A method for quickly identifying leakage faults at the hydraulic end of a reciprocating pump

Abstract

The present invention relates to the field of leakage fault identification technology, and in particular to a method for quickly identifying leakage faults at the hydraulic end of a reciprocating pump. In the preprocessing stage, a suitable bandpass digital filter is first used to bandpass filter the collected vibration data; secondly, the filtered signal is enveloped; then, whether the reciprocating pump is leaking is determined based on whether the characteristic frequency of the envelope spectrum is significant; then, for the specially processed envelope signal, a cyclic window with a fixed duty cycle is used to identify the position with the largest sum of envelope values ​​within the window to locate the location of the reciprocating pump leakage; finally, the key information in the algorithm is visualized. The various key thresholds and parameters involved in the algorithm are adjusted based on a large amount of data collected on-site, and then fed back to the field for testing and application. Application results show that the accuracy of the present invention in identifying leakage faults at the hydraulic end of a reciprocating pump is above 90%, which has important reference value and practical economic value for guiding timely maintenance on-site in oil and gas fields.

Description

Technical Field

[0001] The present invention relates to the technical field of leakage fault identification, and in particular to a method for quickly identifying a leakage fault at the hydraulic end of a reciprocating pump. Background Art

[0002] Most onshore oilfields in my country suffer from poor reservoir properties and low formation energy, yet the scale of development continues to expand. Driven by the production imperative of "sufficient and effective water injection," and driven by the goal of achieving stable and increased oilfield production, water injection volumes have increased rapidly. The primary method for maintaining formation pressure in oilfields is to pump surface water into the formation using a water injection system to boost pressure. The plunger pump unit at a water injection station is the heart of the system, and its performance directly impacts production efficiency and profitability. This underpins the crucial role of plunger pumps in condition monitoring and fault diagnosis for oilfield production equipment. Plunger pumps operate under high loads, corrosive media, and complex structures, making them prone to failure. Currently, most oilfield operators rely on a reactive approach of planned maintenance, manual inspections, and emergency repairs. This approach fails to accurately assess the operating status of the plunger pumps, nor proactively predict potential failures. This not only reduces plunger pump utilization and service life, but also creates hidden dangers for the company's operations. During the oilfield modernization drive, the need for unmanned water injection stations was raised. This requirement relies on the technical foundation of condition monitoring and fault early warning for each piece of production equipment. This indirectly requires a clear understanding of the operating mechanisms, degradation trends, fault conditions, and their causes for the mechanical equipment.

[0003] The cutting-edge theories and achievements in reciprocating machinery fault diagnosis research at home and abroad indicate that current research on reciprocating machinery fault diagnosis methods focuses primarily on signal extraction and processing, with insufficient attention paid to fault mechanisms and identification under different operating conditions based on reciprocating machinery dynamics research. Furthermore, dynamic analysis of reciprocating machinery is often conducted as a theoretical study or for the purpose of reciprocating machinery design.

[0004] In many oilfields in my country, reciprocating pump maintenance still relies on empirical judgment, even relying on hearing close to the pump's valve box to identify valve failures. These methods are not only inaccurate but also severely damage hearing. Extensive research and retrieval of relevant domestic and international literature reveals that reciprocating pump fault diagnosis typically involves extracting vibration signals for processing and analysis, such as time-domain statistical analysis, frequency-domain analysis, and time-frequency analysis. While some research has been conducted on reciprocating pump fault diagnosis, and some equipment has been developed, such as a fault diagnosis system for the 3NB-1300C drilling pump developed by China University of Petroleum (East China) in collaboration with Sinopec and Jiangsu Polytechnic University, a comprehensive diagnostic system for field applications, particularly for hydraulic end failures, remains lacking. To address these issues, studying the mechanisms of reciprocating pump hydraulic end failures and quickly and accurately identifying hydraulic end leakage faults is crucial for improving the efficiency of reciprocating pumps in oilfields and ensuring safe production. Therefore, we propose a method to quickly identify leakage faults at the hydraulic end of a reciprocating pump. Summary of the Invention

[0005] Based on the technical problems existing in the background technology, the present invention proposes a method for quickly identifying leakage faults at the hydraulic end of a reciprocating pump. By tracking the vibration caused by the leakage, the success rate of on-site early warning is close to high, which solves the problem that the existing leakage identification technology not only has low accuracy but also makes it difficult to detect early leakage of pumps and valves.

