Intelligent testing system and method for wear of filling slurry conveying pipeline based on big data

Abstract

The present invention belongs to the technical field of intelligent testing systems and discloses a big data-based intelligent testing system and method for wear in filling slurry conveying pipelines. The system comprises a pipeline testing module, a database module, and a data analysis module. The database module includes a multidimensional pipeline database and a preprocessing module. The database module receives and stores pipeline signals transmitted by the pipeline testing module and constructs a multidimensional pipeline database. The preprocessing module receives pipeline signals and performs data preprocessing to form characteristic signals. The present invention's big data-based intelligent testing system and method for wear in filling slurry conveying pipelines can address the issues of untimely wear monitoring of filling slurry conveying pipelines and wear assessment accuracy generally falling short of actual requirements.

Description

Technical Field

[0001] The present invention relates to an intelligent testing system, and in particular to an intelligent testing system and method for wear of a filling slurry conveying pipeline based on big data. Background Art

[0002] Wear monitoring and control of filling slurry conveying pipelines are crucial in mine filling operations, but existing technologies have significant flaws. Traditional methods rely on regular manual inspections and single sensors to collect basic parameters such as local pressure and flow rate, which makes it difficult to fully reflect the true state of pipeline wear. Due to the lack of comprehensive analysis of multi-dimensional data such as slurry concentration, particle grading, and pipe characteristics, the accuracy of wear assessment is generally lower than actual needs, leading to deviations in maintenance decisions. Existing wear prediction models are mostly based on static empirical formulas and cannot dynamically respond to changes in complex working conditions. Measured data show that prediction errors often exceed 30%, especially when the rheological properties of the slurry suddenly change, increasing the risk of failure.

[0003] Furthermore, manual control suffers from a significant response lag, with abnormal operating conditions often delayed for over two hours. At one iron mine, this resulted in an unplanned downtime rate of up to 17%. Data silos further hampered system performance. The lack of effective integration of heterogeneous data sources, including historical operating data, equipment parameter libraries, and geological data, led to a data utilization rate of less than 40% at one gold mine, severely limiting the scope for intelligent upgrades.

[0004] The existing technology (such as CN202410395694.7) has constructed a wear detection model based on historical data, but the model has poor analytical ability when processing signals with different temporal and spatial characteristics, and the wear detection model based only on historical wear data cannot adapt to complex and changeable pipeline wear scenarios; the existing technology (CN202411176764.6) judges the wear of the pipe wall by processing the acoustic wave signal and the acoustic wave signal reflected back by the pipe wall, but cannot predict the wear state of the pipe wall, and cannot detect other wear states of the pipeline. Summary of the Invention

[0005] In view of the shortcomings of the above-mentioned prior art, the purpose of the present invention is to provide an intelligent testing system and method for the wear of filling slurry conveying pipelines based on big data, which is used to solve the problems of untimely wear monitoring of filling slurry conveying pipelines and generally lower wear assessment accuracy than actual needs. The present invention systematically overcomes the limitations of traditional technologies by integrating multi-dimensional data acquisition, dynamic simulation analysis and intelligent closed-loop control. Specifically, this solution constructs an intelligent monitoring system based on a multi-physical field sensor network and an adaptive learning framework. Its technical implementation path includes four core modules: a high-density sensor array to achieve holographic perception of pipeline status, a distributed data processing architecture to complete spatiotemporal alignment and feature extraction of multimodal data, a hybrid modeling method that integrates the dual advantages of physical mechanism and data-driven, and a closed-loop optimization mechanism to achieve the co-evolution of parameter control and model iteration.

[0006] The present invention provides a big data-based intelligent testing system for wear of filling slurry conveying pipelines, comprising:

[0007] A pipeline testing module forms a pipeline signal;

[0008] The database module includes a multi-dimensional pipeline database and a preprocessing module. The database module receives and stores pipeline signals transmitted by the pipeline testing module and builds a multi-dimensional pipeline database by integrating historical pipeline signal data sets and external pipeline operation data obtained through the network. The preprocessing module receives pipeline signals and performs data preprocessing to form characteristic signals.

[0009] Data analysis module: The data analysis module builds a pipeline wear simulation model based on a multi-dimensional pipeline database. Characteristic signals are input into the pipeline wear simulation model to perform dynamic wear analysis and generate simulation prediction signals and control parameter signals. The control parameter signals are transmitted to the pipeline testing module to adjust the slurry delivery parameters. The pipeline testing module generates a slurry control signal.

[0010] Feedback optimization module, the feedback optimization module receives and compares the slurry control signal and the simulation prediction signal, extracts the difference characteristic value analysis between the two to generate a feedback optimization signal, and inputs it to the data analysis module and the database module; when the feedback optimization module compares the slurry control signal and the simulation prediction signal, after the control signal acts on the actual system, the data acquisition unit continuously monitors the key parameters of the pipe wall to form a real-time feedback signal. After the system aligns the feedback signal with the simulation prediction signal in time and space, it applies DMD technology to extract the dominant time and space modes of the two, and evaluates the control effect by calculating the mode amplitude correlation coefficient. When the correlation coefficient is lower than the preset threshold, the difference feature deep analysis process is started. When performing the difference feature deep analysis process, the phase space of the slurry control signal with differences is reconstructed, and the maximum Lyapunov exponent is calculated to judge the chaotic characteristics of the system. The key modal components are extracted based on singular value decomposition to form difference eigenvalues. When performing difference eigenvalue analysis, a causal reasoning framework is introduced to construct a Bayesian network with slurry parameters as dependent variables and wear characteristics as result variables. The conditional probability distribution is estimated by the Markov chain Monte Carlo method, and the control parameters whose contribution to wear differences exceeds the preset value are identified. A feedback optimization signal containing a parameter adjustment priority list is generated, and the parameter weights of the pipeline wear simulation model are updated through an incremental learning mechanism.

