Detecting anomalies early – Lowering energy costs

Monitoring solutions are unquestionably an important tool paving the way toward a digitalized system, improved maintenance and energy-efficient processes. To this end, basic data such as vibrations, temperatures, operating hours and information on the operating point are key to success. This data first must be available, however, as well as correctly analyzed and interpreted. The overriding objectives are clear – identify incorrectly operated components and predict signs of wear and damage to avoid cost-intensive downtime and reduce energy costs. In this context, the question remains as to how digital solutions can facilitate analysis and interpretation, and how this data can be optimally used.

Transforming data into information

The magic word in this context is anomaly detection. Here, the objective is to be able to make predictions about anomalies based on current operating parameters, allowing precautions to be taken before an unscheduled downtime occurs. This method is also known as predictive maintenance. Predictive maintenance links and evaluates existing machine data using statistical models and intelligent algorithms.

How do operators implement intelligent, yet practical predictive maintenance, however? Reliable, centrally collected digital status data on the machinery in operation is first required. Once this data has been obtained, the status data should be evaluated and visualized – ideally in automated fashion. After the evaluation has been made, users can derive key information. For example, trends can already be recognized based on vibration patterns which can, in turn, point to a potential problem even before actual damage occurs. These trends indicate what type of pump problem exists. Gradually elevated vibrations over the course of a few weeks or months, for example, indicate that there is a technical problem with a pump. This is of course not noticeable to the observer because the pump is running well. Suboptimal pump operation cannot be hidden from an intelligent algorithm, however. This makes it clear that digitalization brings about sizeable benefits and, thus, also optimizes or can even transform maintenance management. Operators can now benefit from transparency, operating reliability and availability.

Easy entry into predictive maintenance

Studies conducted by the VDMA (German Mechanical and Plant Engineering Industry Association) have shown that 70 percent of unscheduled system downtimes can be detected in advance and avoided by leveraging monitoring solutions. With a digital all-round service, operators can optimize their maintenance management as well as receive targeted support from experts.

Getting started with intelligent predictive monitoring of pumps is easy for users. The sensor unit can be bonded to the pump or other rotating equipment by the customer or the solution provider. The sensor unit records vibrations and temperatures and sends them to a cloud. The data collected is then analyzed using patented algorithms and subsequently visualized on the web portal or in the app.

A clearly arranged dashboard provides users with an overview of their pumps at any time and from anywhere. They also receive a great deal of ancillary information on temperatures, vibrations, operating hours and load conditions/profiles. Documentation on the pumps can also be viewed, and a history of all maintenance activities can be configured. This, then, establishes a comprehensive history file for the pump. Whoever is in need of a reminder for bearing replacement or lubrication can also define this accordingly.

Should a deviation from normal operation be detected, users will be provided with a corresponding notification in the form of a warning or alert. Even notifications are “intelligent” now. This means that the algorithm used sets dynamic limits based on the individual pump’s vibration behavior and output. While the static limit values of a DIN standard were used previously, now the limits are carefully balanced in line with the algorithm.

Preventing unscheduled system downtime

The combination of sensors and smart algorithms along with the expert knowledge of the solution provider make it possible to detect signs of wear at an earlier point, for example, or avoid possible consequential damage as a result of maintenance measures which have not been performed. Unscheduled system downtime is avoided. Maintenance work can also be better prepared as users can create maintenance schedules for each machine. This also makes it possible to do a more efficient job of spare parts stocking. All of this translates to better operating reliability and greater availability as well as considerable cost savings.

Adding to this is a monitoring centre with vibration experts contributing their knowledge. To this end, they monitor and analyze the data stream for the user. The team at the monitoring centre identifies changing trends in the measurement data at an early stage, examines these in detail if so required and recommends repair and maintenance measures. In this context, it is not always about damage that may or may not have occurred, but also efficient pump operation. One fifth of the pumps equipped with this system could be repaired based on this measurement data. In five percent of the pumps fitted with this system, the initial measurement already revealed anomalies.

In developing this solution, the provider placed particular focus on IT security. Since the solution operates completely autonomously and independently of the customer’s actual network, an intrusion into the customer’s network is impossible. The solution provider is also careful to ensure maximum security in the handling of the data itself: All data acquired is encrypted using TLS 1.2 and sent to the provider’s gateway via the MQTT-SN application protocol. The data is then forwarded to the provider’s cloud via the secure mobile network.

Quickly connecting data silos

Should an evaluation or visualization tool already be installed on a customer network, the monitoring solution allows the user to synchronize the data of the solution provider with these systems using a standardized REST API data interface. Via the data interface, operators can conveniently integrate all information generated by the monitoring solution into their company’s existing solutions. This, in turn, makes it possible to use and manage data centrally as well as combine it with other data.

Practical applications

The pump monitoring data interface can be set up quickly. In one example, the solution provider implemented and tested the data interface in just two weeks for a customer in the chemical industry in order to monitor the pumps in the central compressed air supply system. The data collected from the pumps is now continually fed into the data centre via this interface. Interpreting the data is easy thanks to the standardized interface. The user knows at a glance whether the temperatures or vibrations are okay or whether a limit value has been exceeded. The user can also determine whether only a warning or an alarm is involved. The collected data is supplemented by other data on downtimes and production to provide relevant recommendations for action to maintenance staff and the system operator.

A large water supplier experienced elevated vibration levels last summer, at which point the experts at the monitoring centre contacted the customer and informed him about the condition of the pump and that bearing damage was imminent. The user then replaced the mechanical seal and a bearing. Soon after, the pump was quickly put back into operation. Major damage was prevented.

At a waste water plant, experts from the monitoring centre repeatedly noticed brief operation with very high vibration levels in excess of 13 mm/s. A glance at the frequency analysis revealed indications of unbalance. The customer then took a more in-depth look at the pump and discovered an excessive bearing clearance. Here, too, monitoring helped avoid consequential damage.

Identifying energy guzzlers – Minimizing carbon footprint

In addition to facilitating maintenance management, the monitoring solution also provides information on energy efficiency. After all, 75 percent of all pumps are not run at their optimum operating point! Pumps therefore consume much more energy than necessary. This not only leads to very high costs – which will continue to increase in the years to come – but also directly conflicts with the defined EU target of becoming climate neutral by the year 2050. The monitoring solution identifies the energy guzzlers among fixed-speed pumps in a system as well as optimization potential.

This, in turn, is based on the intelligent algorithm for determining the operating point. The load profile identifies the load range in which the pump is currently being run – part load, optimum, and overload – or whether the pump is currently out of operation. This helps inefficiencies to be detected and, consequently, potential energy or CO2 savings to be identified. The solution provider also offsets the emissions attributed to the production, use and waste disposal of the monitoring solution. This is yet another step toward sustainable system operation.

See also:

Philipp Bruhns02.10.2023