Document Processing solutions, and our AWS for Industrial page to discover how to accelerate design, optimize operations, reinvent supply chains, and transform your business using AWS.

With innovation at our core and your success as our priority, together we can reshape the future of manufacturing. Let’s make it happen.