[0006] The present invention provides the following technical solution: a method for quickly identifying a leakage fault at the hydraulic end of a reciprocating pump, comprising the following steps:

[0007] S1. Obtain a key phase signal from a speed sensor and calculate the speed based on the key phase signal; when the speed is higher than a threshold, obtain a vibration signal from a vibration sensor;

[0008] S2. Design a window function within a speed cycle based on the speed, and propose an initial window with a phase of zero;

[0009] S3. Input the window function and envelope signal into the leakage decision algorithm. The leakage decision includes two parts: calculating the leakage rate and the leakage phase. According to whether the characteristic frequency of the envelope signal spectrum is significant, a Boolean value of leakage is obtained, and the leakage rate of the most recent appropriate number of samples is calculated in real time. The envelope value in the window is calculated cyclically to find the position when the envelope value in the window is the largest. Combined with the initial phase determined by the installation position of the speed sensor, the leakage phase is calculated in real time and converted into the position of the leakage cylinder.

[0010] Preferably, the specific process of step S1 multi-signal joint acquisition is as follows:

[0011] S11, obtaining key phase signals from the speed sensor and data acquisition card arranged on site;

[0012] S12, setting the threshold to convert the key phase signal into a level signal;

[0013] S13, performing differential processing on the level signal to calculate the rotation speed;

[0014] S14, selecting an appropriate rotation speed as a threshold value based on the time of a single sample and computer performance, and determining whether the rotation speed is appropriate;

[0015] S15: If the speed is too low, return to step S11;

[0016] S16. If the rotation speed is appropriate, vibration data is obtained from multiple vibration sensors and data acquisition cards arranged on site;

[0017] S17, restoring actual vibration data according to the data acquisition card's data storage method;

[0018] S18. Retain the vibration data to be processed in combination with the rotation speed.

[0019] Preferably, the specific process of form design in step S2 is as follows:

[0020] S21. Extracting the number of sampling points of a single speed cycle according to the speed reaching the threshold;

[0021] S22, combining the sample sampling length and the number of sampling points in S21, determining the number of windows for a single sample;

[0022] S23, analyzing the proportion of the pump crankshaft sector area to determine the duty cycle of the window;

[0023] S24, constructing an initial window with a phase of zero;

[0024] S25 , combining the number of sampling points in step S21 and the initial window in step S24 , and shifting the phase.

[0025] Preferably, signal preprocessing is also included, and the specific process is as follows:

[0026] Extract leak samples and normal samples from a large number of samples collected on site;

[0027] Perform time-frequency analysis and comparison on leakage samples and normal samples to determine characteristic frequency bands;

[0028] The multiple vibration signals obtained are band-pass filtered, and the filtering frequency band is determined by the characteristic frequency band;

[0029] Perform Hilbert envelope on the filtered signal;

[0030] The envelope signal is then processed to make it not less than zero.

[0031] Preferably, the specific process of leak decision in step S3 is as follows:

[0032] S31, cyclically calculating and storing the accumulated value of the envelope signal within the window according to the processed envelope signal and the phase-shiftable window function;

[0033] S32. Extracting the index when the accumulated value is maximum from the stored accumulated values, and calculating the phase of the window function;

[0034] S33, performing spectrum analysis on the envelope signal;

[0035] S34, judging the leakage state according to the significance of the characteristic frequency in the envelope signal spectrum;

[0036] S35, the leakage status is stored in the form of a Boolean value;

[0037] S36, calculating the leakage rate of the most recent N samples;

[0038] S37, when the state is leakage, the phase is converted to leakage phase according to the window function of step S32; when the state is non-leakage, the phase is uniformly set to zero;

[0039] S38. Locate the leakage position according to the leakage phase in step S37.

[0040] Preferably, a visual display module is also included for displaying the filter signal, envelope signal, speed signal, leakage Boolean value, leakage rate, leakage phase and effective value of the filter signal, so that algorithm developers can verify the accuracy of the leakage.

