BoB: Your Personal Assistant for Predictive Maintenance
Think about all the machines you use each day. The hairdryer in the morning, the car to drive to work, and the heating system in the evening, to name a few. We depend on machines to make our lives easier, which means we highly rely on them to function. But how can we manage this reliability?
Reliability is managed by understanding machine data in a process known as predictive maintenance.
Predictive maintenance relies on the actual condition and trends/slow deviations from the normal state of the equipment to predict future conditions. It uses sensor monitoring to collect real-time data and send alerts according to the operating status of the machine. Statistical models detect abnormalities and failure patterns to anticipate when future maintenance will be required, before machine failure.
The BoB Assistant from Watteco is an example of an IoT device that can be deployed on various types of machinery across many industries such as manufacturing, energy and facility management. The device has an embedded temperature sensor and a 3 axis accelerometer, which detects orientation, shake, motion, positioning, shock, or vibration. Its embedded analysis algorithms understand and monitor equipment autonomously without any particular configuration. It operates across a LoRaWAN network for maximum range, deep indoor coverage and minimal energy consumption, all while sending only encrypted data. Operating at temperatures from -20°C to +55°C and with a battery life of over 2 years, it supports many applications.
A predictive maintenance case with the BoB Assistant and akenza
In this particular use case, we installed the BoB Assistant to an air circulation vent. While air circulation vents may be relatively simple machines, their proper functioning is critical, especially in crowded indoor areas like concert halls. A malfunctioning vent can significantly impact the comfort and safety of occupants, making proactive maintenance crucial to avoid disruptions and potential risks.
The BoB Assistant is connected to the akenza platform, where the data is collected and visualized. Custom rules and alerts can be created in the platform based on the sensor data.
The Learning Phase
The BoB Assistant is equipped with an accelerometer and a machine learning algorithm to recognize vibration patterns and faults. It does not require previous knowledge of the target machine (in this case an air vent) to operate. Once it is installed, it will start “listening” to the machine. It takes up to seven days to learn vibration patterns and understand what the normal vibration patterns are. After the learning phase, it uses its pattern knowledge to recognize a normal vibration versus a vibration anomaly.
Predictions and Anomaly
With the learned vibration patterns, the BoB Assistant calculates vibration trends and can predict if these will reach a critical threshold, sending an alert before that. In addition to that, we created a rule in the akenza platform that detects if the anomaly level crosses a certain threshold, sending an alert immediately for maintenance before any potential damage may occur. This is important as staying at a high level of an anomaly for an extended period could cause a larger, more costly complication.
By accurately identifying vibration patterns, the BoB Assistant can also track the operating hours of the machine, sensing when it is in use and when it is idle. This valuable data allows for proactive maintenance planning, as one can monitor the number of operating hours in the akenza platform and indicate when the machine is due for a service. This helps to prevent unplanned downtime and costly breakdowns, ensuring that the machine is serviced at the optimal time. It also helps to make the service interval dynamic based on the actual operating hours of the machine.
Integrate BoB into your solution
Whether you think about smaller household devices or larger industrial machines, the overlying principle of machine reliability still applies. However, the BoB Assistant is most suitable for use on expensive, critical equipment such as HVAC systems, CNC mills, compressors, transformers, generators, and pumps where the cost of the device can be easily amortized. By forecasting what type of machine failure will occur and when, unscheduled equipment downtime is eliminated. Other benefits include increased production capacity, reduced maintenance costs, and an increased equipment lifespan.
To explore the full potential of the BoB Assistant, akenza provides a payload decoder, built-in data visualization, and an easy connection to a dashboard builder such as Grafana, or a third-party connector such as Azure IoT Hub, Power BI, or Tableau.