[0011] In one embodiment of the present invention, the pipeline testing module includes a test pipeline unit, a slurry circulation unit and a data acquisition unit. The slurry circulation unit controls the slurry to be delivered to the test pipeline unit. The data acquisition unit is arranged in the test pipeline unit. The data acquisition unit collects various data when the slurry flows through the test pipeline unit and forms a pipeline signal.

[0012] In one embodiment of the present invention, when a database module integrates historical pipeline signal data sets with external pipeline operation data, it includes a data fusion stage, a data cleaning stage, and a data repair stage. In the data fusion stage, a distributed stream processing engine is used to access and align the formats of multi-source heterogeneous pipeline signals in real time. The pipeline signals include but are not limited to pressure waveforms, vibration modal spectra, and pipe wall thickness measurement data under different geological conditions, pipe diameter specifications, and conveying media. In the data cleaning stage, an adaptive threshold filtering algorithm is used to eliminate pulse noise interference, and trend terms and random fluctuation components in the pipeline signals are separated based on time series decomposition technology. In the data repair stage, a spatiotemporal correlation modeling method is used to construct a tensor filling model with pipe segment location as the spatial dimension and operation cycle as the time dimension, and the alternating least squares method is used to optimize the objective function to complete data repair. The processed standardized data is spatially indexed according to the pipe segment topology structure, and a time series database is used to compress and store high-frequency sampling data to form a database module.

[0013] In one embodiment of the present invention, when the preprocessing module receives the pipeline signal and performs data preprocessing to form a characteristic signal, multi-scale feature extraction is first performed on the pipeline signal, and the time domain signal of the pipeline signal is decomposed into different frequency band subspaces using wavelet packet transform, and the energy entropy of each subband is calculated as the frequency domain feature quantity. At the same time, the intrinsic mode function component of the signal is obtained through empirical mode decomposition, and its instantaneous amplitude-frequency characteristics are extracted to form a time-frequency joint feature vector; after the feature vector is normalized, the principal component analysis algorithm is used to reduce the feature dimension, and the feature components whose cumulative contribution rate exceeds the preset threshold are retained to form a feature signal. The generated feature signal is pushed to the data analysis module in real time through the message queue.

[0014] In one embodiment of the present invention, the construction of the pipeline wear simulation model adopts an integrated learning method to improve the prediction accuracy by stacking multiple base models. The pipeline wear simulation model includes a first-layer base model, a second-layer base model and a third-layer base model. The first-layer base model contains a fluid dynamics equation solver based on physical mechanisms to calculate the slurry flow field distribution and particle collision energy; the second-layer base model uses a gradient boosting decision tree to capture the nonlinear relationship in historical data; the third-layer base model is a three-dimensional convolutional neural network to process the microscopic wear morphology features in the pipeline cross-section detection image, and finally dynamically adjusts the weight coefficients of each base model through a meta-learning framework to form a pipeline wear simulation model.

[0015] In one embodiment of the present invention, when performing dynamic wear analysis, a sliding time window mechanism is used to divide the continuous pipeline signal into analysis segments of fixed length according to the time-varying characteristics of the slurry conveying parameters. After the characteristic signal in each time window is input into the simulation model, the output contains the predicted value of the pipe wall thickness loss rate for the next three operating cycles, and forms a simulation prediction signal.

[0016] In one embodiment of the present invention, a multi-objective optimization algorithm is used to generate a slurry control signal. The defined objective function variables include but are not limited to three optimization dimensions: minimizing wear rate, maximizing conveying efficiency, and minimizing energy consumption cost. The optimal solution set is found through the Pareto frontier search algorithm, and the parameter combination with the highest comprehensive score is selected based on the fuzzy decision theory. The fuzzy membership function dynamically adjusts the shape parameters according to the real-time working conditions and forms a slurry control signal.

[0017] The present invention also includes a big data-based intelligent testing method for wear of a filling slurry conveying pipeline, comprising:

[0018] S100: real-time collection of pipeline signals of slurry flow in the delivery pipeline;

[0019] S200: Receive and store pipeline signals, build a multi-dimensional pipeline database by integrating historical pipeline signal data sets and external pipeline operation data obtained through the network, and perform data preprocessing on the pipeline signals to generate characteristic signals;

[0020] S300: Constructing a pipeline wear simulation model based on a multi-dimensional pipeline database, inputting characteristic signals into the simulation model for dynamic wear analysis, and generating a simulation prediction signal including predicted wear parameters and a slurry control signal representing optimal operating parameters;

[0021] S400: transmitting a slurry control signal to an execution unit of a slurry delivery system to adjust slurry parameters;

[0022] S500: Acquire the adjusted pipeline operation feedback signal in real time, compare the feedback signal with the simulation prediction signal, extract the difference feature value between the two, and generate an optimization instruction including a parameter correction strategy;

[0023] S600: Synchronously feed back the optimization instructions to the pipeline wear simulation model and the multi-dimensional pipeline database, driving the simulation model to iteratively update parameters.

[0024] The present invention provides a big data-based intelligent testing system and method for wear of slurry conveying pipelines. This system constructs an intelligent monitoring system based on a multi-physics field sensor network and an adaptive learning framework. Its technical implementation path includes four core modules: a high-density sensor array enables holographic perception of pipeline status; a distributed data processing architecture completes spatiotemporal alignment and feature extraction of multimodal data; a hybrid modeling approach integrates the dual advantages of physical mechanisms and data-driven approaches; and a closed-loop optimization mechanism achieves the co-evolution of parameter control and model iteration. The system captures multi-dimensional physical signals such as pipe wall vibration acceleration, radial strain, and surface temperature through electromagnetic-acoustic composite sensors deployed at key pipeline nodes. Using a high-speed industrial bus to achieve microsecond-level timing synchronization, the system constructs a dynamic monitoring network covering the entire length of the pipeline. The data processing layer uses a stream-batch integrated computing framework to perform preprocessing operations on the original signal, including wavelet denoising, outlier removal, and feature vector construction. It then uses spatiotemporal coding technology to map discrete sampling points into a continuous three-dimensional field distribution model. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0026] Figure 1 This is the system architecture diagram of the intelligent testing system for wear of filling slurry conveying pipelines based on big data;

[0027] Figure 2 This is a flowchart of the steps of the intelligent testing method for wear of slurry conveying pipelines based on big data. DETAILED DESCRIPTION

[0028] The following describes the embodiments of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. The details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other unless they conflict.

