The EWA Sensor condition monitoring solution - A short technical review

1.   Motivation

As a pump manufacturing company, we are approached on a weekly basis by companies offering condition monitoring solutions for our pump systems. The business potential of condition monitoring and customer service agreements are huge, and this has triggered an explosion in the numbers of companies offering Condition Monitoring Solutions (CMS) that can detect every possible machine fault in a rotating machinery. As a customer, it can be difficult to navigate between all the CMS offerings, but a simple and fundamental question is the following: “Can your system detect the rotation direction of my machine?”. For a rotating machine, the RPM and Rotation Direction are fundamental parameters, and if the CMS system is not able to measure these two parameters, the foundation for further detection of unbalance, bearing failures, gear issues etc. becomes a little frivolous. It is a good measure of the insight and understanding of rotational machinery that the CMS provides, and will flag, a “wannabe” CMS system. A good CMS takes many years to build and demands a deep technical knowledge about applications, signal analysis and sensor design. In this relation another interesting measure is the ISO 18436 certification level of developers. No certification is for me a clear flag, that they are in it for the quick win - I would never consider to adapt to such a system.

One particularly system has recently caught my attention - the solution from EWA Sensors. EWA stands for Early Warning Analytics and it is a small start-up company in Denmark. The system measures a range of parameters from RPM, Rotation Direction, unbalance, cavitation etc. to gear issues and the founder is ISO 18436 Cat-IV certified from both Möbius Institute and Technical Associated from Charlotte PO - so the company passed the acceptance test 😊

It is my purpose with this review to provide insight into a new exciting CMS platform, and inspire others to make similar review of other CMS - to share insight.

2.   The EWA Platform

The founder of EWA sensor has chosen to follow another approach, than we normally see in this age of IoT devices. The sensor is a pure edge device - no cloud support and no human-in-the loop. All analytics are performed by the sensor using a build-in Cortex-M7 microcontroller. It is a wired sensor for Modbus communication and 24VDC power. All parameters are accessible for the customer through Modbus/SCADA. The sensor measures vibration, magnetic field and temperature, and apply a combination of Vibration Signature Analysis and Motor Current Signature Analysis, which make the parameters set very robust. It is a true plug-and-play platform with no pre-configuration of parameters, like motor pole pairs - the sensor will find the number by analysis. The sensor is designed as a pressure capsule and can be used in both dry installed and submerged (10m) applications. For Water Utility applications, there is general no benefit in using wireless IoT - if we are going to use 50m of cable to the gateway, it is not a problem to use additionally 4 meters to the sensor. The sensor is equipped with 4 LEDS - for Power, Rotation Direction, Modbus communication and Alarm/Machine Health:

The sensor calculates a large number of condition monitoring parameters (accessible through Modbus), but I will in this review only focus on two parameters, that is new to the marked.

2.1 Rotation Direction

Many CMS seek to provide very high-level conclusions in relation to alignment, cavitation, bearing condition etc., but miss fundamental low-level operation parameters. When we are talking condition monitoring for a rotating machinery, it is important to know - is my machine rotating or is it standing still - and if it is rotating - what is the rotation direction - CCW (negative RPM) or CW (positive RPM). Many CMSs cannot tell if the machine really is rotating - vibration on the line frequency is not a clear sign of an operating machine - it could be the machine next to you. The EWA sensor measures the rotation direction of the motor every second and provide this information both through an LED on the sensor with CW (green), CCW (red) and no rotation (no light) and as a Modbus parameter.  It sounds very simple, but I don’t see this feature very often. It is an important parameter, as rotation direction can be shifted after a service, if the power cables are incorrectly reconnected by mistake. Numerous real-life cases exist with sewage pumps, that have been operating for years with wrong rotation direction, without anybody has noticed anything, other than lower performance. If a CMS is not able to measure Rotation Direction, it is not a condition monitoring system for rotation machinery, but a condition monitoring system for vibrating machinery. As a side-effect of measuring the rotation direction, the EWA sensor also detects whether the installation is a horizontal or vertical machine.

2.2 High resolution RPM

The rotation speed of the pump impeller is often regarded as a fixed number, but nothing could be more misleading. RPM is a time function, and the RPM fluctuation of a pump impeller contains a lot of information of the pump loading and the liquid homogeneity - this will be evidenced when the RPM is estimated with high time resolution, like every second.  

The EWA sensor estimates both the RPM of the rotor rotation and the magnetic field rotations, thereunder also the motor pole pair number.  RPM estimation from a pure vibration spectrum can be very difficult in environments with simultaneously operating machines (like a booster station), but this combination of Vibrations Signature Analysis and Motor Current Signature Analysis makes the RPM estimation very robust. The following illustration shows the magnetic (blue) and vibration (orange) RPM measurements from a sewage pump performing deragging to reduce clogging:

It is clear, that the deragging is successful, as the fluctuation of the impeller RPM is reduced (the orange curve) after deragging.

As the sensor also measures the rotation direction of the pump, the RPM is a signed variable. This is illustrated in the following plot:

From the RPM measure it is clear, that every operation cycle is ended by a short reverse rotation - following the requirements from the pump manufacturer.

Another interesting feature of the EWA sensor, that I have not seen from other CMS providers, is the submerged robustness. The sensor has been pressure tested to 30m and approved to 10m. We are currently testing this with our mixer program, and the results look very promising:

The sensor provides us with deep insight into an application, that is normally not accessible. Information about slip, RPM, 3D RMS, GMF, etc. and corresponding baselines are now calculated with one second interval, and this provides us with an insight that we did not have before. This is valuable for both the pump manufacture (pump performance, design considerations) and for the customer (machine health and process performance).

3. Conclusion

The EWA Sensor is offered as a one-time-buy with no subscription fee, which makes the economic part simple. It is targeting customers with their own SCADA or cloud system for display of parameters, baseline and alarm handling. The sensor is glued on the outside of the pump unit, or could be built into the pump, like into the terminal box. The approach is generic and will work on machines with an electric motor - no specific brand is required.  All in all, an innovative condition monitoring solution, that is different from what is currently being offered by the CMS marked.