Industrial IoT showstopper: Shaky Wi-Fi on the factory floor

 

Challenges and costs of Wi-Fi connectivity on the factory floor

Wi-Fi prefers the cozy conditions of the average office environment. It is comfortable with desks, coffee machines, computers, meeting rooms, and filing cabinets. But when faced with the harsh reality of the factory floor, it begins to flounder. The factory floor is full of metal obstacles that can disrupt fragile radio waves and prevent equipment from reliably connecting to the network.

While the Internet of Things (IoT) has the potential to bring huge benefits to manufacturing sites, the risk of outages in the data stream connecting machines and equipment to backend systems means automation, logistics, and monitoring systems may not be able to function properly.

One solution could be to strengthen wireless connectivity by installing additional routers, hardware, or mobile network connections. However, this comes with a cost, and additional costs are toxic for IoT solutions. Scalability depends heavily on keeping hardware costs in check.

Akkodis’ approach to IoT: less is more

The alternative, choosing a ‘less is more’ approach, is exactly what a team of Akkodis engineers implemented for a German manufacturing company.

IoT devices have diverse applications across various industries and sectors. Common use cases include smart automation, industrial monitoring and control, transportation and logistics, and energy management. Monitoring plays a crucial role in many of these applications, in which IoT devices are installed near the data source to be monitored.

Akkodis was tasked with developing a device that could record the sound of conveyor belts and sorters and send it to a computing system. The system would then crunch the data, detect anomalies, and thus predict malfunctions before they occurred.

A predictive maintenance system can help reduce the downtime of a production line – a manufacturing company’s most valuable asset – and save time and money. As any factory manager knows, nothing is more important than optimizing the Overall Equipment Efficiency – the measure of how well a manufacturing asset is utilized: 100% is the goal.

Akkodis engineers redefine maintenance predictions

The German manufacturing company needed the IoT listening device to help meet that goal. Akkodis engineers developed a device containing a sound chip, microphone, light barrier, and standard microcontroller components such as CPU, RAM, and various interfaces.

The light barrier identifies the start of the conveyor belt/sorter and thus maps the recorded sound data to the corresponding machine part. Measurement data and timestamps are transmitted to a cloud system using MQTT (Message Queuing Telemetry Transport). It is a highly lightweight publish/subscribe IoT messaging protocol which is ideal for connecting remote devices with a small code footprint and minimal network bandwidth. After transmission to the cloud, the data is used to calculate maintenance predictions for the respective conveyor belt/sorter parts.

Overcoming IoT data transmission constraints: Akkodis’ novel solution

During development, the engineers faced a number of challenges. The recorded raw sound data amounted to several gigabytes per day. All of the data was essential as the maintenance prediction algorithm relied heavily on receiving as much data as possible, uninterrupted and with minimal loss, to function properly. That meant stable wireless connectivity and a lot of bandwidth, leading to obstacles in the form of price and practicalities.

The cloud system had a message size limit of 4kB, as well as a limit on the number of messages it could receive in a given time frame. To ensure cost-effective scalability, the engineers had to come up with an innovative data transmission approach to deliver sound data reliably to the cloud, even in a demanding environment. The answer was to optimize data compression, manage transmission frequency, and track messages efficiently.

The connectivity solution developed by the Akkodis engineers reduces the amount of data to be transferred. By utilizing Fast Fourier Transformation (FFT) on the recorded sound data and further compressing it, message frequency decreases to one message per device per minute.

The FFT and further compressions are performed exclusively on one core of the IoT device. With only two CPU cores, the second core handles other essential tasks such as managing the Wi-Fi connection, constructing MQTT messages for the cloud system, time synchronization, task supervision, and more.

Implementing efficient traffic regulations for IoT connectivity

Shrinking the data was not all that was required. Efficient traffic regulations had to be put in place too. If a sensor disconnects and then attempts to send accumulated data immediately upon reconnection, the risk of triggering another disconnection heightens. That would lead to a cascade into more connectivity issues for nearby sensors.

Employing continuous transmission with a consistent data quantum sidesteps that risk, stabilizing Wi-Fi connections and drastically reducing the probability of connection losses. The system transmits two messages in succession instead of one, ensuring a uniform data volume and uninterrupted data transmission. After the data stored during the disconnect has been transferred, the transfer rate drops back to one message per second.

Defeating a showstopper with a reliable industrial IoT solution

By thoroughly analyzing the problem at hand, the engineers were able to offer the client an efficient listening device to monitor the condition of manufacturing equipment using secure connectivity through a specially designed algorithm, and a cloud system able to analyze the data and provide reliable maintenance predictions. In other words, an efficient and reliable industrial IoT solution.