Effective maintenance of complex industrial processes is critical to maintaining productivity and reducing costs. Condition monitoring (CBM) is an alternative to regular maintenance schedules that uses data from sensors in the Industrial Internet of Things to monitor trends in performance before failures occur. This article will explore this topic and recommend suitable power modules for sensor nodes.
Developing countries are moving towards a high level of industrial automation, enabling high output with high flexibility and low cost. Likewise, the goal of data centers, distribution warehouses, and infrastructure is to operate “with the lights off” with little or no labor (unmanned factories) to reduce labor costs. "Industry 4.0" or the Industrial Internet of Things (IIoT) is an integral part of this, pushing intelligence to the "edge" of the process, closer to where it needs to be monitored and controlled to speed up reaction times, so there is a need for monitoring and sensors versus central control A communication network is required between them. This could be through the "cloud", by collecting and analyzing data, and feeding it into control algorithms to optimize processes in so-called smart factories.
The benefits of a smart factory are many, not only enabling low-cost products and services, but also significantly reducing energy consumption. But what if something goes wrong? We're not talking about "the rise of the machines" here, we're talking about a simple mechanical failure, or a false processor failure due to interference from cosmic rays, or an "accidental failure" for a myriad of reasons. With much of the world's electrical infrastructure over 25 years old, rising failure rates are a concern. Based on this, system designers provide the redundancy required for critical components to cope with single or double failures, and select large margin components for reliable operation depending on the application. However, "wear and tear" is a reality for mechanical parts and even electronics, capacitors dry out and age, surge limiters get damaged, and semiconductors develop lattice defects over time.
must be maintained
We can wait for a failure to fix the problem at the source, adopting a "if it ain't broke, don't fix it" attitude, but failures rarely happen only at a convenient time. This "corrective" maintenance is an attractive approach for non-critical functions, such as a lighting cluster with a blown LED, or as in our previous example of multiple redundant systems, where the system is not affected by a single failure. Of course, this premise is that you must know that a failure has occurred, so monitoring is the key. Sufficient parts inventory and manpower must be ensured at all times, and it must also be on call "just in case".
Most processes rely on regular "preventive" maintenance and inspections to better maintain throughput. Simple mechanical systems require regular replacement of filters and oil, inspection of bearings or adjustment of clearances, etc. On the electronics side, functioning but old fuses, arresters and electrolytic capacitors may be replaced. Deciding when to act is not easy, doing it too late can lead to destructive failures, and replacing too early can throw good parts away creating unnecessary work and wasted costs. So timing is a challenge, arranging work based on usage time, usage, or just intuition and experience. High-availability systems typically perform a Failure Mode, Effects, and Criticality Analysis (FMECA) to predict failure frequency and impact on a scientific basis. This minimizes parts inventory and schedules maintenance work at the right time.
Condition-based maintenance is ideal
When faced with large and complex processes, it is difficult to accurately schedule preventive maintenance, so the alternative is Condition Maintenance (CBM). It is ideal to replace or adjust according to the remaining service life after measurement. This involves understanding the current condition of the part, which is still performing well but has started to enter the "wear and tear" phase. For example, an oil filter might have enough lube flow today but lose 10% of its filtering effect, but if you know that a 50% loss is not OK and it will be 10 weeks before it reaches that state, then you can schedule a Replace after 8 weeks. Likewise, real-time analysis of changes in the vibration characteristics of a motor or machine can predict when a bearing will fail.
Performance changes and trends can be detected by parameters such as liquid level changes, vibration signatures, infrared thermal imaging for non-contact temperature measurement, oil turbidity, current and voltage characteristics of power equipment, ultrasonic leak detection, arc and corona detection. Ozone sensor, etc. If CBM is implemented on an existing system, the cost of adding this level of intensive monitoring may be prohibitive in the short term, but the savings in the long term are very clear. Fortunately, implementing IIoT big data scheduling and optimization also yields data for CBM analysis. Any sensor-specific functionality required for CBM is relatively easy to connect to processor nodes at the edge of existing processes. CBM data is inherently slow to change, and the requirements that add to IIoT computing and communications are negligible, whether wired or wireless. The differences in maintenance regimes are shown in Figure 1.
Figure 1: The effect of maintenance regimes on process performance and reliability
CBM needs reliable power supply for sensors
If the available power source varies widely, such as solar power charging the battery, a regulated DC output is required. Simple linear regulators that are too inefficient can drain the battery very quickly. In this case, RECOM's R-78Exx-1.0 is the ideal switching regulator module with up to 97% efficiency in outdoor solar applications and also suitable for monitoring mobile equipment such as railway axle bearings.
The solution for node power using local AC power is usually a small AC/DC converter, sometimes up to 277VAC. These can be used in conjunction with the upcoming IO-Link industrial sensor system, a digital bidirectional serial interface using standard M12 connectors. The system requires 24V and has a maximum load of 410mA per node, so RECOM's onboard AC/DC converters of the RAC10, RAC20 and RACM40 series meet the 40W requirement for four IO-Link ports. The lower power RAC03 series can also be used to wirelessly connect a process controller to remotely control the process and signal when temperature is out of specification.
The example shown in Figure 2 has a volume of only 40x25x25mm when finished. Process feedback is not limited to temperature; noise, fluid flow or gas sensors can easily be combined.
Figure 2: Small AC/DC converters power sensors, microcontrollers, and bidirectional wireless circuits
Gas leaks, noise and IR temperature sensors are typically ceiling mounted, and the lighting circuit is supplied with a suitable AC supply of 115V, 230V or 277VAC, which is the phase voltage in a 480V three-phase system. Data transfer using long-range LPWAN radios (LoRa, sigfox, KNX-RF, etc.) or cellular networks (5G, NB-IoT, GSM, etc.). Since power requirements are generally low, RECOM's universal input, 5W onboard RAC05-K/277 is a good choice.
To achieve the goal of autonomy, some sensor nodes can be powered by harvested energy, which may be solar or acoustic, radio frequency, vibration or temperature gradients. Voltage sources can be very low so efficient power conversion is required to boost voltages suitable for sensors and processors. The RECOM REH-3-31.8 module is designed for this job, operating from an input voltage as low as 50mV while achieving dual 3.3V/1.8V outputs. This part includes maximum power point (MPP) tracking for photovoltaic cells and can be coupled with batteries or supercapacitors for energy storage.
Condition monitoring is an ideal way to maintain process availability at the lowest cost, while easily integrating with the IIoT. Reliable and economical power conversion is an important element of CBM and RECOM offers a wide range of product portfolios to meet all application requirements.