Fracturing equipment state monitoring and fault diagnosis system and method

Abstract

The invention discloses a fracturing equipment state monitoring and fault diagnosis system and method, and belongs to the technical field of oil and gas field fracturing equipment intelligence. The invention aims to solve the problems that in the prior art, monitoring depends on a single signal, fault early warning lags behind, and the misjudgment rate is high. The method comprises the following steps: collecting multi-source operation data of the fracturing pump in real time; establishing a theoretical pressure indicator diagram, and comparing the theoretical pressure indicator diagram with an actual indicator diagram generated by real-time data to realize first-stage fault judgment; an improved wavelet threshold noise reduction method is adopted to process the signals, and time domain and frequency domain features are extracted; and inputting the processed feature data into the combined diagnosis model for second-stage fault judgment. The joint diagnosis model combines a principal component analysis (PCA) model used for uncalibrated data anomaly detection and a BP neural network model used for calibrated data fault classification. Through deep fusion of the mechanism model and the data driving model, environmental interference is effectively resisted, and the diagnosis accuracy and the operation and maintenance efficiency are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent technology for oil and gas field fracturing equipment, specifically relating to a fracturing equipment condition monitoring and fault diagnosis system and method. Background Technology

[0002] As a core piece of equipment in the development of unconventional resources such as shale gas and tight oil and gas, fracturing pumps are responsible for injecting high-pressure fracturing fluids (such as proppant-carrying fluids and acid fluids) into the formation to form a complex fracture network. Their technical performance directly determines the construction efficiency, safety, and economy of hydraulic fracturing operations. With the growth of global energy demand and the development of unconventional oil and gas resources towards deeper formations (burial depth greater than 3500m), ultra-high pressure (pressure exceeding 100MPa), and large displacement (displacement greater than 3m³ / min), fracturing pumps face multiple technical challenges under extreme operating conditions.

[0003] Typical fracturing pumps are mostly three-cylinder or five-cylinder plunger-type structures, consisting of a power end (including crankshaft, gearbox, bearings, etc.) and a hydraulic end (including pump head, plunger, suction valve, discharge valve, etc.). During operation, the power end is subjected to huge alternating loads, while the hydraulic end is exposed to a complex stress environment of high pressure, high stroke, high abrasiveness, and chemical corrosion for a long time. This extreme working condition can easily lead to premature failure of key components (such as plunger seals, pump valves, crankshaft bearings), and the failure mode often exhibits multi-dimensional coupling characteristics of mechanical and hydraulic systems. Currently, the industry generally uses alarm systems based on single signals and fixed thresholds for condition monitoring of fracturing pumps, mainly relying on monitoring discrete data of vibration, pressure, or temperature. This method has significant technical limitations. Existing systems mostly analyze the vibration of the power-end bearings or the pressure fluctuations of the hydraulic end independently, lacking a systematic assessment of the mechanical-hydraulic coupling characteristics. Studies have shown that when the vibration signal exceeds the threshold (e.g., acceleration > 10g), the bearing wear has entered the middle and late stages (clearance > 0.5mm), and the risk of sudden shutdown increases by more than 3 times. In addition, single signals are easily affected by environmental interference such as manifold vibration transmission and load fluctuations in fracturing truck operations, with a misjudgment rate exceeding 25%. Fault data from North American shale gas fields in 2021 showed that cases of pressure pulses being misjudged as valve failures accounted for 32% of invalid maintenance. The conservative threshold strategy also forces operators to adopt a "preventive shutdown" mode, with an average of 120 hours of unplanned downtime per fracturing pump per year, and maintenance costs accounting for 18%-25% of the total equipment operating costs.

