A control system for online identification of the hydraulic end seal status of a fracturing pump

Abstract

This invention provides a control system for online identification of the hydraulic end seal status of fracturing pumps, comprising: a data acquisition module, a data processing module, a status diagnosis module, a PHM module, and an output module. This invention aims to address the industry pain points of monitoring the hydraulic end seal status of fracturing pumps by constructing a fault prediction and health management system that integrates online identification, lifespan prediction, and maintenance recommendations. It identifies early leakage states of the seals, adapts to different operating conditions to predict remaining lifespan, and forms a closed loop for seal status management, thereby improving the operational stability of fracturing pumps.

Description

Technical Field

[0001] This invention relates to the field of equipment condition monitoring and fault prediction technology, particularly to the condition identification and health management technology of core equipment in fracturing operations, specifically a control system for online identification of the hydraulic end sealing status of fracturing pumps. Background Technology

[0002] In oil and gas extraction, fracturing pumps are the core power equipment for fracturing operations. As a vulnerable and critical component, the working condition of hydraulic end seals directly determines the operational stability and safety of fracturing pumps. As oil and gas field development moves towards deep and ultra-deep wells, fracturing operation conditions are becoming increasingly harsh (high pressure, high stroke rate, and highly corrosive media), and the requirements for monitoring and managing the hydraulic end seal status of fracturing pumps are constantly increasing.

[0003] In recent years, research on fracturing pump condition monitoring in the industry has gradually shifted from traditional manual inspections to online monitoring. Methods for monitoring sealing conditions based on pressure and temperature signals have emerged, and some technologies have incorporated vibration signal analysis, enabling preliminary identification of sealing faults and promoting the intelligent upgrading of fracturing pump monitoring technology. However, existing technologies still have many limitations. For example, most monitoring systems can only perform post-failure diagnosis of sealing faults, failing to identify early leakage conditions and providing early warnings. Some systems use a single model to calculate the remaining life of the seals, without considering the differences in seal damage evolution under normal and high-pressure conditions, resulting in low life prediction accuracy. Furthermore, existing technologies lack a complete fault prediction and health management (PHM) system, failing to achieve closed-loop management from condition identification to life prediction to maintenance recommendations. This leads to frequent sudden seal failures, causing unplanned downtime, increasing operating costs, and even safety accidents due to seal leaks.

[0004] Therefore, there is an urgent need in this field for a control system for online identification of the hydraulic end sealing status of fracturing pumps to solve the above-mentioned technical problems. Summary of the Invention

[0005] This invention provides a control system for online identification of the hydraulic end seal status of fracturing pumps. It aims to address the aforementioned industry pain points in monitoring the hydraulic end seal status of fracturing pumps by constructing a fault prediction and health management system that integrates online identification, life prediction, and maintenance recommendations. The system identifies early leakage states of the seal, adapts to different operating conditions to predict remaining life, forms a closed loop for seal status management, and improves the operational stability of fracturing pumps.

[0006] This invention provides a control system for online identification of the hydraulic end sealing status of a fracturing pump, comprising:

[0007] The data acquisition module is used to collect the operating status parameters of the hydraulic end of the fracturing pump in real time;

[0008] A data processing module, connected to the data acquisition module, is used to preprocess and extract features from the acquired operating status parameters to generate feature data that can characterize the sealing state.

[0009] The status diagnosis module is connected to the data processing module and has a built-in sealing status recognition model. It is used to identify the current sealing status of the hydraulic end online based on the feature data. The sealing status includes normal status, early leakage status and severe leakage status.

[0010] The PHM module, connected to the condition diagnostic module, is used to predict the remaining service life (RUL) of the seal based on the identified seal condition and its changing trend, and to generate maintenance recommendations.

[0011] The output module is connected to the status diagnostic module and the PHM module respectively, and is used to output the sealing status, remaining service life (RUL), and maintenance recommendations.

[0012] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0013] 1. This invention realizes online hierarchical identification of sealing status. By collecting multi-dimensional parameters through multi-source sensing and combining Mahalanobis distance to eliminate parameter correlation and dimensional differences, a comprehensive sealing index is obtained through nonlinear transformation. This index can distinguish between normal, early leakage, and severe leakage states, solving the problem that existing technologies cannot identify early leakage and enabling early warning of faults.