[0029] It should be noted that the illustrations provided in the following embodiments are merely schematic illustrations of the basic concept of the present invention. Therefore, the illustrations only show components related to the present invention and are not drawn according to the number, shape, and size of components in actual implementation. In actual implementation, the type, quantity, and proportion of each component may be changed arbitrarily, and the component layout may also be more complex.

[0030] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it will be apparent to those skilled in the art that the embodiments of the present invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring the embodiments of the present invention.

[0031] The present invention relates to an intelligent testing system and method for wear of filling slurry conveying pipelines based on big data, which is mainly used in pipeline transportation work scenarios. As one of the core methods of modern industrial material transportation, pipeline transportation plays a key role in the fields of mine filling, metallurgy, petrochemicals, etc. When the filling slurry flows at high speed inside the pipeline, the continuous friction between the solid particles and the pipe wall causes progressive wear. This wear process has complex characteristics of nonlinearity and multi-factor coupling, and traditional detection methods are difficult to achieve accurate monitoring and trend prediction. The periodic manual detection method currently commonly used in the industry has significant lag, and problems are usually not discovered until the thinning of the pipe wall exceeds the safety threshold, resulting in frequent unplanned shutdowns and rising maintenance costs. Although online monitoring equipment based on technologies such as vibration sensing and ultrasonic thickness measurement has gradually become popular in recent years, its problems such as single data analysis dimension and insufficient model generalization ability have not been fundamentally solved. In particular, when dealing with wear prediction under multiple working conditions and the combined effects of multiple materials, existing systems generally exhibit defects such as poor adaptability and high false alarm rate. The main limitations of existing technology systems are reflected in three aspects: the insufficient spatiotemporal resolution of data acquisition, with most systems adopting a sparse point distribution strategy, making it difficult to capture the wear gradient distribution characteristics across the entire pipeline area; the lack of deep integration of multi-source heterogeneous data in data processing, with key parameters such as vibration signals, pressure waveforms, and material properties often stored in isolated forms, failing to construct correlation models reflecting wear mechanisms; and the reliance on static empirical models at the decision support level, which are unable to adapt to the complexities of real-world operating conditions such as dynamic changes in slurry composition and degradation of pipe performance. A typical case study shows that the conventional monitoring system used by a copper mine had indicated that the wear rate was within a safe range for three consecutive months. However, actual testing revealed that the localized wear at the bend had exceeded the critical value. The root cause was that the system algorithm failed to recognize the turbulent impact characteristics implicit in the high-frequency vibration signal, resulting in the critical warning information being mistakenly eliminated by the noise filtering mechanism.

[0032] See Figure 1-Figure 2, showing the intelligent testing system and method for filler slurry pipeline wear based on big data of the present invention. The intelligent testing system for filler slurry pipeline wear based on big data of the present invention includes a pipeline testing module, a database module, and a data analysis module. The pipeline testing module generates pipeline signals. The database module includes a multidimensional pipeline database and a preprocessing module. The database module receives and stores the pipeline signals transmitted by the pipeline testing module, constructing the multidimensional pipeline database by integrating historical pipeline signal datasets with external pipeline operation data acquired through the network. The preprocessing module receives the pipeline signals and performs data preprocessing to generate characteristic signals. The data analysis module constructs a pipeline wear simulation model based on the multidimensional pipeline database. The characteristic signals are input into the pipeline wear simulation model for dynamic wear analysis, generating simulation prediction signals and control parameter signals. The control parameter signals are transmitted to the slurry circulation unit to adjust slurry delivery parameters, and the pipeline testing module generates slurry control signals. The feedback optimization module receives and compares the slurry control signal and the simulation prediction signal, extracts the difference between the two, analyzes the characteristic values, generates feedback optimization signals, and inputs them to the data analysis module and the database module.

[0033] like Figure 1 As shown in the figure, in a big data-based intelligent wear testing system for slurry pipelines, the pipeline testing module serves as the front-end perception layer of the entire system, undertaking the critical task of acquiring real-time raw data on pipeline operating conditions. This module consists of three components: a test pipeline unit, a slurry circulation unit, and a data acquisition unit. Through a highly coordinated hardware layout and intelligent control logic, it accurately captures and dynamically controls pipeline wear characteristics under complex operating conditions. The test pipeline unit is the core vehicle for physical experiments. Its design must balance realistic operating condition simulation with parameter adjustability. It typically adopts a modular structure, enabling rapid changes in pipe material, diameter, and connection type to meet diverse experimental requirements. For example, for highly abrasive slurries, the test pipeline may utilize wear-resistant sections lined with ceramic composites, while for high-pressure transmission conditions, high-strength alloy steel may be selected. Contoured structures (such as corrugations and protrusions) can be pre-installed within the pipeline to simulate localized wear hotspots commonly found in real-world applications. Stress monitoring devices at flange connections simultaneously monitor the impact of mechanical vibration on pipe wall integrity in real time. The installation angle, bending radius and other parameters of the test pipe can be dynamically adjusted through the electric adjustment mechanism to reproduce the impact of spatial direction changes commonly seen in mine filling systems on the slurry flow state.

