Boosting Productivity in Manufacturing with AI

Boosting Productivity in Manufacturing with AI

Table of Contents

1. Introduction
2. What is Machining?
3. Challenges in the Machining Industry
4. AI Solution for Cost Estimation
   • Kabuku's Data Management Tool
   • Uploading Blueprints and Estimation
   • Detection of Processing Steps
   • Benefits of the Tool
5. Building the Tool with Google Cloud AI
   • AutoML Vision for Machining Operations Detection
   • Training the Model
   • Integration with TensorFlow.js
6. Continuous Learning with User Feedback
   • UI for Expert Feedback
   • Rerunning the ML Training Process
   • Cloud AI Platform Pipelines for Continuous Training
7. The Power of Active Learning
8. Conclusion
9. Get Started with Google Cloud AI

Article

AI in Manufacturing: Increasing Productivity with Kabuku's Data Management Tool

In today's fast-paced manufacturing industry, efficiency and productivity are key factors for success. One area that has historically posed challenges is cost estimation in the machining process. The manual task of determining the individual cuts and holes required for each part can be time-consuming and act as a bottleneck for machine shops. However, Kabuku, a startup focused on modernizing manufacturing, has developed a solution using AI to revolutionize the cost estimation process.

What is Machining?

Before delving into the solution, let's first understand what machining entails. Put simply, machining is the process of cutting raw materials, such as metal, wood, or plastic, into the desired final Shape and size. Modern machine shops rely on computer-controlled precision tools for various operations, including cutting, drilling, boring, and milling. Machining is an essential part of the manufacturing industry, enabling the creation of intricate and precise parts.

Challenges in the Machining Industry

One significant challenge in the machining industry has been the time-consuming task of cost estimation. When machine shops receive an order for building parts, they need to carefully analyze the provided blueprints to determine the specific cuts and holes required for each individual part. This manual process often requires the expertise of machining professionals dedicated to cost estimation. As a result, it can become a bottleneck for the overall productivity of the machine shop.

AI Solution for Cost Estimation

Kabuku, recognizing the need for a more efficient cost estimation process, introduced a data management tool integrated with AI. The tool addresses the challenge of time-consuming manual cost estimation by automating the process and significantly increasing productivity.

Kabuku's data management tool simplifies the cost estimation process by handling the data management of the machining process using AI. The workflow is straightforward – users upload the blueprint in a web browser, specify the region of interest for estimation, and press the Cost Estimation button. The tool then detects and highlights every object that requires processing, providing detailed information about the processing steps, their quantity, individual costs, and the total estimated cost for the entire part. Additionally, users can add specific details for each part or request expert input for necessary adjustments.

Benefits of the Tool

The implementation of Kabuku's data management tool offers several benefits to machine shops. Firstly, the tool allows machine shops to respond quickly to cost estimation requests from their customers, improving the overall customer experience and creating more business opportunities. Moreover, the tool acts as a central repository of expert knowledge, allowing easy knowledge sharing with junior staff members and learners. This scalability enhances business operations and streamlines cost estimation processes.

Building the Tool with Google Cloud AI

To develop their innovative data management tool, Kabuku leveraged the power of Google Cloud AI. Specifically, they utilized AutoML Vision, a technology that automates several steps in the machine learning (ML) lifecycle, to detect the machining operations from the blueprints.

The development process involved collecting thousands of labeled blueprint images for drilling, boring, and cutting operations. Traditionally, building an object detection system for production use would require a considerable amount of time and resources. However, by utilizing AutoML Vision, Kabuku significantly reduced the development time. They simply uploaded their training images, pressed the Train button, and within half a day, they obtained an object detection model generated by AutoML Vision. The model came with a detailed report on model accuracy and various metrics.

To ensure seamless integration and accessibility, Kabuku utilized TensorFlow.js to run the object detection model within a web browser. This choice allowed machining professionals to efficiently detect tens of processing parts within a Second, enhancing productivity and expediting the cost estimation process.

Continuous Learning with User Feedback

One crucial requirement for Kabuku's data management tool is the ability to continually learn from machining experts. Practices and Patterns used in blueprints can vary for each customer project, and feedback from experts is essential for accurate cost estimation. To address this, the tool incorporates a user interface that allows experts to fix any misdetected parts and provide additional details.

Based on the feedback received, the tool triggers a continuous ML training process to improve the model. Kabuku achieved this by utilizing Cloud AI Platform Pipelines, which enables the building and orchestration of ML pipelines. The pipeline automates several tasks, including data pre-processing, validation, model training, and deployment. This seamless integration ensures that the model learns from expert feedback and captures the unique practices and patterns used in each case. As a result, the accuracy of the model improves over time, enhancing productivity and overall efficiency.

The Power of Active Learning

The continuous feedback loop between machining experts and the ML training pipeline is a prime example of active learning in ML. By continuously integrating expert feedback, Kabuku's data management tool becomes more accurate and tailored to specific customer projects. The active learning approach amplifies the productivity gains from using AI in the machining industry, making it a valuable tool for machine shops seeking to optimize their operations.

Conclusion

Kabuku's data management tool has demonstrated how AI can transform the cost estimation process in the machining industry. By leveraging Google Cloud AI, Kabuku successfully automated and optimized the previously manual and time-consuming task. The tool offers machine shops increased productivity, scalability, and improved customer experiences. Furthermore, the integration of technologies like AutoML Vision and TensorFlow.js ensures fast and accurate object detection, enabling efficient cost estimation. With the power of active learning and continuous integration through Cloud AI Platform Pipelines, Kabuku's tool continually evolves and adapts to the unique needs of its users, further enhancing productivity and efficiency in the machining industry.

Get Started with Google Cloud AI

Embark on your AI Journey today with the comprehensive tools offered by Google Cloud AI. Whether you're looking to automate processes, improve decision-making with machine learning, or enhance customer experiences with AI-powered solutions, Google Cloud AI has the tools and technologies to support your initiatives. Explore AutoML Vision, Cloud AI Platform Pipelines, and more to unlock the full potential of AI in your industry.