Predictive maintenance (PdM) is outgrowing industrial applications to transform consumer products in part because our products make this new field more accessible and cost-effective, thus enabling makers and startups to offer solutions that will change the way we relate to technology. Predictive maintenance (PdM), also referred to as preventive maintenance or additive manufacturing, defines the implementation of sensors, connectivity, and processing to improve maintenance costs, uptime, and safety. While automation and manufacturing environments have been using such solutions for a while, costs and the ease of using such solutions are now ready to create a new wave of business models in consumer segments.
Since this is an area where technology is advancing very fast, let’s consider some typical non-industrial use cases. Consider an HVAC system in a typical house. Why wouldn’t consumers use PdM to ensure the system doesn’t fail at the peak of hot weather or the bottom of cold weather? Why not use a PdM service to ensure air-quality maintenance happens at the right time and according to sensor data rather than some arbitrary schedule? Manufacturers would also want to know that their appliances/machines are installed and used properly, for the purpose of warranty or service warning.
What Changed in Predictive Maintenance Solutions?
Recent advances in sensor and connectivity technologies are allowing predictive maintenance systems to get significant feature improvements along with reductions in size and power consumption. PdM solutions can now be compact enough to avoid interfering with the machines they’re monitoring, they can run for multiple years on a battery, and they can connect to the Cloud for remote decision/action.
ST offers many PdM solutions that range from very flexible solutions developed using the STM32 Open Development Environment (ODE) to very specific solutions dedicated for PdM only. To implement PdM using STM32 ODE one can start with an STM32 Nucleo board, such as NUCLEO-F401RE, and an X-Nucleo board based on the types of sensors needed. One example that could be a good starting point for sensing motion or environment: X-NUCLEO-IKS01A2 incorporates an LSM6DSL 3-axis accelerometer + 3-axis Gyro, an LSM303AGR 3-axis magnetometer + 3-axis accelerometer, an HTS221 humidity and temperature, and an LPS22HB pressure sensor. This X-Nucleo board also includes a 1 Mbit EEPROM to perform edge processing and optimize the communication payload with the cloud.
Once the user selects the correct Nucleo and X-Nucleo boards, the next step is to install the STM32 ODE software on a PC and download the correct Function Packs. These Function Packs allow users to capture sensor data to monitor the health of the system and to predict further actions or possible failures. The software platform and Function Packs included along with our evaluation platform allow users to program windowing such as flat top, hanning or hamming, and program FFT averaging and overlapping. In addition to FFT, users can also perform time domain analysis using the included software. Finally, users can convert microphone data from PDM to PCM, extract sound pressure or perform FFT for acoustic emission analysis.
What Is the Next Challenge?
Like ‘the last mile problem’ in telecommunications and supply chain, PdM has its own challenges because one size never fits all solutions. Every PdM implementation has to first go through a learning curve to understand the problem it is trying to uncover. As a result, every implementation is specific to the machine and environment even though it may use the same hardware. This is why software application developers play a very important role in creating end-to-end solutions that combine hardware, software and cloud connectivity.
STMicroelectronics has pioneered many solutions and partnered with numerous third parties to deliver end-to-end PdM solutions for motors and other industrial systems. The Latest advances in PdM utilize motion sensors, distance/ranging sensors, and environmental sensors that, coupled with strong connectivity, processing solutions, and a complete software/firmware ecosystem, enable users to get started on defining state-of-the-art solutions.