[0034] Specifically, the slurry circulation unit, as a power and medium supply system, has the core function of precisely controlling the physical properties and flow parameters of the slurry. This unit typically includes key components such as a high-pressure plunger pump, a slurry mixing chamber, a concentration control device, and a temperature control system. The high-pressure plunger pump utilizes variable frequency drive technology, enabling stepless adjustment of the slurry flow rate within a range of 0.5-8 m / s. A digital flow meter is used to achieve closed-loop control of the delivery rate. The slurry mixing chamber is equipped with a dual-axis agitator and an ultrasonic disperser to ensure uniform suspension of solid particles (such as tailings and aggregates) in the liquid medium. The concentration control device monitors the slurry solids content in real time using an online density meter and dynamically balances the concentration using a dry powder feeder and dilution water valve, achieving a regulation accuracy of ±1.5%. The temperature control system maintains the slurry within a set temperature range (typically 5-60°C) using a coil-and-tube heat exchanger to simulate the effects of different seasons or underground environments on the slurry's rheological properties. Specifically, the unit integrates a particle size analyzer and rheometer, enabling real-time acquisition of slurry particle size distribution and apparent viscosity data, providing multi-dimensional input parameters for subsequent wear mechanism analysis. The data acquisition unit, serving as an information perception network, comprehensively monitors the pipeline's operating status through a distributed sensor array. Sensor nodes are arranged according to a spatial topology at key locations along the test pipeline, including vulnerable areas such as straight sections, elbows, and reducers. The pressure sensor utilizes a piezoelectric principle, sampling at a 1000Hz frequency to capture dynamic pressure fluctuations on the pipe wall, with a measurement range of 0-25MPa and a resolution of 0.01MPa. Vibration monitoring utilizes a combination of a triaxial accelerometer and a fiber Bragg grating (FBG) sensor. The former captures the mechanical vibration spectrum at a high sampling frequency of 10kHz, while the latter analyzes the pipe strain distribution through wavelength offset, achieving a spatial resolution of up to 5mm. Pipe wall thickness measurement utilizes a hybrid technology of pulsed eddy current and electromagnetic ultrasonic testing, achieving online, non-contact thickness measurement with an accuracy of 0.1mm and unaffected by surface coatings or deposits. The environmental parameter sensor group continuously records changes in ambient temperature, humidity, and pipe surface temperature gradients, providing auxiliary data for multi-physics field coupling analysis. All sensor signals are transmitted to the data acquisition card via anti-interference shielded cables. After 24-bit A / D conversion and digital filtering, they are processed into standardized data packets with a time synchronization error of less than 1ms, ensuring the spatiotemporal alignment of multimodal data.

[0035] In one embodiment of the present invention, the three subunits achieve deep collaboration via industrial Ethernet and fieldbus. When the test process is initiated, the slurry circulation unit first prepares slurry according to preset parameters (e.g., 65% concentration, 4 m / s flow rate) and establishes stable flow. The test pipeline unit simultaneously adjusts to the target inclination angle (e.g., 30° elevation) and bend radius (e.g., 3D elbow) to simulate the spatial characteristics of an actual transmission line. The data acquisition unit then initiates full-channel sampling, uploading raw signals such as pressure pulsation, vibration spectrum, and pipe wall thickness changes to a central controller in real time. During this process, an intelligent control algorithm continuously analyzes data features and dynamically adjusts slurry parameters to cover a wider range of operating conditions. For example, if abnormal vibration energy accumulation is detected in a particular bend, the system automatically triggers a gradient pressurization program, gradually increasing the pumping pressure until a shift in the pipe wall resonant frequency is observed, thereby determining the critical structural strength point of that pipe section. This active testing mode overcomes the limitations of traditional passive monitoring and can stimulate potential failure modes under controllable conditions, providing a rich data sample for wear mechanism research.

[0036] like Figure 1 As shown in Figure 2, the construction and data processing flow of the database module are the core foundation for intelligent analysis. This module integrates real-time signals from the pipeline testing module, historically accumulated pipeline operation datasets, and external pipeline operation data acquired through the network to build a pipeline database covering multiple dimensions and spatiotemporal scales. The data integration process begins with data fusion, utilizing a distributed stream processing engine to connect heterogeneous data streams from multiple sources. These data sources include pressure waveforms, vibration modal spectra, pipe diameter specifications, conveying medium properties (such as slurry concentration and particle size distribution) under different geological conditions, and periodic measurements of pipe wall thickness. Due to differences in acquisition frequency, data format, and communication protocol among different sensors, the system uses a customized format alignment algorithm to uniformly encode metadata such as timestamps, spatial coordinates, and data types, and establishes a spatiotemporal correlation index between the data. For example, millisecond-level pulse signals from the pressure sensor and hourly-level data from the pipe wall thickness meter are time-aligned using an interpolation algorithm, while vibration spectrum data from different pipe sections are spatially mapped using the pipeline topology.

[0037] Specifically, the data cleaning phase employs an adaptive threshold filtering algorithm to eliminate impulse noise interference. This algorithm dynamically calculates statistical features (such as mean and variance) within a signal window and, in conjunction with a sliding window mechanism, adjusts the filtering threshold in real time. For sudden spike noise, the system employs morphological filtering to identify and remove anomalous pulses while preserving the true pressure fluctuation characteristics. To address baseline drift, a common problem in vibration signals, time series decomposition techniques are used to separate the original signal into trend, periodic, and random fluctuation components. The trend term reflects stiffness degradation caused by pipe fatigue, the periodic term corresponds to the natural frequency of pump operation or slurry pulses, and the random term is used to capture anomalous events. The data repair phase focuses on addressing missing data caused by sensor failure or communication interruptions. By constructing a spatiotemporal correlation tensor infill model, the pipeline network is treated as a three-dimensional structure (pipeline segment location, operating time, and measurement parameters). The system uses alternating least squares to optimize the objective function and iteratively estimate missing values. For example, if a pipeline segment loses 12 hours of vibration data due to an optical fiber break, the system reconstructs the data distribution for that period by comparing the vibration modes of adjacent segments and the temporal evolution of historical data under the same operating conditions. After processing, the standardized data is indexed using an R-tree based on the spatial topology of the pipe segment. High-frequency sampling data is compressed and stored in a time-series database. Difference encoding and run-length encoding are used to reduce storage space to less than 30% of the original data. The preprocessing module extracts multi-scale features from the raw pipeline signal and decomposes the time-domain signal into frequency subspaces using a wavelet packet transform. Taking the pressure pulsation signal as an example, a five-level decomposition using the db4 wavelet basis is performed, resulting in 32 frequency band components. The energy entropy of each subband is calculated to characterize the complexity of the frequency-domain energy distribution. For non-stationary vibration signals, empirical mode decomposition is used to adaptively decompose them into multiple intrinsic mode functions (IMFs). The instantaneous amplitude envelope and Hilbert spectrum of each IMF component are extracted to form a joint time-frequency feature vector. For example, after EMD decomposition of the vibration signal at a certain bend, the amplitude abrupt change of the third-order IMF component in a specific frequency band (e.g., 800-1200 Hz) was identified as a typical characteristic of particle swarm collisions. After Z-score normalization, the feature vectors are then subjected to dimensionality reduction using principal component analysis. Principal components with cumulative contributions exceeding 95% are retained, compressing the original 128-dimensional features to 18 dimensions, significantly improving the computational efficiency of subsequent models. The reduced feature signals are pushed to the data analysis module in real time via a Kafka message queue, ensuring a throughput of over 100,000 feature data streams per second.

