Pumps used in process and manufacturing plants are often vulnerable to failure from cavitation. In the past, it has been difficult to detect the early signs of cavitation before substantial damage occurs, but a new industrial intranet/internet of things (IIoT) system using innovative technologies addresses this issue. With early detection, plant personnel can spot developing problems and take proper corrective action to avoid pump damage or total failure.
With this system, a high-sensitivity differential pressure transmitter collects unfiltered pressure data from the suction side of a pump. Advanced analysis distills this differential pressure data to create actionable information and identify plausible root causes for pump cavitation occurrences.
The systems automatically monitor unfiltered pressure data to detect minute fluctuations — a precursor to more serious cavitation issues. With data-driven information, maintenance engineers can develop the best preventive maintenance plans for reliability programs to improve pump operations.
IIoT systems effectively apply advanced analytics to enable early detection of pump cavitation, and applying these strategies improves process operations, reduces maintenance costs and avoids unscheduled downtime.
Let’s begin by looking at typical process plant pump operations, along with the impact of cavitation incidents.
Problems and causes for poor pump operations
Process industry plants are complex operations using expensive equipment to safely react and move feedstocks and products throughout the plant site and to storage areas. Pumps are key rotating equipment for handling process fluids in these plants. For critical service applications, operation must be uninterrupted, so the most important pumps in a plant are often duplicated because failure or reduced operation can cause a significant decrease in productivity or a total plant shutdown.
Cavitation is the implosion of gas bubbles near the pump impeller due to an undesired pressure profile within the pump. This bubble collapse exerts energy on the impeller face and is detected as shock, noise and/or vibration. These implosions can exert enough energy to erode metal from the impeller surface.
As illustrated in Figure 1, there are various stages of cavitation, and early detection is difficult. Often, cavitation is not discovered until operations or maintenance staff notice acoustic or vibration anomalies. By that time, substantial damage has occurred to the pump and often to connected equipment such as the motor driving the pump, local monitoring instruments and upstream and downstream piping.
Cavitation is a leading problem for pumps, so process plants are engineered to prevent this condition from occurring. However, processing operations, asset conditions and environments change over time, providing opportunities for cavitation to develop. In addition, improper actions by plant personnel can initiate or contribute to cavitation incidents.
To protect pumps, a better monitoring approach is needed — one focused on understanding pressure profile variations.
Advanced cavitation detection system overview
Early cavitation incidents are difficult to detect, even when using handheld vibration monitoring devices. There are no general indexes on cavitation strength definitions, but minute changes in the pressure profile are known to be early markers for cavitation-inducing conditions.
Detecting conditions within pipes and processing equipment provides greater insight for operations and maintenance staff. As shown in Figure 2, an advanced differential pressure transmitter able to detect minute variations in pressure is connected to the suction side of a pump, and the transmitter sends data to a controller via a high-speed digital communication link.
The transmitter senses minute pressure variations using a silicon resonant sensor, and this highly precise data acquisition enables cavitation detection. Performing cavitation detection in real-time requires 100 millisecond-cycle data acquisition by the controller from the transmitter, along with high-speed real-time calculation functionality in the controller.
The controller analyzes the internal unfiltered data generated by the differential pressure transmitter regarding pressure profiles in the pump and connected piping. It also applies cavitation detection logic to turn raw data into actionable information. With IIoT connectivity, pressure/cavitation information is sent to maintenance and reliability engineers for viewing via a PC or other device station for early detection of cavitation.
Because the controller contains a web server, information can be sent to any device capable of supporting a web browser. This information is transmitted over the internet or a company intranet. Alternatively, the information can be sent to a PC via Modbus.
This real-time information reveals very early cavitation conditions. More importantly, the information can be applied to troubleshoot root causes. The advanced analytics interprets fine-pressure changes stemming from cavitation events, and it alerts users before vibration or noise events are noticed.
The level of damage possible from cavitation is directly related to the level of changes in the pressure profile over time. With this history, reliability and maintenance staff can identify and categorize critical events. Using data-driven information, maintenance engineers can implement prevention methods to avoid future incidents.