[0014] 2. This invention improves the accuracy of predicting the remaining life of seals. For two typical operating conditions, namely conventional and high-pressure, two different damage accumulation models are designed, namely power law and exponential, to adapt to the damage evolution characteristics of seals under different operating conditions, avoid the prediction bias of a single model, and provide life data support for health management.

[0015] 3. This invention forms a complete PHM closed-loop management system. From multi-source data acquisition, feature extraction, and status diagnosis to life prediction, maintenance suggestion generation, and result output and alarm, each module works together to not only output status information in real time, but also generate targeted maintenance suggestions based on remaining life and operating conditions. This realizes the transformation from reactive maintenance to proactive maintenance, greatly reducing the probability of sudden failure of seals and reducing unplanned downtime.

[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention, but do not constitute a limitation thereof; in the drawings:

[0018] Figure 1 This is a schematic diagram of the structure of a control system for online identification of the hydraulic end sealing status of a fracturing pump provided by the present invention;

[0019] Figure 2 This is a flowchart of the sealing state recognition model in this invention;

[0020] Figure 3 This is a flowchart of the PHM module in this invention. Detailed Implementation

[0021] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0022] Example 1:

[0023] This invention provides a control system for online identification of the hydraulic end sealing status of a fracturing pump. Please refer to [link to relevant documentation]. Figure 1 ,include:

[0024] The data acquisition module is used to collect the operating status parameters of the hydraulic end of the fracturing pump in real time;

[0025] The data processing module, connected to the data acquisition module, is used to preprocess and extract features from the acquired operating status parameters to generate feature data that can characterize the sealing state.

[0026] The status diagnosis module is connected to the data processing module and has a built-in sealing status recognition model. It is used to identify the current sealing status of the hydraulic end based on feature data. The sealing status includes normal status, early leakage status and severe leakage status.

[0027] The PHM module, connected to the Condition Diagnostics module, is used to predict the remaining service life (RUL) of the seal based on the identified seal condition and its changing trends, and to generate maintenance recommendations.

[0028] The output module connects to the condition diagnostic module and the PHM module respectively, and is used to output the sealing status, remaining service life (RUL), and maintenance recommendations.

[0029] Specifically, this embodiment constructs a process for assessing the sealing status of the hydraulic end of a fracturing pump, from data acquisition to processing, diagnosis, life prediction, and output. The data acquisition module, the system's perception layer, acquires real-time operating parameters of the hydraulic end of the fracturing pump, providing a data foundation for subsequent analysis. The data processing module performs signal conditioning and feature extraction on the raw acquired parameters, filtering out feature data that effectively characterizes the sealing status and eliminating invalid information. The status diagnosis module analyzes and processes the feature data using a built-in sealing status recognition model, enabling online identification of three states: normal sealing, early leakage, and severe leakage. The PHM (Predictive Health Management) module predicts the remaining service life of the seals based on the status diagnosis results and status change trends, and generates targeted maintenance recommendations based on the prediction results, achieving an upgrade from fault diagnosis to health management and avoiding unplanned downtime. The output module, as the system's interaction layer, displays and alarms the sealing status, remaining service life, and maintenance recommendations, providing intuitive decision-making basis for on-site operators, thus forming a closed-loop online sealing status identification and health management system.

[0030] In one implementation, the data acquisition module includes:

[0031] Multiple pressure sensors are installed in the suction chamber and discharge chamber of the hydraulic end, respectively, to collect fluid pressure signals inside the pump;

[0032] At least one flow sensor is installed on the discharge line at the hydraulic end to collect the pump discharge flow signal;

[0033] At least one temperature sensor is installed on the hydraulic end housing or in the internal flow channel to collect the temperature signal of the hydraulic end;

[0034] At least one vibration acceleration sensor is installed on the hydraulic end valve box or pump head body to collect vibration signals of the pump head body.

[0035] Specifically, in this embodiment, pressure sensors are distributed in the suction chamber and discharge chamber. The collected pressure signals can directly reflect the pressure difference changes on the sealing surface. When the seal leaks, it will cause abnormal pressure distribution in the cavity, which is a core characteristic parameter of the sealing state. The flow sensor is installed in the discharge pipeline, and the collected discharge flow signal is used to help determine the flow loss caused by the seal leakage. The temperature sensor is installed in the shell or internal flow channel. Wear and leakage of the seal will generate frictional heat or fluid turbulence heat, which will cause local temperature rise. Therefore, the temperature signal provides a thermal characterization of the sealing state. The vibration acceleration sensor is installed in the valve box or pump head body. The seal failure will cause changes in the vibration characteristics of the pump head body and valve box. The vibration signal provides a dynamic characterization of the sealing state.

