Artificial Intelligence in T&M Devices

Many new technologies for testing and measuring are waiting to enter the market in 2021, which will monitor and predict the chance of breakdown automatically. The new Artificial Intelligence (AI) based Test & Measurement devices will not only ease the job of maintenance but also drastically reduce the downtime… - By P.K Chatterjee

Although 2020 was a gloomy year as far as the progress of the overall human civilization is concerned, there are plenty of innovations that will ease the future course of progress. The field of Electrical Engineering has evolved further – though a fairly good amount of time has been lost due to repeatedly declared lockdown to control the spread of COVID – 19 at all parts of the world.

In 2020, globally, inclusion of Artificial Intelligence (AI) in Electrical T&M instruments has boosted up the capability, reliability and functionality of many T&M devices. Several research works too have been conducted in this field. Let us have a look at a few of the interesting developments in this area.

PCB testing

Today’s versatile use of Printed Circuit Boards (PCBs) and the increasing need to reduce their form factors are leading to more and more stringent requirements with regard to design and quality assurance. In good PCB design, it is very important to avoid any kind of electrical interference and ensure electromagnetic compatibility. The Fraunhofer Institute for Applied Information Technology FIT has developed a modular AI platform to optimize the design and testing of PCBs, thereby reducing the requisite outlay by as much as
20 per cent.

According to a communiqué from the Institute, “When designing a PCB for a new application, every care is taken to make best possible use of the available space and to position components as close together as possible without risking a failure. At present, this process relies largely on the experience of engineers, whose designs must then be tested in real trials. A further complication is that results are not stringently documented, meaning that error-prone designs undergo repeat testing, which leads to increased costs.”

Drawback of the current system: As per the communiqué, given their complex design, PCBs must be manufactured to extremely exacting specifications. For this reason, each completed PCB undergoes, at the very least, Automated Optical Inspection (AOI). This uses image analysis techniques in order to determine that the PCB has been produced as per design and therefore does not have any technical defects. At present, however, this method generates a high false negative rate, i.e., a lot of fully functional PCBs are incorrectly classified as defective.

These supposedly defective PCBs must then be inspected once again by hand, either visually or by means of measuring equipment. In other words, an unacceptably high false negative rate means that non-defective PCBs are being rejected and then require re-inspection, which in turn results in higher costs. On the other hand, if this rate is too low, the follow-up costs are high as a result of defective components entering the supply chain. It is difficult to achieve an ideal true positive or false negative rate based on human inspection, since human errors also enter into the equation.

The development: The work at Fraunhofer FIT shows what a future inspection process can look like. As in conventional Automated Optical Inspection (AOI), a camera records images of the PCB. This improves the quality of the decisions made by the algorithms. Here, it is vital that the modules are provided with high-quality training data. Initially, the modules for machine learning and deep learning are fed with a good selection of data.

“The modular design means we can harness several algorithms, which continually enhance their own performance. Data generated by ongoing automated inspection of components flows back to the algorithm. This then provides the basis for a process of self-learning by the artificial intelligence module. This permanent feedback enhances the database and optimizes the true negative rate. Early estimates from industry indicate this could reduce the use of production resources by around 20 per cent,” explained Timo Brune, Project manager at Fraunhofer FIT.

The modular platform comprises modules for machine learning, deep learning and artificial intelligence. In turn, each module features algorithms trained to undertake different tasks. For example, an algorithm from the first type of module (left) not only classifies input data but also extracts features from that data, which then provides the input for further modules… Source: Fraunhofer FIT

One of the appreciable advantages of these modules is – user can train the modules themselves, on the basis of their own process and production data. This means that companies retain control of their own data and are not required to send it to an external server, for example. This toolkit of algorithms can be combined at will for application to specific problems.

Once trained, the algorithms can also be used to design new PCBs. This ends the lengthy and costly procedure of trial and error – whereby components are arranged on the board until the optimal configuration is found. Instead, an algorithm helps to predict which configuration, from a host of options, offers the best functionality.

The PCB application is merely one instance of where a modular, self-enhancing algorithm platform can enhance design and quality assurance. In fact, this approach from Fraunhofer FIT can be applied to many other electrical systems. Here, too, processes can be optimized in order to achieve significant savings in time and production costs.

Data logging and recording

Yokogawa is a leading manufacturer of recorders, and also holds a solid track record in the consulting industry with the use of machine learning to predict and analyze both equipment anomalies and product quality for manufacturers. The company’s Artificial Intelligence (AI)-enabled versions of the GX series panel-mount type paperless recorders, GP series portable paperless recorders, and GA10 data logging software, which are components of the highly operable and expandable SMARTDAC+ data acquisition and control system. These can be used for monitoring and recording of voltage, current and many other process variables; evaluation of performance in equipment management, production, and product development; safety and reliability evaluation during product quality inspection processes.

According to the company, there is a rising interest in the manufacturing sector in the use of AI to prevent equipment malfunctions and maximize productivity. At the same time, AI technologies have a steep learning curve, and the introduction of existing AI-enabled products and the analysis of their data is typically beyond the expertise of novices to this field.

(Top) GA10 display (Bottom L2R) e-RT3 Plus and GX/GP Source:  Yokogawa

Recorders are used in manufacturing and R&D to acquire, display, and record data on voltage, current, temperature, flow rate, pressure, and other process variables. In order to improve productivity and product quality, Yokogawa is now building user-friendly AI-related functions into its recorders, data logging software and controllers. Just to cite an example, these functions now enable GX/GP series recorders to draw waveforms on screen that are predicted based on the real-time analysis of collected data – so that users can anticipate and correct problems early on, before they have a chance to escalate. Also, the company’s e-RT3 Plus edge computing platform has been enhanced with the addition of support for Python, a programming language that is widely used in AI R&D.

A new approach to support maintenance

Senzoro has developed a novel approach to predictive maintenance combining ultrasound with artificial intelligence. Ultrasound has been used by NASA for many years and Senzoro has combined this as a world first with artificial intelligence for the broad mass market.

As per the company, there is a lot of generic knowledge available, but the actual competitive advantage is achieved with specific data from your own systems and machines. To do this, combine the specific data with the publicly available knowledge, whereby the forecast of the Remaining Service Life (RUL) becomes more and more accurate. The earlier you start to ‘digitize’ the health of your assets, the greater is your advantage. It’s difficult or impossible to accelerate real data over time. There are test stands and ‘synthetic data’, but these can never be compared with real data from factories.

They say ‘PF (Potential Failure) curve’ describes how early you can recognize a potential failure. It is well known that ultrasound is the earliest technology to detect potential failures . The human senses are more in the ‘late stage’ (for example, if you already hear the “rattling noise”, the failure is not far away). Modern technologies give you a time buffer to carry out the exchange as efficiently and cost-effectively as possible. The technology has a high potential to detect the chance of failure of the bearings and other components in motors, generators and other electrical equipment.

Conclusion

The year (2020) of lost-productivity has created an avenue for better productivity in the year 2021 and henceforth. In the small span of this article, all the developments could not be included, thus there are many other technologies available now, which will help in improving productivity of electrically operated machinery. May the new technologies find the widest applications!


By P.K Chatterjee

Leave a Reply