[0041] The present invention provides a method for quickly identifying leakage faults at the hydraulic end of a reciprocating pump, which includes five parts: multi-signal joint acquisition, window design, signal preprocessing, leakage decision-making, and visual display. First, after acquiring the key phase signal, the pump speed is calculated in real time; if the speed is too low, the current sample is not processed; if the speed meets the standard, the window design and signal preprocessing are entered in parallel; the input of the leakage decision is the cyclic window function and the envelope signal after special processing, and the output is the leakage rate and leakage phase; the content of the visual display is the key information of the algorithm, which is used by the algorithm developers to verify the accuracy of the algorithm. By tracking the vibration caused by the leakage, the success rate of on-site early warning is close to high, which solves the problem of existing leakage identification technology, which not only has low accuracy but also has difficulty in detecting early leakage of pumps and valves. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is a flow chart of the method of the present invention;

[0043] Figure 2 is a flow chart of multi-signal joint acquisition of the present invention;

[0044] Figure 3It is a flow chart of the form design of the present invention;

[0045] Figure 4 is a flow chart of signal preprocessing of the present invention;

[0046] Figure 5 is a flow chart of leakage decision making of the present invention;

[0047] Figure 6 This is a diagram showing the main waveform examples of the present invention; DETAILED DESCRIPTION

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0049] The present invention provides a technical solution that mainly includes five parts: multi-signal joint acquisition, window design, signal preprocessing, leakage decision-making, and visual display. Multi-signal joint acquisition mainly involves acquiring key phase signals and vibration signals; window design mainly constructs a window function within a speed cycle; signal preprocessing performs characteristic frequency band filtering and envelope analysis on the vibration signal; leakage decision-making comprehensively determines whether the current vibration signal has leakage characteristics based on the window function and envelope signal; and visual display mainly presents key information of the algorithm in graphical form, facilitating comparison and verification of the algorithm's reliability.

[0050] like Figure 1 As shown, a method for quickly identifying a leakage fault at the hydraulic end of a reciprocating pump includes the following steps:

[0051] S1. Obtain a key phase signal from a speed sensor and calculate the speed based on the key phase signal; when the speed is higher than a threshold, obtain a vibration signal from a vibration sensor;

[0052] S2. Design a window function within a speed cycle based on the speed, and propose an initial window with a phase of zero;

[0053] S3. Input the window function and envelope signal into the leakage decision algorithm. The leakage decision includes two parts: calculating the leakage rate and the leakage phase. According to whether the characteristic frequency of the envelope signal spectrum is significant, a Boolean value of leakage is obtained, and the leakage rate of the most recent appropriate number of samples is calculated in real time. The envelope value in the window is calculated cyclically to find the position when the envelope value in the window is the largest. Combined with the initial phase determined by the installation position of the speed sensor, the leakage phase is calculated in real time and converted into the position of the leakage cylinder.

[0054] like Figure 2As shown, the specific process of step S1 multi-signal joint acquisition is as follows:

[0055] S11, obtaining key phase signals from the speed sensor and data acquisition card arranged on site;

[0056] S12, setting the threshold to convert the key phase signal into a level signal;

[0057] S13, performing differential processing on the level signal to calculate the rotation speed;

[0058] S14, selecting an appropriate rotation speed as a threshold value based on the time of a single sample and computer performance, and determining whether the rotation speed is appropriate;

[0059] S15: If the speed is too low, return to step S11;

[0060] S16. If the rotation speed is appropriate, vibration data is obtained from multiple vibration sensors and data acquisition cards arranged on site;

[0061] S17, restoring actual vibration data according to the data acquisition card's data storage method;

[0062] S18. Retain the vibration data to be processed in combination with the rotation speed.

[0063] Multi-signal acquisition involves acquiring both key phase and vibration signals. Key phase signals are acquired by using a high-frequency data acquisition card to convert the analog signal sensed by the speed sensor into a digital signal. The key phase signal is compared with a threshold and converted into a level signal. This level signal is then differentially processed to determine the pump's real-time speed. Leakage in reciprocating pumps is typically reflected in the pump's operating status, which can be characterized by the pump's speed. Pump leakage is associated with vibration signals, so vibration signal acquisition is controlled by the speed. If the speed is too low, the current sample is ineffective in determining leakage. If the speed meets the required criteria, vibration data is acquired from multiple vibration sensors and data acquisition cards deployed on-site. The actual vibration data is then restored using the data acquisition card's data storage method. The vibration data requiring processing is retained based on the speed, thereby acquiring multiple vibration signals.