[0036] In one implementation, the data processing module includes:

[0037] The signal conditioning unit is used to filter and amplify pressure signals, flow signals, temperature signals, and vibration signals;

[0038] The feature extraction unit, connected to the signal conditioning unit, is used to extract frequency domain features and time domain features from the conditioned signal.

[0039] Frequency domain characteristics include spectral entropy obtained from vibration signal spectral analysis. , ;in, For the first One frequency component, For the first The proportion of the energy of each frequency component to the total signal energy This represents the total number of frequency components.

[0040] Time-frequency domain features include the energy spectral density reconstructed from the pressure signal after wavelet packet decomposition. , ;in, For the first Wavelet packet reconstructed signal for each frequency band For time.

[0041] Specifically, in this embodiment, the signal conditioning unit filters and amplifies the analog signals collected by the multi-sensor unit. Filtering can remove noise signals such as electromagnetic interference and mechanical vibration at the scene, while amplification can amplify the weak effective signal to an amplitude suitable for subsequent analysis, ensuring the effectiveness and signal-to-noise ratio of the signal. The feature extraction unit extracts frequency domain and time-frequency domain features from the conditioned clean signal. These features can effectively uncover the sealing state information hidden in the signal and are more recognizable than the original time domain signal.

[0042] In the formula, Let i be the i-th frequency component obtained after Fourier transforming the vibration signal. Let be the ratio of the energy of the i-th frequency component to the total energy of the vibration signal. This represents the total number of frequency components after the Fourier transform of the vibration signal; and and The results were obtained by performing a fast Fourier transform on the conditioned vibration signal. It is obtained by calculating the ratio of the energy of a single frequency component to the sum of the energies of all frequency components; The information entropy term is used to calculate the energy proportion of each frequency component of the vibration signal, reflecting the uniformity of the frequency energy distribution; spectral entropy. Based on the frequency domain characteristics of information entropy theory, when the seal is normal, the frequency energy distribution of the pump head vibration is relatively uniform. When the value is large, and the seal leaks or wears, the frequency energy of the vibration signal will concentrate in a specific frequency band, reducing the uniformity of distribution. The value decreases, through The changes in these parameters can quantitatively characterize the frequency characteristics of the vibration signal, and thus reflect the sealing status.

[0043] In the formula, The time-domain signal obtained by reconstructing the j-th frequency band after wavelet packet decomposition of the pressure signal is obtained by performing wavelet packet decomposition on the conditioned pressure signal, decomposing the signal to a preset frequency band, and reconstructing the wavelet packet coefficients of the j-th frequency band; where... The energy density of the reconstructed signal in the j-th frequency band is The energy spectral density represents the total energy of that frequency band by integrating the energy density over the time domain; Based on the time-frequency domain features of wavelet packet decomposition, it can analyze pressure signals in a two-dimensional space of time and frequency. Leakage in the seal causes significant changes in the energy of the pressure signal in a specific frequency band. By extracting the corresponding frequency band... It can capture the time-frequency characteristics of pressure signals, thus achieving effective characterization of the sealing state.

[0044] In a preferred embodiment, the feature extraction unit performs wavelet packet decomposition on the pressure signal using the Daubechies 6 (db6) wavelet basis function for a 3-level wavelet packet decomposition. The signal is decomposed into 8 equal-width frequency bands, and the j-th node of the 3rd level (e.g., the detail component corresponding to node (3,1) or (3,2)) containing leakage feature information is selected for single-branch reconstruction to obtain the reconstructed signal. Based on the sampling theorem and the physical characteristics of fracturing pumps, this preset frequency band typically covers the mid-to-high frequency range of 500Hz to 2000Hz, which is most sensitive to high-frequency fluid pulsations and shocks caused by seal leaks.

[0045] In one implementation, please refer to Figure 2 The sealing condition identification model is used to calculate the comprehensive sealing index. Comprehensive sealing index Used for quantitative evaluation of sealing condition; the processing procedure of the sealing condition identification model is as follows:

[0046] Constructing feature vectors ,in This represents the pressure difference between the intake and exhaust chambers under the current operating conditions. For the corresponding standard pressure difference, The reference spectral entropy is the value under normal sealing conditions. This is the reference energy spectral density for the corresponding frequency band under normal sealed conditions; The current temperature. Reference temperature;

[0047] Calculate eigenvectors relative to the mean of the feature vector of the normal state Mahalanobis distance , ;in, Covariance Matrix This data was obtained from historical normal operation statistics.