[0004] Despite attempts to introduce supplementary monitoring methods such as infrared thermal imaging and acoustic emission, bottlenecks remain, including difficulties in fusing multi-source heterogeneous data, low accuracy in fault feature extraction, and a lack of effective remaining life prediction models. Therefore, the industry urgently needs a new technology system that can deeply integrate high-precision sensor data, multi-physics coupling models, and intelligent algorithms to achieve health management of fracturing pumps throughout their entire lifecycle, improve fault identification rates, and reduce operation and maintenance costs. Summary of the Invention

[0005] The purpose of this invention is to solve the problems of existing fracturing pump monitoring technology, such as reliance on single-dimensional monitoring signals, delayed fault early warning, and high operation and maintenance costs, and to provide a fracturing equipment condition monitoring and fault diagnosis system and method based on multi-source information fusion and dynamic coupling analysis.

[0006] The objective of this invention is achieved through the following technical solution: The fracturing equipment condition monitoring and fault diagnosis system includes a multi-source information acquisition module, a mechanism model analysis module, a signal processing module, and a joint fault diagnosis module. The multi-source information acquisition module consists of a sensor group, a PLC, a voltage / current type multi-channel acquisition card, a switch, and a gateway, used to acquire real-time data from various measuring points of the fracturing equipment and integrate and forward it. The mechanism model analysis module generates theoretical pressure indicator diagrams and actual pressure indicator diagrams based on the rated and actual parameters of the tested equipment, respectively, and performs a first-level fault state judgment by comparing the differences. The signal processing module processes high-frequency vibration and high-frequency pressure signals, uses an improved wavelet algorithm for signal denoising, and then extracts and reduces the dimensionality of fault feature data in the time-frequency domain. The joint fault diagnosis module receives the fault feature data and performs a second-level fault state judgment, integrating at least two diagnostic models. It selects the corresponding diagnostic model for analysis based on whether the data has a pre-defined fault type.

[0007] The present invention also provides a corresponding method for monitoring the condition and diagnosing faults of fracturing equipment, characterized by comprising the following steps: Step 1: Real-time acquisition of operational data from fracturing equipment, including vibration, pressure, axial force, and key phase displacement; Step 2: Establish the theoretical pressure indicator diagram and diagnostic model of the fracturing equipment, and generate the actual pressure indicator diagram based on the collected real-time operating data. By comparing the characteristic differences between the theoretical pressure indicator diagram and the actual pressure indicator diagram, the first-level fault state judgment is performed. The theoretical pressure calculation formula for the liquid in the working chamber during the intake and exhaust process is as follows: in These are the liquid pressures within the working chamber during the inhalation and exhalation processes, respectively; These are the reference pressures for the inhalation and exhaust sides, respectively; The specific weight of the fluid; These are the head losses along the fluid in the suction and discharge pipes, respectively. These are the local head losses of the fluid in the suction and discharge pipes, respectively. These represent the fluid losses inside the pump during the suction and discharge processes; The average flow velocity inside the pipe; The flow rates within the suction and discharge tubes; It is the acceleration due to gravity; Step 3: Perform signal processing on the collected high-frequency data, including noise reduction, feature extraction, and feature dimensionality reduction; (1) Wavelet noise reduction Since the high-frequency signals generated during the operation of fracturing equipment are sometimes non-stationary, and wavelet transform, being a time-frequency analysis method, is adaptive to signals and is well-suited for processing non-stationary signals, this invention addresses the problem of effective information loss that can easily occur when the fixed threshold denoising method processes multi-layer wavelet coefficients. This invention employs a locally adaptive threshold, which can be dynamically adjusted according to the number of wavelet decomposition layers. Its calculation formula is as follows: in This is the original vibration signal; This represents the total number of wavelet decomposition levels. This represents the current wavelet decomposition level. These are the low-frequency coefficients of the nth layer; For the i-th layer, the high-frequency coefficients are: The adaptive threshold for the i-th layer; The standard deviation of noise; The length of the processed signal; To address the issues of signal discontinuity caused by hard thresholding functions, constant bias produced by soft thresholding functions, and computational complexity of traditional semi-soft thresholding functions, an improved semi-soft thresholding function is used to quantize wavelet coefficients. Its expression is as follows: in This represents the current wavelet decomposition level. It is a symbolic function; These are the high-frequency coefficients of the i-th layer after threshold filtering; For the i-th layer, the high-frequency coefficients are: The threshold value is used; It is a natural constant.

