Subscribe to
our newsletter.
Need support?
Get in touch

Improve your product quality using BCAP virtual sensors

This moisture estimation and prediction example shows how process industry operators can improve their ability to monitor product quality indicators to reach a more stable production process. In this case, by applying data-analytics and machine learning to infrequent moisture measurement data as well as other readily available data, frequent measurements can be obtained without need for any physical installation.

Customer

Industry:Chemical Industry
Region:DACH

Challenge

Process industries are often challenged by waste of material and energy due to inadequate measurement and monitoring of key physical or chemical parameters  at critical steps and time frames of the production process. This can be because of high sensor cost, inaccessible asset areas, time-consuming laboratory sample testing, or other reasons.

Solution

The virtual sensor is one component of the BCAP product portfolio. Needing no physical installation, it can also be used in unreachable asset areas. A virtual sensor is a data-driven solution which provides real time insight into process, product, and asset parameters. This solution can be based on already-available physical sensor data, infrequent manual and laboratory measurement data, and control system set values. 

Benefit

Virtual Sensor provides insights which complement existing insights and help process industries to achieve Overall Equipment Effectiveness (OEE) improvements and reduce costs.

A use case about product quality improvement

Challenge: How to determine powder moisture content without physical sensors

For production of one particular powder material, moisture content (or dryness) governs product quality and process energy consumption. A belt filter was used within the solid-liquid separation unit. Besides low moisture content, other decisive factors for the operators of this belt filter were continuously stable dewatering processes and maximum reliability.

The final product’s moisture content is a key quality factor; frequent, timely data on this can greatly help operation. However, online measurement is not feasible because sensors cannot be installed where required. Moisture levels can only be measured manually by a three-hour laboratory sample analysis. The product could easily get out of specification within this time, reducing overall process stability, product quality and wasting energy and material. Any resulting production rerun will incur further cost, energy, and material losses.

Solution: A data driven, real time insight into product, process, and asset parameters

The answer is a virtual sensor – a data driven solution which provides real time insight into the process, product, and asset parameters.

To deploy the Virtual Sensor, first the target KPI (here, a product with low moisture content) and the parameters influencing it are defined, while considering the physical sensor data and relevant but infrequent manual measurements. No physical installation is required.

In the belt filter, data from related physical sensors is used with historical manufacturer-provided manual laboratory measurements. The virtual sensor provides predictive moisture levels at one-minute intervals. Enhanced by BCAP recommender models, it also provides recommendations on process adjustments to stay within process specifications.

Possible adjustments include belt speed, pressure, and volume of material fed onto the belts. This can result in significant process and quality stabilization improvements. Continued manual measurements may be desirable, as these can yield data for ongoing improvement of the virtual sensor model. However, these would only need conducting at 6-hourly or daily intervals, rather than 3-hourly intervals.

„The use case of monitoring moisture content during the belt filter dewatering process demonstrates how useful the Virtual Sensor can be in improving product quality. However, this technology is also valid for other application areas such as asset lifetime monitoring.“

Dr. Amir Gheisi, Senior Product Manager

Benefits: Data to facilitate product quality optimization and reduced energy use

The client wanted to obtain data that would reveal their product’s moisture content in real time, allowing them to make process adjustments to optimize product quality and reduce energy use. Bilfinger Digital Next offers virtual sensors as part of the BCAP platform. The virtual sensor models created can be applied to other similar asset sets with minimal effort.

Customer Benefits

  • Improves product quality
  • Reduces energy and material consumption
  • Improves process stabilization
  • Reduces man-hours on manual measurements

„As an example of asset lifetime monitoring, a virtual sensor can provide continuous monitoring of wear areas for refractory linings – an environment with extremely high temperatures. This is done using readily available data, such as material throughput, temperature profiles, product analysis and inspection/maintenance data. Accordingly, we can enable operators with recommendations on how to make process adjustments and eliminate unwanted and costly downtime.“

Dr. Amir Gheisi, Senior Product Manager