[0048] The Mahalanobis distance is mapped to a comprehensive sealing index through a nonlinear transformation. , ;in, and This is the preset sensitivity adjustment constant.

[0049] Specifically, feature vectors It is a four-dimensional normalized eigenvector, containing normalized values ​​of four dimensions: pressure difference, spectral entropy, energy spectral density, and temperature. The pressure signal is obtained by subtracting the pressure signals collected by the pressure sensors in the suction chamber and the discharge chamber. Data is collected directly from a temperature sensor. , Calculated by the feature extraction unit described above; , , , When the hydraulic end seal of the fracturing pump is in normal condition, the baseline values ​​of the corresponding parameters can be set by experience or obtained by statistical fitting of a large amount of historical normal operation data. At the same time, it is necessary to normalize each parameter to eliminate the dimensional differences of different parameters (such as pressure in MPa, temperature in ℃, and spectral entropy as dimensionless) so that each characteristic parameter is comparable and avoids the characteristic analysis bias caused by the dimensional difference of a single parameter.

[0050] This is the mean vector of the four-dimensional eigenvectors under normal sealing conditions of the fracturing pump. The covariance matrix corresponding to the eigenvectors; historical normal operation data refers to the eigenvector sample set collected during the first operating cycle after the fracturing pump has new seals installed and has gone through the break-in period, when it is running stably and without any signs of leakage;

[0051] Sure and The process is as follows:

[0052] During the above-mentioned normal operation, feature vectors from N sample points are continuously collected. Outliers or anomalies caused by sensor interference, communication interruptions, etc., are eliminated to ensure that the sample data used in the calculation can truly reflect the distribution of the normal state; the collection time span should usually cover at least one complete job cycle to ensure the statistical representativeness of the sample;

[0053] Mean vector It is obtained by taking the arithmetic mean of the feature vectors of N samples, that is... Covariance matrix It is obtained by calculating the covariance between each feature, i.e. .

[0054] In a preferred embodiment of the present invention and It is set as a static baseline value, meaning it is calculated once based on the initial normal operating data after the first run or maintenance, and remains unchanged unless there are major changes to the equipment hardware. The purpose of this is to establish a fixed health baseline, so that subsequent Mahalanobis distance calculations have a stable and consistent reference system.

[0055] It is the inverse of the covariance matrix. is the transpose of the difference between the eigenvector and the mean vector; where, and This was obtained by statistically calculating the historical feature vector data of the hydraulic end seal of the fracturing pump under normal conditions. It is obtained through matrix inversion; in the formula... Used to reflect the deviation between the current feature vector and the mean vector of the normal state. The weighted sum of squared deviations, covariance matrix The correlation between various feature parameters is considered, and its inverse matrix achieves a weighted correction of the feature space; Mahalanobis distance. It is a multidimensional spatial distance that takes into account the correlation of feature parameters and dimensional normalization, and is used to characterize the degree of deviation between the current feature vector and the set of feature vectors in the normal state.

[0056] and It is obtained through simulation tests of fracturing pump seal failures and fitting of field operation data, and can be adjusted according to the diagnostic sensitivity requirements of actual application scenarios; in the formula, This is the normalization and sensitivity adjustment term for the Mahalanobis distance. The larger the value, the more sensitive the model is to changes in Mahalanobis distance; This is a nonlinear exponential transformation term used to map the Mahalanobis distance to the interval (0,1]. This is the final nonlinear transformation term, used to convert the mapped result into a comprehensive sealing index; where the comprehensive sealing index... It is a dimensionless exponent obtained by nonlinear transformation of the Mahalanobis distance, used to quantitatively evaluate the sealing condition, with a value range of [0,1). With comprehensive sealing index There is a positive correlation. The larger the value, the greater the deviation of the current feature vector from the normal state, and the worse the sealing condition. The larger the value, the better. The reason for using nonlinear transformation in this embodiment is twofold: first, to map the dimensionless continuous value of the Mahalanobis distance to a fixed interval of [0,1), making the evaluation results of the sealing state more intuitive and easier to judge; second, to achieve this through exponential nonlinear transformation and sensitivity adjustment constants. The model's sensitivity to changes in sealing condition can be adjusted according to actual needs, enabling accurate identification of early leaks (early leaks). A slight increase can be amplified through nonlinear transformation, making... It presents identifiable increments, avoiding the problem of unclear early leakage characteristics caused by linear transformation.