[0008] (2) Time-frequency domain feature extraction After denoising the high-frequency signal, multi-dimensional feature extraction is performed on the signal to comprehensively characterize the equipment status from the operational data. Specifically, time-domain features reflecting the statistical distribution and energy characteristics of the signal are extracted using the calculation formula for time-domain features, including maximum value, minimum value, peak-to-peak value, mean, variance, standard deviation, root mean square, kurtosis, and skewness. Then, the signal is transformed to the frequency domain using Fast Fourier Transform (FFT), and frequency-domain features reflecting its frequency structure and energy distribution are further extracted, including mean square frequency, centroid frequency, and frequency variance, thus forming a multi-dimensional feature vector.

[0009] (3) Feature dimensionality reduction Preferably, for feature dimensionality reduction, PCA dimensionality reduction method is used to map the original high-dimensional feature space to a low-dimensional principal component space. This method preserves the variance information of the original data to the greatest extent while eliminating the correlation between features, thus providing high-quality input for subsequent diagnostic models.

[0010] Step 4: The fault characteristic data is input into the joint fault diagnosis module for second-level fault status judgment. The joint fault diagnosis module, based on whether the fault characteristic data contains a pre-defined fault type, calls the corresponding internally integrated diagnostic model for analysis and outputs the diagnostic result. The joint fault diagnosis module integrates two models with different mechanisms: Based on the diagnostic model with calibrated fault data, a BP neural network model is preferred. The BP neural network includes an input layer, multiple hidden layers, and an output layer. It is trained using the backpropagation algorithm and can effectively learn and identify fault modes of complex systems, and perform accurate fault classification and prediction. The construction process is as follows: (1) Forward propagation calculation: Input the training set, calculate the output value through weights and thresholds, and the calculation formula is as follows: in For network layer indexing; For the node index of the current layer; This is the output of the i-th node in the k-th layer; For activation functions; This is the input for the i-th node in the k-th layer; This is the node index of the previous level; This represents the number of nodes in the previous layer. The connection weight between the j-th node in the previous layer and the i-th node in the k-th layer; This is the output of the j-th node in the (k-1)-th layer; This is the bias of the i-th node in the k-th layer.

[0011] (2) Error backpropagation: Calculate the error between the actual output value and the predicted output value, and adjust the weights and thresholds using the gradient descent method to reduce the error. The error backpropagation formula is: in This represents the update amount of the connection weights from the i-th neuron to the j-th neuron; The learning rate; For loss function Weights The partial derivatives of , i.e., the gradient; The weight adjustment formula is: in The weights connecting the i-th neuron and the j-th neuron; This represents the number of iterations. The weight value at the t-th iteration; This represents the amount of weight update; The learning rate; For loss function Weights The partial derivative of , i.e., the gradient.

[0012] (3) The BP neural network predicts the fault type, and the calculation formula is as follows: in This is the predicted output of the BP neural network; The trained BP neural network model; This is the standardized test data.

[0013] Based on diagnostic models using uncalibrated fault data, Principal Component Analysis (PCA) is the preferred model. Judgment is made by comparing control limits and statistical measures. The construction process is as follows: (1) Calculate the Q control limit using the following formula: in Q-control limit; It is the first moment; It is the second-order moment; For shape parameters; This is the critical value for the standard normal distribution.

[0014] (2) Calculate the SPE statistic. The formula is as follows: in Let SPE statistic be the value of the i-th sample. This is the standardized i-th test set sample; The projection matrix; This is the transpose of matrix P.

