In the oil and gas industry, nothing is more important than ensuring that electrically driven motors and motor-driven machines are operating at peak performance 24/7. The average industrial facility employs a large number of electrically driven assets, and 10% to 20% of those are extremely critical to the everyday processes involved in regular operations.

With so much depending upon those systems, unexpected failures can be prohibitively expensive. Not only can companies lose money due to unrecoverable lost time and production, but they also have to shoulder emergency maintenance costs and excess energy consumption from inefficiently operating machines that continue to run until breaking down. Additionally, the potential safety and environmental liabilities arising from catastrophic failures are inestimable, and it can take an organization many years to recover from such failures.

Considering everything that is at stake, oil and gas operators need affordable asset performance monitoring tools that can give them reliable and actionable information quickly. This includes not only electrical or mechanical failure detection but also asset monitoring tools that can accurately offer recommendations to improve energy efficiency. In the past companies have traditionally relied on vibration analysis to monitor the motors; however, vibration analysis is costly, invasive, and sometimes can deliver less-than-reliable and inconsistent information when it comes to the state of mechanical systems.

In contrast, continuous monitoring through electric waveform analysis is much more accurate and reliable and does not require access to the equipment itself. Rather than monitoring the effect of degraded assets on production over time, waveform analysis roots out

the cause of mechanical failures well before they happen because it consistently monitors the condition of an operation’s assets and checks it against established baselines. The result is real-time monitoring of not only critical equipment within each asset class but an entire facility’s motor population.

Using electrical waveform analysis

Many companies are reluctant to implement continuous reliability assessment systems because of increased cost and potential for a lengthy installation process. With the Veros electrical waveform analysis, however, a small monitor is installed inside a facility’s switch room, and setup time takes a matter of hours rather than days or weeks. With this technology, engineers and executives can track small and slowly changing variations in the electrical waveforms. This high degree of sensitivity to patterns provides more helpful and immediate results. Additionally, the information can be hosted in a company’s own datacenter or a secure private cloud and offers a clear and detailed dashboard view of detected faults, history, trends, and detailed information about the specific part that is expected to fail.

System protocols

The predictive intelligence platform (PIP) collects waveform data and uses proprietary machine learning algorithms to analyze and interpret those data continuously in real time. The software identifies and distinguishes the sources of waveform distortion and whether the distortion is caused by changes in incoming grid power, driven process, or asset conditions. Electrical problems are detected by empirically developing and tracking system impedance models. Classification and isolation of faults are accomplished by a combination of machine learning methods based on classifiers and specific spectral fingerprints of faults.

In addition, the electrical waveform analysis software learns the specific fingerprint signatures of monitored power trains, sets condition alarm thresholds, and then automatically enters the assessment mode to continually check for issues. Because it detects abrupt changes to and preexisting conditions of the monitored assets, the electrical waveform analysis assesses equipment by comparing actual observed behavior to previous levels. This helps engineers identify failures before they happen rather than relying on manual inspections to uncover potential (or already existing) problems.

Nothing is installed on the machines monitored; instead, predictive intelligence monitors (PIMs) continuously acquire electrical waveforms at the motor switches at high sampling rates. Data are transferred wirelessly to the PIP server, which continuously analyzes these waveforms, identifying impending faults and assessing energy efficiency. The system then produces predictive and actionable intelligence from each asset that operators want to monitor, detecting both electric and mechanical faults. Thus, the waveform analysis can detect anything from misalignment in bearings to eccentricity in rotors.

That information also may be fed into an organization’s enterprise asset management (EAM) system, which in turn will kick off workflows based on those results such as ordering replacement parts and initiating related approval processes. To ensure that the entire asset team is always apprised of asset status, future problem alerts can be sent on a regular basis to anyone who needs to know.

Real-world systems

Chaparral Energy Inc. implemented the Veros waveform analysis in its operations in the Midcontinent and Permian basin. Chaparral is nonintrusively monitoring 13 motors ranging from 350 hp to 1,500 hp, 10 COcompressors, two refrigeration compressors, and one COpump.

A failed COcompressor could cause COvolumes delivered to the pipeline to fall by 50% or more, which would result in lower product volume and costly repairs. To head off that possibility, Chaparral installed a continuous monitoring solution that employed waveform analysis to detect any potential issues.

Using a dedicated server to access the software, the project team installed 13 noninvasive PIMs in the Liberal, Kan., COcompression facility to monitor the compressors. Not long after going live, the new software detected a problem in three of the compressors. Rather than having to worry about all of the compressors failing, this allowed the operator to focus its limited resources on the problems with just those three.

Chaparral now uses the PIP as a real-time and future view of this plant’s operations to establish a more methodical way to decrease the company’s operating expenses and avoid losses in production. Moreover, Chaparral’s engineers and facility managers shifted to a predictive maintenance system, allowing them to avoid taking equipment offline unless it is completely necessary. In the end the software helped Chaparral save approximately US $100,000 by identifying key mechanical problems, disregarding false alarms, and avoiding unnecessary repairs on entire systems.

The Veros continuous electrical waveform analysis, integrated with EAM solutions, not only provides actionable data that oil and gas operators can use in the day-to-day monitoring and maintenance of their electrical and mechanical assets, but it also saves valuable time and money in the long run by helping operators avoid needless repairs and head off potential failures before they happen.

Comments