[0057] In one embodiment, the sealing state identification model further includes:

[0058] By comparing comprehensive sealing index The relationship between the data and the first and second preset thresholds is used to identify the sealing status online.

[0059] like If the first preset threshold is reached, it is identified as a normal state;

[0060] If the first preset threshold The second preset threshold is then identified as an early leakage state;

[0061] like If the second preset threshold is reached, it will be identified as a serious leakage state.

[0062] Specifically, the first and second preset thresholds are obtained through bench tests, field fault simulation tests, and statistical analysis of actual operating data on the hydraulic end sealing state of fracturing pumps: a comprehensive sealing index is collected under three conditions: normal sealing, early leakage, and severe leakage. The sample data was analyzed using statistical methods (such as cluster analysis and threshold optimization) to determine the critical values ​​that could most accurately distinguish the three states, which served as the first and second preset thresholds. At the same time, the thresholds were fine-tuned according to the model, operating conditions and sealing type of the fracturing pump on site to ensure the adaptability and recognition accuracy of the thresholds.

[0063] If the first preset threshold The second preset threshold is then identified as an early leakage state. At this time, the seal has slight wear or leakage, and the failure can be prevented from escalating through early maintenance.

[0064] like If the second preset threshold is reached, it is identified as a serious leakage state. At this point, the seal is close to failure and the machine must be shut down immediately for maintenance to prevent equipment damage and safety accidents.

[0065] In one implementation, please refer to Figure 3 The PHM module is used to determine the overall sealing index. Based on the current operating parameters, two different nonlinear damage accumulation models are used to calculate the remaining service life (RUL), which are adapted to the conventional operating conditions and high-pressure operating conditions of fracturing pumps, respectively.

[0066] Before calculating the remaining useful life (RUL), the following are also included:

[0067] Define the current cumulative damage amount The relationship with the comprehensive sealing index is as follows: ;in This is the preset damage coefficient.

[0068] Specifically, The preset damage coefficient is used to characterize the influence of the comprehensive sealing index on the damage of the seal. It is obtained by fitting the accelerated life test, fatigue damage test and field operation data of the seal. It can be adjusted according to the material and model of the seal and the operating conditions of the fracturing pump. This is the damage amplification term of the comprehensive sealing index, used to reflect the quantitative contribution of the comprehensive sealing index to the damage amount; This is a nonlinear exponential transformation term used to map the comprehensive sealing index to the (0,1] interval; This is a damage conversion term that converts the mapping result into cumulative damage.

[0069] Current cumulative damage It is a dimensionless index with a value range of [0,1), used to quantitatively characterize the cumulative damage of a seal from the time it is put into operation to the present moment. The closer the value is to 1, the more severe the cumulative damage to the seal, and the closer it is to failure; among them, and A positively correlated nonlinear relationship. The larger the value, the worse the sealing condition, the more severe the wear and leakage of the seal, and the corresponding current cumulative damage. The larger the value, the more significant the damage evolution process of the seal; and the nonlinear characteristics of the damage evolution process itself, namely, slow early damage development and accelerated later damage development, and the comprehensive sealing index. As the result of nonlinear transformation, nonlinear correlation can more accurately reflect the correspondence between the sealing state and the actual damage amount, avoiding the damage quantification distortion caused by linear correlation, and providing an accurate damage quantification basis for subsequent remaining service life prediction.

[0070] In one implementation, the nonlinear damage accumulation model includes a power-law damage accumulation model and an exponential damage accumulation model;

[0071] Define measured pressure For the normal operating conditions of the fracturing pump, a power-law damage accumulation model is used for calculation. :

[0072]

[0073] in, This represents the damage rate constant under normal operating conditions. The damage initiation threshold, , , To preset the nonlinear exponent, This is the current measured pressure. For reference pressure, To activate energy, Let be the ideal gas constant. This is the current measured temperature;

[0074] Define measured pressure For fracturing pumps operating under high pressure, an exponential damage accumulation model is used for calculation. :

[0075]

[0076] in, The damage rate constant under high pressure conditions. The high-voltage damage amplification index, This is the preset critical damage value;

[0077] The PHM module automatically identifies the current operating condition type, matches the corresponding formula to calculate the RUL, and generates maintenance suggestions based on the current RUL and preset maintenance rules.