[0064] like Figure 3 As shown, the specific process of form design is as follows:

[0065] S21. Extracting the number of sampling points of a single speed cycle according to the speed reaching the threshold;

[0066] S22, combining the sample sampling length and the number of sampling points in S21, determining the number of windows for a single sample;

[0067] S23, analyzing the proportion of the pump crankshaft sector area to determine the duty cycle of the window;

[0068] S24, constructing an initial window with a phase of zero;

[0069] S25 , combining the number of sampling points in step S21 and the initial window in step S24 , and shifting the phase.

[0070] The window design's startup is also controlled by speed. If the speed meets the conditions, the number of sampling points per speed cycle is extracted. The number of windows per sample is determined by combining the sample length and the number of sampling points per speed cycle. Based on the pump crankshaft's structural design, the proportion of the crankshaft cam's sector area is calculated to determine the duty cycle of the window function. The initial window phase can be set arbitrarily; for ease of calculation, it is uniformly set to zero. The window function's phase shifts between zero and the number of sampling points per speed cycle.

[0071] like Figure 4 As shown, it also includes signal preprocessing, the specific process is as follows:

[0072] Extract leak samples and normal samples from a large number of samples collected on site;

[0073] Perform time-frequency analysis and comparison on leakage samples and normal samples to determine characteristic frequency bands;

[0074] The multiple vibration signals obtained are band-pass filtered, and the filtering frequency band is determined by the characteristic frequency band;

[0075] Perform Hilbert envelope on the filtered signal;

[0076] The envelope signal is then processed to make it not less than zero.

[0077] Signal preprocessing involves bandpass filtering and signal envelope analysis. The passband design of the bandpass filter depends on the vibration frequency band caused by the leakage. This frequency band requires time-frequency analysis and comparison of leak samples and normal samples to determine the characteristic frequency band range. After bandpass filtering multiple vibration signals, the Hilbert envelope is applied and the envelope signal is processed to ensure it is not less than zero.

[0078] like Figure 5 As shown, the specific process of leakage decision in step S3 is as follows:

[0079] S31, cyclically calculating and storing the accumulated value of the envelope signal within the window according to the processed envelope signal and the phase-shiftable window function;

[0080] S32. Extracting the index when the accumulated value is maximum from the stored accumulated values, and calculating the phase of the window function;

[0081] S33, performing spectrum analysis on the envelope signal;

[0082] S34, judging the leakage state according to the significance of the characteristic frequency in the envelope signal spectrum;

[0083] S35, the leakage status is stored in the form of a Boolean value;

[0084] S36, calculating the leakage rate of the most recent N samples;

[0085] S37, when the state is leakage, the phase is converted to leakage phase according to the window function of step S32; when the state is non-leakage, the phase is uniformly set to zero;

[0086] S38. Locate the leakage position according to the leakage phase in step S37.

[0087] Leakage determination involves calculating the leak rate and the leak phase. The pump's leakage status is determined based on the significance of the envelope signal spectrum. The leak status is stored as a Boolean value, and the leak rate is calculated for the last N samples. Based on the specially processed envelope signal and a phase-shifted window function, the accumulated envelope signal within the window is cyclically calculated and stored. The index of the maximum accumulated value is extracted from the stored accumulated values, and the phase of the window function is calculated. Combined with the initial phase determined by the speed sensor's installation position, the leakage phase is calculated in real time and converted into the position of the leaking cylinder.

[0088] The visualization content is mainly divided into four parts: the first part is the bandpass filter signal, envelope signal and speed signal of the vibration signal; the second part is the spectrum of the envelope signal; the third part is the leakage rate and leakage phase; the fourth part is the effective value of the bandpass filter signal. The main waveform examples of the visualization are as follows: Figure 6 shown. Figure 6 It only means that the visual display can display multiple waveforms, and the specific text and data do not affect the integrity of this technical solution.