[0038] Furthermore, the pipeline wear simulation model was constructed using an ensemble learning approach, achieving multi-level feature fusion by stacking physical mechanism models, machine learning models, and deep learning models. The first-layer base model is a physical simulator based on computational fluid dynamics. It solves the Navier-Stokes equations to simulate the slurry flow field distribution and combines it with the discrete element method to calculate the collision energy distribution of solid particles. This layer outputs key physical quantities including pipe wall shear stress, particle impact frequency, and kinetic energy transfer efficiency. The second-layer base model uses an XGBoost gradient boosting decision tree, which inputs operating parameters (such as flow rate and concentration) and wear labels from historical data to capture nonlinear interactions between parameters. For example, when the slurry concentration increases from 60% to 70%, the decision tree model can identify that the flow rate needs to be reduced by 0.8 m / s to maintain an equivalent wear rate. The third-layer base model is a three-dimensional convolutional neural network, which processes microscopic wear images captured by pipeline endoscopes and extracts surface morphological features such as pits and scratches using multi-scale convolution kernels. Ultimately, a meta-learning framework dynamically adjusts the weight coefficients of each base model. For example, the CFD model is given a higher weight under stable flow conditions, while the CNN model's decision contribution is enhanced under complex turbulent conditions, forming an adaptive hybrid simulation model. Dynamic wear analysis utilizes a sliding time window mechanism to process the continuous data stream. The window length is dynamically adjusted based on the frequency of changes in slurry conveying parameters, typically adaptively selected within a range of 10 minutes to 2 hours. The characteristic signals within each time window are normalized and input into the simulation model, which outputs a predicted value for the pipe wall thickness loss rate over the next three operating cycles (typically 72 hours). For example, if the vibration energy entropy within the current time window increases by 15% compared to the previous period, the model, combined with the flow field simulation results, predicts that the wear rate on the outer side of the elbow will increase to 0.12 mm / h within the next three days, exceeding the safety threshold of 0.1 mm / h, thereby triggering a warning signal. The Monte Carlo method is used to quantify uncertainty in the prediction process. A wear rate probability distribution curve is generated through 500 random samplings, providing a confidence interval reference for risk-based decision-making. In terms of iterative optimization of the simulation model, the system introduces an online learning mechanism. Whenever new pipeline inspection data is stored, the model incremental training process is automatically started. For example, during an on-site inspection, it was found that the actual wear amount was 8% lower than the predicted value. The system added this difference data to the training set, fine-tuned the convolution kernel weights of the CNN network through the back-propagation algorithm, and reduced the wear rate estimate under this working condition in the next round of prediction. This continuous learning mechanism allows the model prediction error to gradually converge over time. Actual data from a copper mine showed that after 6 months of operation, the average prediction error dropped from the initial 9.2% to 4.7%. The generation of slurry control signals relies on a multi-objective optimization algorithm, and the defined objective function covers three dimensions: wear rate, transportation efficiency, and energy consumption cost.Specifically, the wear rate objective uses an exponential penalty function. When the predicted wear rate approaches a safety threshold, the objective function value increases sharply, forcing parameter adjustments. The delivery efficiency objective is positively correlated with slurry flow rate and concentration, but is limited by pump power. The energy cost objective is linked to pumping pressure and operating time, forming a nonlinear constraint. A Pareto frontier search algorithm is used to find the optimal solution set, screening 200-300 non-dominated solutions from millions of parameter combinations. The final parameters are then selected based on fuzzy decision theory. The fuzzy membership function dynamically adjusts based on real-time operating conditions. For example, energy consumption is prioritized during periods of high equipment load, while wear suppression is significantly prioritized during safety alerts. In one real-world case, the system completed multi-objective optimization calculations within 30 seconds, generating a control solution that reduced the flow rate from 4.5 m / s to 4.1 m / s while simultaneously adding 0.3% drag reducer. This resulted in a 22% reduction in wear rate while increasing energy consumption by only 5%, achieving the optimal balance.