[0078] Specifically, this embodiment uses measured pressure. With reference pressure The ratio is the dividing criterion. The rated operating pressure of the fracturing pump is determined by the equipment's factory parameters.

[0079] In the power-law damage accumulation model The damage rate constant under normal operating conditions is used to reflect the damage development rate of the seal under normal operating conditions. It is obtained by fitting the accelerated life test of the seal under normal operating conditions. The damage initiation threshold is the value at which the comprehensive sealing index reaches, at which point the seal begins to produce quantifiable damage. This value is obtained from statistical analysis of seal fatigue tests. , , The preset nonlinear indices, which characterize the nonlinear influence of the comprehensive sealing index and pressure on damage evolution, are obtained by fitting experimental data. The fatigue activation energy of the sealing material is obtained from material handbooks and sealing material tests. is the ideal gas constant, and is a universal physical constant (with a value of 8.314 J / (mol・K)). This is a preset critical damage value, meaning the seal is considered to have failed when it reaches this level of damage. It is determined by seal failure testing and industry standards, and its value is typically close to 1. In the formula, The nonlinear contribution of the comprehensive sealing index to damage evolution is only when > When this value is positive, the seal is damaged. This is the normalized nonlinear contribution term for pressure, used to reflect the effect of pressure on damage under normal operating conditions; This is used to reflect the effect of temperature on the fatigue damage of sealing materials; the higher the temperature, the faster the damage rate. This is the change in damage from the current value to the critical value, used to reflect the incremental damage required for the seal to go from its current state to failure; The damage evolution rate term under normal operating conditions reflects the damage increment per unit time. This power-law damage accumulation model is based on power-law damage evolution theory and is applicable to operating conditions where damage development is slow and the damage rate and damage amount have a power-law relationship. Under normal operating conditions, the fracturing pump pressure is low, and the wear and leakage of the seals develop slowly, with damage evolution conforming to power-law characteristics. This model can accurately calculate the remaining service life of the seals under normal operating conditions. ;

[0080] In the exponential damage accumulation model, The damage rate constant under high-pressure conditions reflects the damage development rate of the seal under high-pressure conditions. It is obtained by fitting the seal through accelerated life testing under high-pressure conditions. > This means that the seals are damaged faster under high pressure. The high-pressure damage amplification index characterizes the degree of pressure amplification of damage under high-pressure conditions and is obtained by fitting high-pressure fault simulation experiments; the meanings and acquisition methods of the other parameters are consistent with the power-law model; in the formula, This is the pressure damage rate term under high pressure conditions, used to reflect the significant increase in damage rate caused by high pressure. This is a nonlinear term representing the change in damage from the current value to the critical value, used to reflect the characteristics of accelerated damage evolution under high pressure. This is the inverse correction term of the comprehensive sealing index, i.e. The larger the size, the worse the sealing condition and the shorter the remaining service life;

[0081] The exponential damage accumulation model, based on the exponential damage evolution theory, is applicable to operating conditions where damage develops rapidly and the damage rate increases exponentially with operating parameters. Under high-pressure operating conditions, the pressure of the fracturing pump increases significantly, the contact stress on the sealing surface and the fluid erosion effect are significantly enhanced, and the wear and leakage of the seals accelerate exponentially. The damage evolution conforms to exponential characteristics. This model can accurately calculate the remaining service life of the seals under high-pressure conditions. ;

[0082] It should be noted that the damage evolution characteristics of seals differ between conventional and high-pressure operating conditions of fracturing pumps. Under conventional conditions, the pressure is low, and seal damage develops slowly. The damage rate exhibits a power-law relationship with the damage amount and operating parameters. Using an exponential model would lead to an overestimation of the lifespan prediction result and an excessive underestimation of the damage rate. Under high-pressure conditions, the pressure is high, and seal damage develops exponentially. The damage rate is highly sensitive to changes in operating parameters. Using a power-law model would lead to an underestimation of the lifespan prediction result and an excessive overestimation of the damage rate. Therefore, the dual-model design in this embodiment matches the lifespan prediction model with the actual damage evolution characteristics of the seals under different operating conditions, significantly improving the accuracy of remaining service life prediction. Accurate lifespan prediction is the core foundation for achieving seal failure prediction and health management. A single model cannot take into account the damage characteristics of both operating conditions, leading to distorted prediction results and failing to effectively guide on-site maintenance.

