Scalability in industrial analytics: How automation and integration drive success

Posted: August 29, 2024

In today's competitive industrial landscape, scalability is not just an advantage—it's a necessity. Companies need solutions that facilitate growth, seamlessly integrating new assets and data streams without sacrificing performance or accuracy. With the enhanced capabilities of the AVEVA digital twin for asset reliability, enriched with artificial intelligence and machine learning, companies can achieve new levels of operational efficiency and reliability to prevent unplanned downtime and performance leaks.

Now more than ever, companies are turning to AVEVA™ Predictive Analytics to deploy predictive maintenance programs to improve asset reliability and performance. When used in conjunction with AVEVA™ PI System™ and a hierarchical asset framework, companies are automating model-building which, in turn, is transforming predictive maintenance strategies to enable new levels of scalability.

Why scalability matters

To keep up with the ever-changing market and regulatory demands, industrial operations must continually scale. But why is scalability so important?

Scalability can improve:

• Efficiency: By automating processes, companies can reduce the time and effort required to build and deploy models, allowing them to focus on higher-value activities.
• Consistency: Automated model-building ensures all models are built and deployed using the same standards and parameters, reducing the risk of errors.
• Future-proofing: As companies grow and acquire new assets, scalable solutions ensure that predictive maintenance can keep pace without requiring significant additional resources.
• ROI: Scaling predictive maintenance efforts across the organization ensures quick time to value.

Automated model-building: Revolutionizing efficiency

To create a scalable operation that seamlessly integrates new assets, companies must avoid manual processes as much as possible—and that includes manually creating predictive models. Traditional methods often require extensive user input to cleanse data, configure models, and deploy those models across multiple assets, resulting in a laborious and error-prone process. Not only does this consume valuable time, but it also introduces the potential for inconsistencies and mistakes.

Now, with automated model-building capabilities in AVEVA Predictive Analytics, users can quickly and easily deploy predictive models—without the risk of traditional errors. Using a predictive model template, users can create models for all assets of the same type in a single action. This innovative feature automates the retrieval and cleansing of the training dataset, defines monitoring and fault diagnosis parameters, and optionally deploys the models. By leveraging the AVEVA PI System as a fundamental building block, companies can further streamline this process and enhance end-to-end asset reliability.

Southern Company was only capable of deploying 20 models per day—which wasn’t fast enough to move with the speed of the business. Using AVEVA’s automated model-building capabilities in conjunction with AVEVA PI System, the company can now deploy approximately 20 models in just two minutes. Not only did this approach dramatically increase productivity, it also improved efficiency and ROI while ensuring a consistent and accurate modeling process across all assets.

An integrated approach to model-building

How do AVEVA’s automated model-building capabilities work? Thanks to deep integrations between AVEVA Predictive Analytics and AVEVA PI System, users can create or leverage an existing asset framework in AVEVA PI System. Since the predictive analytics models hold a similar structure as the asset framework, the analytics leverage a subset of the structure with asset information and sensor information to automatically build, train, and deploy predictive models.

Together, these solutions underpin a digital twin for asset reliability that unifies, organizes, and contextualizes engineering, operations, and asset reliability data in a single industrial intelligence platform. With this information at their fingertips, users can make better, faster, more informed decisions.

AVEVA Predictive Analytics and AVEVA PI System are designed to meet the evolving needs of modern industries. By automating model-building and leveraging deep integration with existing systems, together both solutions ensure efficient, consistent, and scalable predictive maintenance.