[0039] like Figure 1 As shown, the feedback optimization module uses dynamic modal decomposition (DMD) to compare the differential characteristics of the control signal and the predicted signal. First, the time series data of both are mapped into a high-dimensional phase space, and the dominant spatiotemporal modes are extracted using singular value decomposition (SVD). For example, the decay mode of the actual wear rate after control differs from the predicted exponential decay mode, indicating that the model does not fully account for the cumulative effects of material fatigue. The differential eigenvalues are analyzed using a causal reasoning framework, and a Bayesian network is constructed to infer key influencing factors. In one analysis, the system identified slurry viscosity measurement errors as the primary cause of the prediction deviation. The data review process was then initiated, revealing a systematic error of 0.8 Pa·s in the viscosity sensor. The optimization command then updated the sensor calibration coefficients and triggered parameter retraining of the simulation model, improving subsequent prediction accuracy by 12%. The implementation of this entire technology chain has significantly improved the intelligent level of pipeline wear management. At the data level, the efficient integration and repair technology of multi-source heterogeneous data breaks through the information island limitations of traditional systems; at the model level, the integrated learning framework realizes the complementary advantages of physical mechanisms and data-driven methods; at the control level, dynamic optimization and feedback mechanisms ensure the system's adaptability under complex working conditions. Industrial trials at an iron ore mine have shown that the system reduces the warning time for abnormal pipeline wear from 48 hours to 72 hours compared to traditional methods, and achieves a 96.3% accuracy in predicting wear on key parts. At the same time, by optimizing the control strategy, the system reduces pumping energy consumption by 18%, saving over 1.5 million yuan in electricity costs annually. These practical results verify the effectiveness and engineering practical value of the technical solutions described in claims 3 to 6, and provide a complete technical solution for intelligent pipeline operation and maintenance.

[0040] Specifically, the generation and optimization of slurry control signals is the core of the entire intelligent control chain. This process utilizes a multi-objective optimization algorithm to achieve a dynamic balance between wear control, conveying efficiency, and energy costs. The key lies in constructing a mathematical model that reflects complex operating constraints and designing an efficient solution strategy. The system's defined objective function encompasses three main dimensions: minimizing wear rate requires suppressing slurry particle impact damage on the pipe wall; maximizing conveying efficiency requires maintaining a reasonable slurry flow rate and concentration; and minimizing energy costs involves optimizing pumping power. These three objectives are nonlinearly coupled. For example, increasing flow rate increases conveying volume but increases wear and energy consumption, while decreasing concentration reduces particle impact but can lead to slurry segregation, impacting filling quality. To resolve these multi-objective conflicts, the system employs a modified NSGA-II algorithm for Pareto front search. This algorithm uses fast non-dominated sorting and congestion calculation to identify the optimal solution from a vast number of parameter combinations. During implementation, the algorithm initializes a randomly generated population of 5,000 individuals, each encoding 10 control parameters, including flow rate, concentration, and pump pressure. During the iterative process, population diversity is maintained by simulating binary crossover and polynomial mutation operations. Constraint handling mechanisms are also introduced to eliminate invalid solutions that violate pipe strength or equipment power limits. After 100 generations of evolution, the algorithm outputs a Pareto front containing 200-300 non-dominated solutions, which form an optimal trade-off surface in the three-dimensional target space. Fuzzy decision theory plays a key role in this stage, dynamically adjusting the shape parameters of the membership function to adapt to real-time operating conditions. The system presets triangular and trapezoidal membership functions to correspond to the satisfaction ranges of each target. For example, the membership function for the wear rate target reaches full satisfaction (membership 1) at 0.08 mm / h, and satisfaction drops to 0 when the rate exceeds 0.12 mm / h. During the decision-making process, the fuzzy inference engine combines the membership values of each target to calculate a comprehensive score and selects the parameter combination with the highest total score as the final control solution. To address the dynamic operating conditions, the system monitors external factors (such as grid electricity price fluctuations and the urgency of production tasks) in real time and adjusts the target weight coefficients. For example, the system automatically prioritizes energy cost targets during peak electricity consumption periods, while giving the highest priority to wear reduction during safety alerts. In one real-world case, when the system detected that the wear rate of a pipe section was approaching a threshold, it used fuzzy decision-making to reduce the flow rate from 4.3 m / s to 3.9 m / s and simultaneously add 0.2% drag reducer. This resulted in an 18% decrease in wear rate while increasing energy consumption by only 7%, successfully avoiding unplanned downtime.

[0041] In one embodiment of the present invention, the feedback optimization module uses dynamic modal decomposition (DMD) technology to deeply analyze the discrepancies between control effects and predicted results. When the control signal is applied to the actual system, the data acquisition unit continuously monitors key parameters such as pipe wall thickness and vibration spectrum, generating a real-time feedback signal. The system then spatially and temporally aligns the feedback signal with the simulated prediction signal and applies DMD technology to extract the dominant spatiotemporal modes of both. Specifically, the time series data is constructed into a high-dimensional state matrix, and a low-rank approximation space is obtained through singular value decomposition (SVD). This is then used to determine the eigenvalues and modes of the linear dynamic operators. For example, after a certain control, the decay mode of the actual wear rate exhibits exponential characteristics, while the predicted signal shows linear decay. The calculated correlation coefficient between the modal amplitudes of the two is 0.65, which is below the preset threshold of 0.8, triggering a deep analysis process for the difference characteristics. During this process, the system first reconstructs the phase space of the difference signal, mapping the one-dimensional sequence into a high-dimensional phase space using time-delay embedding. The system's dynamic characteristics are then determined by calculating the maximum Lyapunov exponent. If the exponent is greater than zero, the system is in a chaotic state and requires nonlinear analysis. If the exponent approaches zero, a linear approximation model is appropriate. Subsequently, a singular value decomposition (SWD) was performed on the reconstructed phase space trajectory, extracting the top five dominant modal components as differential eigenvalues. These components often correspond to physical mechanisms not fully captured by the model (such as the cumulative effects of material fatigue or unmodeled environmental interference). During the differential eigenvalue analysis phase, the system employed a causal inference framework to construct a Bayesian network to reveal the causal relationship between control parameters and wear characteristics. The network nodes included 35 variables, including slurry parameters (flow rate, concentration, etc.), equipment status (pump pressure, temperature, etc.), and wear indicators (thickness loss rate, surface roughness, etc.). The conditional probability distribution was estimated using the Markov Chain Monte Carlo method, and Gibbs sampling was used to traverse the parameter space. After 5000 iterations, the network converged to a stable probabilistic network. For example, when analyzing a particular prediction deviation, the network revealed that slurry viscosity measurement error contributed 42% to the wear variance, far exceeding other factors. The system then initiated data traceability and discovered a systematic drift of 0.6 Pa·s in the viscosity sensor, immediately triggering a calibration procedure and updating the historical data labels in the database. Simultaneously, the differential eigenvalues are encoded as feature vectors and fed into the incremental learning module, where the parameters of the pipeline wear simulation model are fine-tuned using an online backpropagation algorithm. In specific implementation, the fully connected weights of the final layer of the 3D convolutional neural network are adaptively adjusted at a learning rate of 0.01, while the subtree structure of the gradient boosting decision tree incorporates new feature interactions by adding new split nodes. This progressive optimization strategy enables the model to quickly adapt to new operating conditions without forgetting existing knowledge. Data from an iron ore mine application shows that after three rounds of incremental learning, the model's prediction error under similar operating conditions dropped from 12.3% to 6.8%.