[0015] (3) Calculate the T² control limits and statistic. The calculation formula is as follows: in The control limit is T². The number of principal components selected; The number of samples in the training data; The 95th quantile of the F distribution is based on k and Nk degrees of freedom; Let T² be the T-statistic for the i-th sample; This is the standardized i-th test sample; The projection matrix; This is the transpose of matrix P; It is the inverse of the diagonal matrix, containing the first k eigenvalues; for Transpose of a vector.

[0016] (4) Comparison of control limits and statistics: If the data sample is detected three times consecutively... > or > This indicates that the sample data is an outlier. After storing all outlier samples, calculate the residual for each sample using the following formula: in Let be the residual of the i-th sample; This is the standardized i-th test sample; Let be the projection of the i-th test sample into the principal component space; This is the transpose of the projection matrix.

[0017] (5) Contribution Analysis: Based on the obtained fault data samples, the contribution of each variable in the samples is calculated sequentially, and the variable with the largest contribution is extracted for analysis to achieve accurate fault location. The calculation formula is as follows: in The contribution of the j-th variable to the i-th sample; Let be the residual of the j-th variable in the i-th sample.

[0018] Further technical solutions include a fault strategy scheduling module and a remote control module to further improve fault handling and monitoring efficiency. After a fault pre-alarm occurs, the fault strategy scheduling module can automatically generate vibration and pressure images, detailed text descriptions, and maintenance strategies corresponding to the current fault point. The remote control module uploads data to the cloud periodically through a host computer deployed in the field instrument room, enabling remote issuance of query and control commands.

[0019] The fracturing equipment condition monitoring and fault diagnosis system and method proposed in this invention have the following beneficial effects: (1) Improve fault diagnosis accuracy and early identification capability: By integrating vibration, pressure, temperature and flow sensor networks, a multi-level dynamic fault feature extraction model is constructed to achieve accurate identification of minor faults; (2) Significantly enhances the system's anti-interference capability and reliability: the false alarm rate is reduced from the industry average of 25% to less than 8%, and the system maintains stable operation under extreme conditions such as ultra-high pressure and high sand ratio; (3) Improve the level of intelligence and adaptability: The dynamic PCA threshold correction mechanism can adapt to complex working conditions such as displacement fluctuation and pressure change. (4) The verification effect is significant and has promotion and application value: when the training sample size exceeds 4000, the false alarm rate is stable within 8%. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the fracturing equipment monitoring system proposed in this invention; Figure 2 This is a schematic diagram of the fracturing equipment monitoring system proposed in this invention; Figure 3 This is a schematic diagram of the fault diagnosis process for fracturing equipment proposed in this invention; Figure 4 A comparison chart of theoretical and actual pressure indications for fracturing equipment; Figure 5 A comparison chart of actual vibration data for each pump chamber at the hydraulic end of a fracturing pump; Figure 6 A comparison chart of theoretical and actual discharge pressures for each pump chamber at the hydraulic end of a fracturing pump. Figure 7 A scatter plot of actual data on the hydraulic end suction pressure of the fracturing pump; Figure 8 A graph showing the change in false alarm rate for fault diagnosis of fracturing equipment. Detailed Implementation

[0021] To enhance understanding of the present invention, the following detailed description, in conjunction with embodiments, is provided. These embodiments are for illustrative purposes only and do not constitute a limitation on the scope of protection of the present invention. This embodiment uses the operating data of a GD500 fracturing pump as an example to verify the fault diagnosis method described in this invention. The main parameters of the equipment during operation are shown in Table 1: Table 1 Operating parameters of GD5000 fracturing pump like Figure 3 As shown, the fault diagnosis process for pressure equipment proposed in this invention adopts a phased and progressive structure: On the left, historical and real-time data are collected and classified, and trained using PCA and BP neural network models respectively to form a multi-dimensional fault diagnosis basic model library; on the right, the multi-source fusion fault diagnosis process forms a closed-loop intelligent diagnosis system. After inputting the data to be tested, it first generates preliminary diagnostic results through multi-source fusion analysis, and triggers branch processes through abnormal state judgment—if an abnormality is detected, the fault type is accurately output and the result verification stage is entered; if no abnormality is detected, it is determined to be a normal state. At the same time, the feedback mechanism of the verification stage can feed new data back into the PCA model to achieve dynamic optimization, ensuring that the system continuously adapts to changes in operating conditions.

