Manufacturing Quality & Performance: The AI Boost

Predictive quality monitoring

Quality monitoring can be time-consuming, especially in laboratory settings. We develop machine learning models to predict quality based on process parameters, requiring the integration of quality and process data (MES-QMS) for effective operation. This forecast information is then incorporated into the operational view of production.

Anomaly detection and root-cause analysis

By identifying deviations in quality parameters and comparing historical situations with similar quality deviations, we can perform root cause analysis to explain these deviations based on process parameters and other factors.

Edge computing

For situations where production data cannot be exported to the cloud or network connections are unreliable, edge computing provides a solution by performing calculations locally. This approach also allows for non-sensitive data storage in the cloud.

Multivariate methods for deviation identification

When traditional SPC (statistical process control) fails to detect deviations due to multiple variables, multivariate methods, including machine learning models, can be utilized to identify more complex deviations.

Root cause analysis of complaints for production optimization

To optimize production, we identify process variables that explain customer complaints. By recognizing diverse quality requirements in a varied customer base, we can optimize processes and production to be more customer-specific, leading to increased profitability.

Customized visualizations for enhanced user experience

When required, we create data visualizations that cater to the specific users of the software. For instance, an operational desk built for production monitoring displays real-time process and quality parameters, as well as alerts for any deviations.

Design thinking for better outcomes

By utilizing service design methods, we identify crucial development targets and problems when users are real people. This approach ensures that development resources are directed towards achieving the most significant results in line with business goals. Additionally, we handle user interface and visual design when needed.

Broad experience across the industrial domain

Veracell has a proven track record in a wide range of quality control, forecast development, and visualization projects. We offer a range of services to the industry, including data processing, machine learning model development, and analytics.

For example, we collaborated on a product development project to create a device that automatically prepares sequencing samples for hospital use. Combining the efforts of device developers and data scientists, we developed a real-time solution to measure the device's performance and the effect of different parameters on quality. This included automatic data collection, AWS cloud service processing, and the development of a user interface for monitoring sample quality and parameters.

In another project, we implemented prediction models for energy consumption and identified repair needs and optimization opportunities for mobile phone network base station components. Using historical data, renewable energy source production forecasting, and time series methods, we were able to visualize component states and create alerts for operators regarding potential anomalies.

At Veracell, we are committed to transforming industries with our expertise in data analysis and integration, ensuring optimal outcomes for our customers.