[0042] Furthermore, the generation of feedback optimization signals not only includes parameter adjustment suggestions, but also involves system-level status assessment and maintenance strategy optimization. After identifying key influencing factors, the system combines the equipment health model to predict the remaining service life and generates a multi-level response strategy: for short-term reversible deviations (such as sensor drift), online calibration and model retraining are automatically performed; for medium-term trend changes (such as material performance degradation), it is recommended to adjust the maintenance cycle or spare parts procurement plan; for long-term systemic risks (such as design defects), it triggers engineering modification suggestions. For example, in one case, the system detected that the wear rate of the bend was continuously higher than the predicted value. After causal analysis, it was found that the insufficient curvature radius of the bend led to the intensification of secondary flow. Finally, it was recommended to increase the curvature radius from 2D to 3D. After implementation, the wear rate in this area decreased by 37%. The entire feedback optimization process forms a closed-loop control system, demonstrating its innovation in three key aspects. First, the combination of dynamic mode decomposition (DMD) and causal reasoning enables traceability analysis from data discrepancies to physical mechanisms, overcoming the limitation of traditional statistical methods that can only detect correlations but not causal relationships. Second, an incremental learning mechanism ensures continuous model evolution, enabling tracking of the time-varying characteristics of the pipeline system. Finally, a multi-level response strategy seamlessly connects micro-parameter adjustments with macro-maintenance decisions, significantly improving overall system reliability. Operational data from a demonstration project demonstrates that this module has reduced unplanned downtime by 55% and unplanned maintenance costs by 41%, fully demonstrating its technical advantages. Technically, the system utilizes a distributed computing architecture to handle massive data processing demands. The multi-objective optimization algorithm is deployed on a GPU cluster, leveraging CUDA parallel computing to accelerate the evolutionary process, enabling searches across a million-parameter space to be performed in under 30 seconds. Dynamic mode decomposition and Bayesian network reasoning are run on a CPU cluster, utilizing the MPI protocol to parallelize large-scale matrix operations. The incremental learning process uses the Elastic Weight Conservation (EWC) algorithm to prevent catastrophic forgetting. By calculating a parameter importance matrix to constrain weight updates, the system ensures the compatibility of new and old knowledge. All optimization instructions and feedback data are seamlessly integrated with the industrial control system via the OPC UA protocol, enabling end-to-end automation from decision-making to execution.

[0043] like Figure 2As shown, the present invention's intelligent testing method for wear of a slurry conveying pipeline based on big data includes the following steps: S100: real-time acquisition of pipeline signals indicating slurry flow within the conveying pipeline. S200: receiving and storing the pipeline signals, constructing a multidimensional pipeline database by integrating historical pipeline signal datasets and external pipeline operation data acquired through the network, and performing data preprocessing on the pipeline signals to generate characteristic signals. S300: constructing a pipeline wear simulation model based on the multidimensional pipeline database, inputting the characteristic signals into the simulation model for dynamic wear analysis, and generating a simulation prediction signal containing predicted wear parameters and a slurry control signal representing optimal operating parameters. S400: transmitting the slurry control signal to an execution unit of the slurry conveying system to adjust the slurry parameters. S500: acquiring a real-time pipeline operation feedback signal after adjustment, performing a feature comparison between the feedback signal and the simulation prediction signal, extracting the difference between the two, and generating an optimization instruction containing a parameter correction strategy. S600: synchronously feeding back the optimization instruction to the pipeline wear simulation model and the multidimensional pipeline database to drive the simulation model to iteratively update parameters.

