Yokogawa Electric announces the release of the Equipment/Quality Predictive Detection Tool. This addition to the OpreX™ Data Acquisition family is an AI-based tool for building equipment and quality anomaly predictive detection systems for GX series, GP series, and GM series SMARTDAC+™ paperless recorders and data loggers. With this software, even users who are not AI specialists will be able to build their own equipment and quality anomaly predictive detection systems for manufacturing sites. It will help them improve production efficiency by identifying equipment defects and deteriorating quality in their plants and other facilities at an early stage.
Background
Recorders and data loggers are used at production and development sites in all kinds of industries to collect, display, and record data on voltage, current, temperature, flow rate, pressure, and other variables. As a leading company in this field, Yokogawa has provided many customers with data consulting services and technologies such as machine learning that can help them predict problems with plant equipment and product quality and analyze and identify causes.
In recent years there has been a rising demand for AI-based solutions to improve production efficiency in plants. However, the hurdles for the application of AI are high as this requires significant expertise in specialized fields like data science. To address this need, Yokogawa has developed the Equipment/Quality Predictive Detection Tool, an easy-to-use AI-based software application for the recorders and data loggers that are commonly used in industry. To use this tool, no specialized AI expertise or consulting is required.
Creating model on existing data without specialized knowledge
A predictive detection model can be created by importing past data into the software and simply flagging it as normal or abnormal, without needing to rely on an AI expert or consultant with knowledge about machine learning, algorithms, etc. Data recorded with Yokogawa and other companies' products can be used. Simulations can be run in advance to see how the AI assesses the data.
Building a predictive system easily
By loading the predictive detection model created by this software into the SMARTDAC+ on site, an equipment and quality anomaly predictive detection system can be constructed. The degree of equipment deterioration can be confirmed before failure by checking the health scores. These health scores enable operators to be informed by alarm or e-mail when equipment needs maintenance, minimizing the likelihood of an unexpected breakdown that can impact production activities.
Cloud or offline version
The Equipment/Quality Predictive Detection Tool will be available as both a cloud and an offline version. The equipment and quality anomaly predictive detection system can be built using either version. The cloud version is more easily available and does not require any installations on a PC.
Sample Applications
1. Managing temperatures and pressures in tire production (vulcanization)
Pressure leaks due to packing deterioration and so on are an issue with the vulcanizers that are used to apply heat and pressure to tire rubber. By using the Equipment/Quality Predictive Detection Tool to monitor changes in a health score that is generated based on a vulcanizer’s pressure readings, signs of packing deterioration can be detected at an early stage.
2. Managing the thermal treatment of aerospace and automotive parts
In the thermal treatment of aircraft and automotive parts, issues like bad burners and inadequate sealing can lead to furnace downtime and poor product quality. The Equipment/Quality Predictive Detection Tool will enable signs of temperature problems to be detected before an alarm is triggered, so users can avoid product loss and predict when to perform maintenance.
3. Managing sterilization of food and drug products
With the vacuum sealing of food and drug products after they have been sterilized, unexpected equipment breakdowns can bring the production line to a halt and result in product loss. The Equipment/Quality Predictive Detection Tool will enable the detection of loose valves and packing deterioration before an alarm is triggered, so users can prevent equipment breakdowns and reduce product loss.
Major target markets are production sites for a broad range of industries, such as steel, power, chemicals, pulp & paper, food, pharmaceuticals, water and wastewater treatment, and electronics. But also consumer electronics, automobiles, semiconductors, and new energy development departments; university and public sector research institutions will benefit from use.
