{"id":779,"date":"2021-08-09T09:41:00","date_gmt":"2021-08-09T07:41:00","guid":{"rendered":"https:\/\/blog.techbase.eu\/en\/?p=779"},"modified":"2021-08-10T09:44:05","modified_gmt":"2021-08-10T07:44:05","slug":"the-performance-of-ai-predictive-maintenance-in-industrial-iot","status":"publish","type":"post","link":"https:\/\/blog.techbase.eu\/en\/the-performance-of-ai-predictive-maintenance-in-industrial-iot\/","title":{"rendered":"The performance of AI & Predictive Maintenance in Industrial IoT"},"content":{"rendered":"\n

IoT technology solves challenges of predictive maintenance by enabling real-time remote status monitoring. Instead of maintaining hardware monthly or every 11 quarters, companies can use IoT health monitoring for direct maintenance. Simple IoT devices can now monitor critical performance metrics in real time and send alerts when problems start.<\/p>\n\n\n\n

Industrial IoT on the site<\/h2>\n\n\n\n

Traditionally, those types of failures were prevented by routine scheduled maintenance by specialized technicians. But as discussed earlier, scheduled maintenance is imperfect. Fear of equipment failure often drives businesses to set aggressive preventative maintenance schedules, which of course raises costs.<\/p><\/blockquote>\n\n\n\n

The cost of equipment failure tends to cascade. The failure of a single piece of machinery can bring an entire assembly line to a grinding halt. And for every hour that production has stopped, hundreds of thousands of dollars might be slipping out the door. That hiccup in production can also result in unfulfilled contracts, angry customers, and upheaval in your supply chain.<\/p><\/blockquote>\n\n\n\n

In fact, according to a recent paper from IBM<\/a>, up to 70% of a company\u2019s investment in preventative maintenance does not affect uptime metrics. This is due largely to the fact that only 11% of machine failures follow an age-degradation pattern. A whopping 89% occur at random.<\/p>Source: https:\/\/blog.particle.io\/the-growing-power-of-iot-in-preventative-maintenance\/<\/a><\/cite><\/blockquote>\n\n\n\n

New features of Edge TPU brought to ModBerry series<\/strong><\/h3>\n\n\n\n

In October 2020, with the release of the latest Compute Module 4 from Rasbperry Pi Foundation, TECHBASE announced an upgraded device from ModBerry 500 series, called ModBerry 500 CM4<\/a><\/strong>. Thanks to the high-performance PCI-Express bus introduced in Compute Module 4 and Raspberry Pi community, the device itself presents support for a wide range of new applications, such as use of Google\u2019s Artificial Intelligence<\/strong> modules at ease.<\/p>\n\n\n\n

Therefore, TECHBASE designed a new device, called ModBerry AI GATEWAY 9500-CM4<\/strong>, utilizing the vertical format of ModBerry 9500, latest Compute Module 4 and Google\u2019s Coral TPU. Installation-ready AI GATEWAY<\/strong> allows direct application in industrial fields.<\/p>\n\n\n\n

\"TECHBASE\u2019s<\/a><\/figure><\/div>\n\n\n\n

AI GATEWAY with Coral TPU enhancement <\/strong><\/h3>\n\n\n\n

Neuron network capabilities enhance CM4-based devices, not only collecting and sending data, but also allows local data change predictions and allows direct management on-site. This feature gives the possibility for various applications, such as data analysing and establishing trends predictions, smart alarms and smart monitoring, local notification control, etc.<\/p>\n\n\n\n

Used Edge TPU<\/strong> coprocessor via PCI-Express bus is capable of performing 4 trillion operations per second (TOPS), using 0.5 watts for each TOPS (2 TOPS per watt). Google Coral easily integrates with Raspberry Pi Compute Module in Linux and optionally in Windows with full support of TensorFlow Lite <\/strong>framework and AutoML Vision Edge<\/strong> solution.<\/p>\n\n\n\n

\"TECHBASE\u2019s<\/a>
TECHBASE\u2019s AI GATEWAY series, world-first industrial gateway utilizing Raspberry Pi Compute Module 4 and Google Coral TPU\n<\/figcaption><\/figure><\/div>\n\n\n\n

ModBerry AI GATEWAY 9500-CM4<\/strong> series also offers a standard PCI module support for various wireless communication protocols, such as:<\/p>\n\n\n\n