[0083] In one implementation, the output module includes:

[0084] The local display unit, located in the fracturing pump control room, is used to display the current sealing status, remaining service life (RUL), and maintenance recommendations in real time.

[0085] The remote alarm unit is used to send alarm signals to the remote monitoring center when an early or serious leak is detected.

[0086] In one implementation, it further includes:

[0087] The communication module, connected to the data processing module, is used to upload operating status parameters and characteristic data to the cloud server for storage and backup.

[0088] The analog-to-digital converter module, connected to the data acquisition module, is used to convert the acquired analog signals into digital signals;

[0089] The storage module is connected to the data processing module, the status diagnosis module, and the PHM module respectively, and is used to store operating status parameters, characteristic data, sealing status identification results, and historical data of remaining service life (RUL).

[0090] In one implementation, it further includes:

[0091] The self-calibration module, connected to the data acquisition module, is used to automatically calibrate the zero-point drift and sensitivity of pressure sensors, flow sensors, temperature sensors, and vibration acceleration sensors periodically or under specific operating conditions.

[0092] It should also be noted that the preset maintenance rules are the core basis for the PHM module to generate maintenance recommendations, and are formulated based on the remaining service life (RUL) of the seals, the seal condition type, and the operating conditions of the fracturing pump; in some implementations, the specific rules are as follows:

[0093] When the sealing condition is normal, if the remaining service life (RUL) is greater than the preset safe service life threshold (determined by the seal design life and on-site operation requirements), a maintenance suggestion of "normal operation, regular inspection" is generated; if the remaining service life (RUL) is less than or equal to the safe service life threshold, a maintenance suggestion of "normal condition but lifespan is approaching, it is recommended to prepare spare parts in advance and plan to shut down for maintenance" is generated.

[0094] When the seal is in an early leakage state, a graded maintenance recommendation is generated based on the remaining service life (RUL). If the RUL is greater than 72 hours, the recommendation is "early leakage, it is recommended to check and replace the seal at the end of this work cycle"; if the RUL is less than or equal to 72 hours, the recommendation is "early leakage, it is recommended to stop the machine within 24 hours to check the seal to prevent the failure from escalating".

[0095] When the seal is in a state of severe leakage, regardless of the remaining service life (RUL), an emergency maintenance recommendation is generated: "Severe leakage, immediately stop the machine and replace the seal to avoid equipment damage and safety accidents."

[0096] All maintenance recommendations can be fine-tuned based on the actual operating conditions of the fracturing pump (such as the current critical stage of high-pressure fracturing). If it is in a critical stage of operation and there is no backup equipment, a recommendation of "temporary sealing maintenance, replacement immediately after operation" can be generated in the early leakage state, taking into account both the continuity of operation and the safety of equipment.

[0097] Therefore, the preset maintenance rules can be flexibly adjusted according to on-site operation requirements, seal model, and equipment management standards to achieve health management of seals. This embodiment does not limit the specific rules.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A control system for online identification of sealing condition of a hydraulic end of a fracturing pump, characterized in that, include: The data acquisition module is used to collect the operating status parameters of the hydraulic end of the fracturing pump in real time; A data processing module, connected to the data acquisition module, is used to preprocess and extract features from the acquired operating status parameters to generate feature data that can characterize the sealing state. The data processing module includes: The signal conditioning unit is used to filter and amplify pressure signals, flow signals, temperature signals, and vibration signals; The feature extraction unit, connected to the signal conditioning unit, is used to extract frequency domain features and time domain features from the conditioned signal. The frequency domain feature comprises a spectrum entropy obtained based on spectrum analysis of the vibration signal , ; wherein, is the energy of the th frequency component, is the proportion of the energy of the th frequency component in the total energy of the signal, and N is the total number of frequency components; The time-frequency domain feature includes a spectrum density reconstructed after wavelet packet decomposition based on the pressure signal , ; wherein, is a wavelet packet reconstructed signal of the th frequency band, and t is time; The status diagnosis module, connected to the data processing module, has a built-in sealing status recognition model, which is used to identify the current sealing status of the hydraulic end online based on the feature data. The sealing status includes normal status, early leakage status and severe leakage status. The sealing state identification model is used for calculating a comprehensive sealing index , and the comprehensive sealing index is used for quantitatively evaluating the sealing state; and a processing procedure of the sealing state identification model is as follows: constructing a feature vector ; wherein is the pressure difference between the suction chamber and the discharge chamber under the current working condition, is the corresponding standard pressure difference, is the reference frequency spectrum entropy under the normal sealing state; is the reference energy spectrum density of the corresponding frequency band under the normal sealing state; is the current temperature, is the reference temperature; The calculated feature vector The Mahalanobis distance relative to the normal state feature vector mean , ; wherein, And the covariance matrix Is obtained by statistical analysis of historical normal operation data;​ Mapping the Mahalanobis distance to a comprehensive sealing index by a non-linear transformation , ; wherein, and are preset sensitivity adjustment constants; The sealing condition identification model further includes: comparing the comprehensive sealing index. The relationship between the sealing state and the first preset threshold and the second preset threshold is used to identify the sealing state online. like If the first preset threshold is reached, it is identified as a normal state; if the first preset threshold is reached... The second preset threshold is then identified as an early leakage state; like The second preset threshold is then identified as a severe leakage state; The PHM module, connected to the condition diagnostic module, is used to predict the remaining service life (RUL) of the seal based on the identified seal condition and its changing trend, and to generate maintenance recommendations. The output module is connected to the status diagnostic module and the PHM module respectively, and is used to output the sealing status, remaining service life (RUL), and maintenance recommendations.