Preventive maintenance strategies and action plan
With digitalization and IIoT, preventive maintenance is incorporating advanced factorial analysis and other predictive analytic techniques to increase the reliability and safety of plant equipment and operations. An advanced preventive maintenance strategy focuses on early cavitation detection for critical-service pumps. Key steps for a preventive maintenance strategy include:
Step 1. Cavitation detection
With predictive maintenance strategies, early warnings and detections provide valuable information. Online monitoring is an effective means to notice the earliest beginnings of cavitation events before damage occurs. More importantly, early detection enables prompt intervention to reduce interruptions to operations, and it minimizes maintenance repairs and costs. Planned maintenance reduces the time to get materials and make repairs, reduces the level of repairs required, and lessens the likelihood of the need for a total pump replacement.
Step 2. Advanced factorial analysis
Detecting that cavitation is occurring is not sufficient to protect key plant assets and operations, because the why and the how of developing cavitation conditions are equally vital. Advanced factorial analysis provides the required insights concerning process or operating changes that initiate gas-bubble formation in process fluids. Operations and maintenance staff can review cavitation histories, operations data and other events to identify and categorize root causes.
Pipe clogging, opening of valves, flow rate changes and more are possible root causes for gas-bubble formation in process fluids. Finding the proper cause for cavitation is valuable, but too often, such events are invisible and not audible, creating the need for advanced analysis.
Step 3. Investigation of root causes and action plan development
Advanced factorial analysis provides data-driven information to recognize problems, investigate root causes and devise an improvement plan for pump operations and maintenance. With real-time history information, operations and maintenance engineers can connect the cause-and-effect actions that initiate early cavitation events. Changes in operator or processing operations actions can inadvertently introduce anomalies that escalate into adverse conditions. Using data trends, advanced factorial analysis tools can be used to expose and highlight precise condition changes directly linked to cavitation. With a true root cause identified, better strategies can be applied to prevent future events.
Step 4. Implement action plan
Based on data-driven information and accumulated knowledge, plant staff can review procedures to minimize, and possibly eliminate, cavitation-inducing conditions. Best-fit preventive/predictive measures can be implemented to avoid cavitation-causing conditions.
Improving skill sets
Process industry firms are under increased pressure to optimize operations, increase profits and cut costs for both operations and maintenance. Better preventive/predictive measures are essential because they can control maintenance costs and minimize equipment and unit unavailability, especially unplanned downtime. Reliability engineers are responsible for implementing these improved measures. Based on years of experience, these engineers develop programs to lengthen the service life of critical-service equipment and schedule preventive maintenance to minimize unscheduled downtime.
But, as more experienced reliability and maintenance staff retire, greater efforts to capture and retain their expertise become critical. A solution is improved IIoT and predictive analysis tools to provide maintenance engineers with greater real-time knowledge to support critical-service equipment.
These tools shorten the time cycle from data extraction to sustained value creation. Advanced factorial analysis tools and pressure-monitoring systems provide the capability to identify and differentiate early cavitation events from other causes.
Advanced warnings can be the difference between making minor repairs with limited production interruptions and total pump failure, causing significant unit or plant outages (see Figure 3). Understanding and interpreting the entire chain of events can drive discovery of the true root cause.
Preventive maintenance is taking advantage of IIoT to protect and lengthen the service life of critical process assets, such as pumps used in process plants. New preventive/predictive maintenance methods with advanced differential pressure transmitters use IIoT, along with advanced factorial analysis tools, to monitor and report fine-pressure anomalies for critical-service pumps. Early detection of cavitation-inducing conditions enables maintenance and operations engineers to take proper corrective action and improve process operations, reduce maintenance costs and avoid unscheduled downtime.
Masaru Kimura is a system products marketing specialist with Yokogawa Electric Corporation. Since joining Yokogawa in 1999, he has worked with manufacturing execution systems for factory automation, plant information management systems, software product planning and the marketing of system products for the IA field. Kimura holds a Master of Computational Intelligence and System Science degree from the Tokyo Institute of Technology.