[0022] like Figure 4 As shown, when the actual discharge pressure is 80MPa, the actual measured value rises slowly in the initial stage, fails to reach the theoretical pressure in the steady stage, and decreases synchronously in the final stage. However, due to the leakage, the rate of decrease and the final pressure also differ from the theoretical value. Figure 5 A comparison chart of actual vibration data of each pump chamber at the hydraulic end of the fracturing pump is provided, showing the actual vibration data of each pump chamber at the hydraulic end of the fracturing pump. Figure 6 This is a comparison chart of the theoretical discharge pressure and the actual discharge pressure of each pump chamber at the hydraulic end of a fracturing pump. The background shows the theoretical discharge pressure of each pump chamber at different times, and the top shows the actual discharge pressure. The comparison between the two can determine whether there is a pressure anomaly. Figure 7 This is a scatter plot of actual data on the suction pressure at the hydraulic end of the fracturing pump; all actual data are taken from the same time axis and correspond to each other. Figure 6 The measured discharge pressure exhibited periodic abnormal pressure fluctuations, which did not match the theoretical pressure fluctuations observed in each cylinder. Figure 5 The vibration curves show periodic peaks of varying degrees in each pump chamber, while... Figure 7 The measured suction pressure did not fluctuate significantly, indicating that there was a risk of early failure in the pump chamber and that the failure was most likely due to the discharge valve. The final model diagnosis determined that the failure was a valve leak in the discharge valve of pump chamber No. 3, and promptly generated vibration and pressure images, detailed text descriptions, and maintenance strategies corresponding to the current failure point.

[0023] pass Figure 8 It can be seen that when the training sample size is 1000~10000, the false alarm rate can be stabilized between 7% and 10%.

Claims

1. A fracturing equipment condition monitoring and fault diagnosis system, characterized in that, include: The multi-source information acquisition module consists of a sensor group, a PLC, a voltage / current type multi-channel acquisition card, a switch, and a gateway. It is used to collect real-time data from various measuring points of the fracturing equipment and integrate and forward it. The mechanism model analysis module is used to establish a theoretical pressure indicator diagram based on the design parameters and working conditions of the fracturing pump, and to generate an actual pressure indicator diagram based on the real-time operating data collected by the multi-source information acquisition module. By comparing the characteristic differences between the theoretical pressure indicator diagram and the actual pressure indicator diagram, the first-level fault state is judged. The signal processing module is used to process high-frequency vibration and high-frequency pressure signals. It uses an improved wavelet algorithm to reduce signal noise, thereby extracting fault feature data in the time and frequency domain and reducing its dimensionality. The joint fault diagnosis module is used to receive the fault feature data and perform a second-level fault status judgment. The joint fault diagnosis module integrates at least two diagnostic models. Depending on whether the data has a pre-defined fault type, the corresponding diagnostic model is selected for analysis.

2. The fracturing equipment condition monitoring and fault diagnosis system according to claim 1, characterized in that: It also includes a fault strategy scheduling module and a remote control module; after a fault pre-alarm occurs, the fault strategy scheduling module can automatically generate vibration and pressure images, detailed text descriptions and maintenance strategies corresponding to the current fault point; the remote control module uploads data to the cloud periodically through the host computer deployed in the field instrument room, enabling remote issuance of query and control commands.