[0089] In this invention, the vibration signal is filtered and enveloped using a characteristic frequency band, and then a cyclic window is used to locate leak faults. Existing vibration identification uses the effective value of the vibration signal. Both methods have been tested repeatedly in oil and gas well sites. Existing techniques use vibration effective value thresholds or trends to detect early leaks, often leading to false alarms or missed alarms. This invention, by tracking the vibrations caused by leaks, achieves a near 100% success rate for on-site early warnings.

[0090] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A method for quickly identifying a reciprocating pump hydraulic end leakage fault, comprising the following steps: S1. Obtain a key phase signal from a speed sensor and calculate the speed based on the key phase signal; when the speed is higher than a threshold, obtain a vibration signal from a vibration sensor; S2. Design a window function within a speed cycle based on the speed, and formulate an initial window with a phase of zero; The specific process of form design in step S2 is as follows: S21. Extracting the number of sampling points of a single speed cycle according to the speed reaching the threshold; S22, combining the sample sampling length and the number of sampling points in S21, determining the number of windows for a single sample; S23, analyzing the proportion of the pump crankshaft sector area to determine the duty cycle of the window; S24, constructing an initial window with a phase of zero; S25, combining the number of sampling points in step S21 and the initial window in step S24, and shifting the phase; S3. Input the window function and envelope signal into the leakage decision algorithm. The leakage decision includes calculating the leakage rate and leakage phase. Based on whether the characteristic frequency of the envelope signal spectrum is significant, a Boolean value is obtained to indicate whether there is a leakage. The leakage rate of the most recent appropriate number of samples is calculated in real time. The envelope value within the window is cyclically calculated to find the position where the envelope value within the window is the largest. Combined with the initial phase determined by the speed sensor installation position, the leakage phase is calculated in real time and converted into the position of the leakage cylinder. The specific process of leak decision in step S3 is as follows: S31, cyclically calculating and storing the accumulated value of the envelope signal within the window according to the processed envelope signal and the phase-shiftable window function; S32. Extracting the index when the accumulated value is maximum from the stored accumulated values, and calculating the phase of the window function; S33, performing spectrum analysis on the envelope signal; S34, judging the leakage state according to the significance of the characteristic frequency in the envelope signal spectrum; S35, the leakage status is stored in the form of a Boolean value; S36, calculating the leakage rate of the most recent N samples; S37, when the state is leakage, the phase is converted to leakage phase according to the window function of step S32; when the state is non-leakage, the phase is uniformly set to zero; S38. Locate the leakage position according to the leakage phase in step S37.

2. A method for quickly identifying a reciprocating pump hydraulic end leakage fault according to claim 1, characterized in that: The specific process of step S1 multi-signal joint acquisition is as follows: S11, obtaining key phase signals from the speed sensor and data acquisition card arranged on site; S12, setting the threshold to convert the key phase signal into a level signal; S13, performing differential processing on the level signal to calculate the rotation speed; S14, selecting an appropriate rotation speed as a threshold value based on the time of a single sample and computer performance, and determining whether the rotation speed is appropriate; S15: If the speed is too low, return to step S11; S16. If the rotation speed is appropriate, obtain vibration data from multiple vibration sensors and data acquisition cards arranged on site; S17, restoring actual vibration data according to the data acquisition card's data storage method; S18. Retain the vibration data to be processed in combination with the rotation speed.

3. The method for quickly identifying a reciprocating pump hydraulic end leakage fault according to claim 1, characterized in that: It also includes signal preprocessing, the specific process is as follows: Extract leak samples and normal samples from a large number of samples collected on site; Perform time-frequency analysis and comparison on leakage samples and normal samples to determine characteristic frequency bands; The multiple vibration signals obtained are band-pass filtered, and the filtering frequency band is determined by the characteristic frequency band; Perform Hilbert envelope on the filtered signal; The envelope signal is then processed to make it not less than zero.

4. The method for quickly identifying a reciprocating pump hydraulic end leakage fault according to claim 1, characterized in that: It also includes a visualization module for displaying the filtered signal, envelope signal, speed signal, leakage Boolean value, leakage rate, leakage phase and the effective value of the filtered signal, allowing algorithm developers to verify the accuracy of the leakage.

Images