Azure Data Hub & Tyre Quality Prediction Model for Bridgestone EU

About Bridgestone

Bridgestone Corporation is a global leader in providing sustainable mobility and advanced solutions. Bridgestone develops, manufactures, and markets a diverse portfolio of original equipment and replacement tyres, tyre-centric solutions, mobility solutions and other rubber-associated and diversified products that deliver social and customer value. These best-in-class offerings are sold to consumers and fleet customers around the world under the trusted Bridgestone and Firestone brand names.

The Challenge

With around a dozen tyre plants and 18,000 employees in Europe alone, Bridgestone EU has a strong emphasis on sustainability and quality with a continuous focus on advancing tyre design and development. In order to improve the current Plant IT and to benefit from opportunities in Digital Manufacturing, they have embarked on a Smart Factory Journey and are looking to embrace the capabilities of Industry 4.0. To achieve that, they need data & A.I. solutions that are data-driven, flexible, scalable, and cloud-native.

One of the challenges related to dealing with manufacturing data of such scale is the security. With a large number of plants and a variety of machines streaming data continuously we needed to figure out a way of storing the data on-premise before sending it in a secure way to the cloud, periodically.

The Approach

We started by immediately rolling out our Azure Data Hub based on reference architecture from Microsoft and provided training and handover to accelerate the transformation. Then, in collaboration with Bridgestone, we organized several workshops during which we looked at their current setup, its technical limitations, and performed a gap analysis to assess what else needed to be included. In order to meet the complexity demands of the organisation, the planned Data Hub needed to be scalable and include data warehousing and real-time data streaming capabilities.

Next to that, we looked at each of the production chain steps from raw materials to finished product in order to see where data can create insight and bring most value. Throughout the process, quality assurance is one of the key steps. Once tyres are finished, they go through several different machines that test for defects or irregularities as each tyre needs to be perfect. In order to optimize the process, we set up several data pipelines, at one of their plants in Spain, that feed information in near real-time about the status and quality of the products

The Results

Within 2 months we had already delivered an Azure cloud-native, scalable Data Hub that had streaming and data warehousing capabilities. To facilitate innovation and tackle complexity, we set up the Data Hub with a ‘sandbox environment’ which allows different teams to automatically request and receive their own isolated and secure environment within the Data Hub where they can experiment, log data, and build use cases which can then be validated and streamlined across the whole organisation faster and easier.

On top of that, using sensor data from the quality assurance machines, we developed and deployed a model that is able to predict and benchmark tyre quality across plants. Tyres were classified into ‘pass’ or ‘fail’ attending to a threshold of several quality measurements which included seasonal patterns. The solution was developed incrementally and deployed using MLflow within Databricks.

The Data Hub solution, which Anchormen specialists developed, helps Bridgestone EU accelerate their digital transformation and brings them one step closer to creating a Smart Factory. By enabling the use of their manufacturing data and adding advanced analytics and A.I. capabilities, we help them deliver not only operational excellence but also increase the scalability and speed of deployment in a connected and secure way.