2. The control system for online identification of the hydraulic end sealing status of a fracturing pump according to claim 1, characterized in that, The data acquisition module includes: Multiple pressure sensors are installed in the suction chamber and discharge chamber of the hydraulic end, respectively, to collect fluid pressure signals inside the pump; At least one flow sensor is installed on the discharge line at the hydraulic end to collect the pump discharge flow signal; At least one temperature sensor is installed on the hydraulic end housing or in the internal flow channel to collect the temperature signal of the hydraulic end; At least one vibration acceleration sensor is installed on the hydraulic end valve box or pump head body to collect vibration signals of the pump head body.

3. The control system for online identification of the hydraulic end sealing status of a fracturing pump according to claim 1, characterized in that, The PHM module is used to determine the comprehensive sealing index. Based on the current operating parameters, two different nonlinear damage accumulation models are used to calculate the remaining service life (RUL), which are adapted to the conventional operating conditions and high-pressure operating conditions of fracturing pumps, respectively. Before calculating the remaining useful life (RUL), the following are also included: Define the current cumulative damage amount The relationship with the comprehensive sealing index is as follows: ;in, This is the preset damage coefficient.

4. The control system for online identification of the hydraulic end sealing status of a fracturing pump according to claim 3, characterized in that, The nonlinear damage accumulation model includes the power-law damage accumulation model and the exponential damage accumulation model; Define measured pressure For the normal operating conditions of the fracturing pump, the power-law damage accumulation model is used to calculate... ; ;in, This represents the damage rate constant under normal operating conditions. The damage initiation threshold, To preset the nonlinear exponent, This is the current measured pressure. For reference pressure, To activate energy, Let be the ideal gas constant. This is the current measured temperature; Define measured pressure For the high-pressure operation of the fracturing pump, the exponential damage accumulation model is used to calculate... : ;in, The damage rate constant under high pressure conditions. The high-voltage damage amplification index, This is the preset critical damage value; The PHM module automatically identifies the current operating condition type, matches the corresponding formula to calculate RUL, and generates maintenance suggestions based on the current RUL and preset maintenance rules.

5. A control system for online identification of the hydraulic end sealing status of a fracturing pump according to claim 1, characterized in that, The output module includes: The local display unit, located in the fracturing pump control room, is used to display the current sealing status, remaining service life (RUL), and maintenance recommendations in real time. The remote alarm unit is used to send alarm signals to the remote monitoring center when an early or serious leak is detected.

6. The control system for online identification of the hydraulic end sealing status of a fracturing pump according to claim 1, characterized in that, Also includes: A communication module, connected to the data processing module, is used to upload the operating status parameters and feature data to a cloud server for storage and backup. An analog-to-digital converter module, connected to the data acquisition module, is used to convert the acquired analog signals into digital signals; The storage module is connected to the data processing module, the status diagnosis module and the PHM module respectively, and is used to store the operating status parameters, feature data, sealing status identification results and historical data of remaining service life (RUL).

7. A control system for online identification of the hydraulic end sealing status of a fracturing pump according to claim 2, characterized in that, Also includes: The self-calibration module, connected to the data acquisition module, is used to periodically or under specific operating conditions to automatically calibrate the zero-point drift and sensitivity of the pressure sensor, flow sensor, temperature sensor, and vibration acceleration sensor.

Images