3. A method for condition monitoring and fault diagnosis of fracturing equipment, characterized in that: Includes the following steps: Step 1: Real-time acquisition of operational data from fracturing equipment, including vibration, pressure, axial force, and key phase displacement; Step 2: Establish the theoretical pressure indicator diagram and diagnostic model of the fracturing equipment, and generate the actual pressure indicator diagram based on the collected real-time operating data. By comparing the characteristic differences between the theoretical pressure indicator diagram and the actual pressure indicator diagram, the first-level fault state judgment is performed. Step 3: Perform signal processing on the collected high-frequency data, including noise reduction, feature extraction, and feature dimensionality reduction; 1) Wavelet noise reduction processing: To address the issue of effective information loss that arises when using fixed threshold denoising methods to process multi-level wavelet coefficients, this invention employs a locally adaptive threshold that dynamically adjusts according to the number of wavelet decomposition levels. in This is the original vibration signal; This represents the total number of wavelet decomposition levels. This represents the current wavelet decomposition level. These are the low-frequency coefficients of the nth layer; For the i-th layer, the high-frequency coefficients are: The adaptive threshold for the i-th layer; The standard deviation of noise; The length of the processed signal; This invention also employs an improved semi-soft thresholding function to quantize wavelet coefficients, the expression of which is: in This represents the current wavelet decomposition level. It is a symbolic function; These are the high-frequency coefficients of the i-th layer after threshold filtering; For the i-th layer, the high-frequency coefficients are: The threshold value is used; It is a natural constant; 2) Time-frequency domain feature extraction The time-domain features that reflect the statistical distribution and energy characteristics of the denoised signal are extracted from the time-domain feature calculation formula; the signal is then transformed to the frequency domain by Fast Fourier Transform (FFT) to further extract the frequency domain features that reflect the frequency structure and energy distribution, thus forming a multi-dimensional feature vector. 3) Feature dimensionality reduction Principal component analysis (PCA) maps the original high-dimensional feature space to a low-dimensional principal component space, preserving the variance information of the original data to the greatest extent while eliminating the correlation between features, thus providing high-quality input for subsequent diagnostic models. Step 4: Input the fault feature data into the joint fault diagnosis module for second-level fault status judgment. The joint fault diagnosis module calls the corresponding internally integrated diagnostic model to analyze and output the diagnostic results based on whether the fault feature data has a pre-defined fault type.

4. The method for monitoring the condition and diagnosing faults of fracturing equipment according to claim 3, characterized in that: The joint fault diagnosis module mentioned in step 4 includes: Diagnostic models based on calibrated fault data: BP neural network model, used to classify and identify data with predefined fault type labels; Diagnostic models based on uncalibrated fault data: Principal Component Analysis (PCA) model, used for anomaly detection of data not identified as known fault types.

5. The method for monitoring the condition and diagnosing faults of fracturing equipment according to claim 4, characterized in that, The BP neural network model includes an input layer, multiple hidden layers, and an output layer. It is trained using the backpropagation algorithm and can effectively learn and identify fault modes in complex systems, enabling accurate fault classification and prediction. in This is the predicted output of the BP neural network; The trained BP neural network model; This is the standardized test data.

6. The method for monitoring the condition and diagnosing faults of fracturing equipment according to claim 4, characterized in that, The PCA principal component analysis model calculates the SPE statistic and T² statistic of the test sample: in Let SPE statistic be the value of the i-th sample. This is the standardized i-th test set sample; The projection matrix; for Transpose of a matrix; Let T² be the T-statistic for the i-th sample; for Transpose of a vector; It is the inverse of the diagonal matrix; The statistic is compared with the control limits obtained from training to determine whether the sample data is abnormal. If abnormal, the sample residuals are calculated. in Let be the residual of the i-th sample; This is the standardized i-th test sample; Let be the projection of the i-th test sample into the principal component space; This is the transpose of the projection matrix; Further contribution analysis was used to pinpoint the key variables causing the anomalies: in The contribution of the j-th variable to the i-th sample; Let be the residual of the j-th variable in the i-th sample.

Images