Artificial intelligence (AI) and the internet of things (IoT) have become such broad industry buzzwords today that the nebulous nature of these topics can make it very difficult to understand the relevance of them for your specific needs. The immediate thought when you read about these topics is that they are for the big companies with big budgets.
However, that’s not entirely the case — while we only hear and read about the big impact stories, there are indeed many “use cases” for more basic tasks that can significantly benefit from some level of AI capability within their system. One such area is in industrial applications, where we often hear about the benefits of predictive maintenance.
A McKinsey Global Institute report from last year, “Notes from the AI frontier: Applications and value of deep learning,” analyzed more than 400 use cases across 19 industries highlighting the broad use and significant economic potential of advanced AI techniques. Predictive maintenance was one of those identified, wherein the power of machine learning can be used to detect anomalies.
It said that the capacity of AI, and particularly deep learning, to analyze very large amounts of high-dimensional data can take existing preventive maintenance systems to a new level. By adding additional data layers such as audio and image data from other sensors — including relatively cheap ones such as microphones and cameras — neural networks can enhance and even replace more traditional methods. Hence, AI can be used to predict failures and allow planned interventions, helping reduce downtime and operating costs while improving production yield.
The use of AI relies on having relevant data. In a factory or on a production line, the combination and analysis of maintenance history, live data such as vibration, temperature, humidity, or any other critical parameter for the machine using IoT sensors, and video images can not only help with failure prevention but also prolong the life of a machine.
Enabling predictive maintenance
One solution for enabling predictive maintenance is the Intelligent Condition Monitoring Box, or ICOMOX, which is an open development platform for condition-based monitoring (CBM) of industrial equipment, assets, and structures. It monitors operating conditions from the surface of the equipment to identify potential faults and reduce risks associated with equipment operation and maintenance.
Provided as a complete solution in a box, it can be used to sense vibration, magnetic field, temperature, and sound. Hence, it can sense a motor, whether it is functioning properly, and arm production managers with vital information that can alert them to any potential issues with the motor that might cause the machine to fail and commission other systems to compensate in case the failure is inevitable.
It provides an open embedded sensor-to-cloud platform but also features embedded software and analytics for early detection of machine failures in condition-based monitoring applications. Hence, everything is included to do a level of analytics, enabling a better understanding of the system, machine, or plant.
A key feature is its multi-sensing capability. This is enabled with a number of off-the-shelf components providing high-quality data for IoT applications and enabling intelligent sensing from the edge of the network. The individual sensors are the ADXL 356, a low-noise, low-power three-axis MEMS accelerometer from Analog Devices; the BMM150 low-power, low-noise three-axis magnetic field sensor from Bosch; the IMD69D130 high-performance microphone with dual backplane MEMS technology from Infineon; and the ADT7410 temperature sensor from Analog Devices.
The ADXL 356 MEMS accelerometer provides accurate and reliable tilt measurements for environments high in shock and vibration without saturating the sensor, an important requirement for tilt measurement applications on heavy equipment, as well as airborne platforms such as unmanned aerial vehicles (UAVs).
The IM69D130 delivers low self-noise (high SNR), wide dynamic range, low distortion, and a high acoustic overload point. It uses Infineon’s dual backplate MEMS technology based on a miniaturized symmetrical microphone design similar to studio condenser microphones and results in high linearity of the output signal within a dynamic range of 105 dB. The microphone distortion does not exceed 1% even at sound pressure levels of 128 dBSPL. The flat-frequency response (28 Hz low-frequency roll-off) and tight manufacturing tolerance result in close phase matching of the microphones, which is important for multi-microphone (array) applications.
Each of the sensors can be configured for warning and alarm levels and timestamp events.
Another key feature of the ICOMOX solution is its ability to transmit vital data in tough industrial environments in which the sensor data might be critical. It operates using SmartMesh, which is a network technology consisting of a scalable self-forming multi-hop mesh of nodes, known as motes, which collect and relay data, and a network manager that monitors and manages network performance and security and exchanges data with a host application. The motes and managers enable a complete wireless sensor network solution.
In the ICOMOX, the Analog Devices LTC5800-IPM SmartMesh IP 2.4-GHz, 802.15.4e system-on-chip (SoC) communications device provides a complete radio transceiver, embedded processor, and networking software for forming a self-healing mesh network. The chip can be utilized as either a wireless mote, e-manager, or access point in a smart mesh network, and the solution claims to ensure greater than 99.999% network reliability in even the most challenging RF environments.
The box is CE- and FCC-certified with IP66 enclosure and is available in a very compact form factor for external and under-hood mounting. It also features various mounting adapters to accommodate a wide range of monitored equipment.
The ICOMOX illustrates an example of how it is possible to bring the benefits of AI, complete with embedded software and analytics, into an industrial environment and implement a technology solution that can withstand tough environments, measure key critical parameters all in one system, and provide data reliably, even in tough RF environments.
Amir Sherman, director of engineering solutions and embedded technology at Arrow Electronics, said that the cooperation between Analog Devices, Shiratech (the company that developed ICOMOX), and Microsoft is helping drive state-of-the-art solutions such as this. This is supported by the new engineering solution center software team in Gdańsk, Poland, available as a complementary service from Arrow in its role as a technology solutions provider.
Arrow Electronics will be showcasing this and other solutions at embedded world that enable developers to harness the power of AI in their applications. The company is evolving beyond its traditional role to offer more of a systems and solutions approach to provide technology, guidance, and support to help implement AI functionality across a wide range of sectors including industrial, health care, and transportation. Arrow provides solutions and services for many AI applications, including smart cities (street lighting, parking, public safety), industrial, robotics, autonomous machines, and agriculture.
More details and products being showcased may be found online at arrow.com/ew2019. In addition, the products mentioned in this article are available at arrow.com. ■