The company I work for, E.ON, is one of the largest energy providers in Europe – not only delivering electrical power to business and residential users, but helping companies find new ways to make their usage more efficient. By continuously gathering, analyzing and interpreting AI-derived, real-time data, our customers can take advantage of amounts of information previously too large to evaluate.
Planning for Energy Use
In manufacturing, for example, energy is one of the main cost elements – sometimes making up to 50% of production costs. However, by tweaking the process based on AI-derived data, we can reduce energy use and optimize productivity and yield, while providing wider operational benefits and lower costs. This is a new approach for operators – it’s more proactive and predictive. By planning based on this data, manufacturers can make better decisions.
For example, let’s say you are a manufacturer of a base material in a continuous process industry, such as paper, steel or glass. You wish to produce a product with demanding quality requirements. By predicting its characteristics into the future, you may decide to delay a scheduled change until you have improved forecast quality to avoid the risk of a rejected product. Alternatively, stable quality predictions may allow you to optimize energy process settings and make the product at a lower cost.
In both cases, you are able to proactively plan and make faster and better informed decisions, in a real-time environment, to maximize facilities and resources.
Managing Process Visualization
One of the ongoing challenges of factory data is its sheer volume and variety. The answer is to reduce this complex mesh of information into simple, usable metrics. A real-time stream of data from various locations can focus key indicators (even 20 or more) into a single dashboard that helps to track the process throughout and resolve problems as they arise.
For example, one screen may combine multiple aspects of the environment such as process parameters, energy consumption, product quality and environmental pollution control. Such visualisation tools allow operators and analysts to grasp data very quickly and spot key trends, providing insights that allow them to react to situations faster.
Can this be done manually? Of course. However, we’ve found that AI-derived data not only provides valuable traditional data, but a number of new insights as well. For example, historical data “snapshots” are regularly gathered to monitor process performance or energy use but such data collection is generally done retrospectively and is labour-intensive. This results in a less granular view which may in turn conceal a potential problem or missed opportunity.
The Value of the Real-Time Data Stream
A real-time data stream driven by an AI model collects information 24/7, giving analysts and operators a much more accurate view of the overall production line. This data may cover all aspects of the environment; for example, linking data from process control, raw material product tracking systems and product quality.
Real-time data is less prone to human error and can be standardized across the enterprise. This results in higher quality, more wide-ranging and more precise information. Analysts become more effective by reducing the time required to manipulate and interrogate the data, and operators have real-time tools to help them make faster and smarter decisions.
Creating New Capabilities
Finally, having a single point of data and analytics creates an environment that can be used to compare process efficiencies from line to line and between various facilities and sites. This enables more effective reporting and helps to develop normalized KPI reports across the business. Such a capability can even be used for setting up prototype runs – tuning the line for best performance and creating settings for new products. As a result, manufacturers are now able to create new best practices to improve business operations and remain competitive.