[0044] Furthermore, the present invention constructs a complete closed-loop control system, which realizes intelligence and adaptability in the whole process from data acquisition to model iteration. The core of this method is to break through the static analysis and experience-dependent mode in traditional pipeline wear management through the synergy of multi-source data fusion, dynamic simulation modeling and real-time feedback optimization, and provide a systematic solution for pipeline health management under complex working conditions. In specific implementation, the system first captures the multi-physical field signals of slurry flow in the pipeline in real time through a distributed sensor network, including pressure pulsation, vibration spectrum, pipe wall thickness change and environmental parameters. These raw data are continuously input into the central processing unit at a sampling rate of tens of thousands of points per second. For example, in the test case of a DN300 pipeline in an iron ore mine, 48 sensor nodes arranged along a 200-meter pipe section generate a data packet containing 128-dimensional features such as pressure, vibration, and temperature every 10 seconds, which is transmitted to the edge computing node via industrial Ethernet for preliminary filtering and timestamp alignment to ensure the timing consistency of subsequent analysis. In the data storage and preprocessing stage, the system adopts a layered architecture to process massive data streams. The raw signal first enters a buffer for format standardization. Data with different sampling frequencies are linearly interpolated to a 10ms time base, and spatial coordinate labels (such as pipeline mileage and station numbers) are attached to each data point. Historical and real-time data are co-stored in a distributed database (such as Cassandra). High-frequency vibration data is stored in a column-based format to optimize query efficiency, while low-frequency thickness measurement data is indexed using time series to support rapid retrospective analysis. The preprocessing module uses wavelet threshold denoising to eliminate on-site electromagnetic interference and applies empirical mode decomposition to separate trend terms from random fluctuations in the signal. For example, the vibration signal from a copper mine pipeline was successfully filtered out, despite the presence of interference from the rock crusher (characteristic frequency 25Hz), retaining the 0.5-15Hz frequency band that truly reflects the slurry flow state. The feature extraction stage utilizes a multi-scale analysis framework, performing wavelet packet decomposition on the pressure signal to obtain the energy entropy of 32 subbands. The time-domain kurtosis index and frequency-domain centroid frequency of the vibration signal are calculated. This generates a 56-dimensional feature vector, which is then compressed to 18 dimensions using principal component analysis to reduce model complexity. The core of dynamic wear analysis lies in the construction of a hybrid simulation model, which innovatively combines the advantages of physical mechanisms and data-driven methods. During the model training phase, the system integrates historical operating data, laboratory wear test data, and an external engineering case library. Through transfer learning technology, the wear patterns of different pipe diameters and materials are mapped to a unified feature space. For example, the wear data of DN200 steel pipes was dimensionlessly processed and used to assist in training the prediction model for DN250 composite pipes, which improved the accuracy of the initial model by 23%. During online prediction, the model uses a sliding time window mechanism to process real-time data streams. The length of each window is dynamically adjusted according to the volatility of the operating conditions (usually 10-30 minutes). The feature vectors within the window are standardized and input into the integrated model.In a gold mine application, when a sudden 20% increase in vibration energy entropy was detected in a pipe bend, the model, combined with flow field simulation results, predicted that the wear rate at that location would rise to 0.15 mm / h within the next 72 hours, exceeding the safety threshold of 0.1 mm / h. This immediately triggered the generation of control instructions. Uncertainty was quantified during the prediction process, and a wear rate probability distribution curve was generated through Monte Carlo simulation, providing confidence intervals for risk assessment. For example, one prediction showed a 95% confidence interval of [0.12, 0.18] mm / h, providing a key basis for operational and maintenance decisions. A multi-objective evolutionary algorithm was used to optimize control parameters, seeking the optimal balance between conveying efficiency and energy costs while ensuring wear safety. The algorithm encoded 10 adjustable parameters, such as flow rate, concentration, and pump pressure, as decision variables. It defined a fitness function that encompassed wear suppression, production capacity requirements, and energy consumption constraints. The NSGA-III algorithm then searched for the Pareto optimal set within a million-level solution space. In actual operation at a lead-zinc mine, the system completed 500 generations of evolutionary calculations in 45 seconds, screening out 153 non-dominated solutions. Fuzzy decision-making then selected a solution that reduced the flow rate from 5.1 m / s to 4.7 m / s and added 0.15% drag reducer. This resulted in a 19% drop in wear rate while increasing energy consumption by only 5%. Control commands were transmitted to the PLC control system via the OPC UA protocol, driving actuators such as the variable frequency pump and dosing device to adjust parameters. The entire process is fully automated, with less than 2% manual intervention.

[0045] In one embodiment of the present invention, a feedback optimization mechanism contributes to the system's self-evolutionary capabilities. After each control is implemented, the data acquisition module continuously monitors the actual wear progression and compares it with the predicted values in spatiotemporal alignment. Difference analysis utilizes dynamic modal decomposition to extract dominant modal features. For example, after a certain control, the actual wear curve exhibits exponential decay, while the predicted model output shows linear decay. The system identifies material fatigue effects not accounted for by the model through the modal correlation coefficient (0.68, below the threshold of 0.8). The causal inference engine then constructs a Bayesian network, analyzing that pipe hardness measurement error is the primary influencing factor (contributing 37%). Triggering the material testing process, it is discovered that the actual Brinell hardness is 8.2 HB lower than the designed value. The system then initiates incremental learning, adding the new material property data to the training set. Using an elastic weight curing algorithm, the neural network parameters are updated, reducing subsequent prediction errors by 6.5%. Furthermore, a new "hardness compensation coefficient" dimension has been added to the optimization instruction library to ensure that control strategies under similar operating conditions automatically adapt to material property changes. A case study at an iron ore mine demonstrated that the system reduced the early warning response time for abnormal pipeline wear from 36 hours compared to traditional methods to 8 hours, achieving a 94.7% accuracy in predicting wear in key areas. By continuously optimizing control strategies, the average pipeline service life was extended from 11 months to 16 months, reducing annual maintenance costs by 2.8 million yuan. More notably, the system's self-learning capabilities enabled the model prediction error to converge from an initial 8.9% to 4.3% after six months of operation, demonstrating its strong adaptability to environmental conditions. This closed-loop "perception-decision-execution-optimization" system not only overcomes the data silos and model rigidity inherent in traditional methods, but also enables continuous system evolution through real-time feedback mechanisms, providing reliable technical support for the development of intelligent mines.

[0046] The present invention's intelligent, big-data-based testing system and method for slurry pipeline wear systematically overcomes the limitations of traditional technologies by integrating multi-dimensional data acquisition, dynamic simulation analysis, and intelligent closed-loop control. Specifically, this solution constructs an intelligent monitoring system based on a multi-physics field sensor network and an adaptive learning framework. Its technical implementation path comprises four core modules: a high-density sensor array for holographic perception of pipeline status; a distributed data processing architecture for spatiotemporal alignment and feature extraction of multimodal data; a hybrid modeling approach that integrates the dual advantages of physical mechanisms and data-driven approaches; and a closed-loop optimization mechanism for the coordinated evolution of parameter control and model iteration.

[0047] Therefore, the intelligent testing system and method for wear of filling slurry conveying pipelines based on big data of the present invention can solve the problems of untimely wear monitoring of filling slurry conveying pipelines and wear assessment accuracy generally lower than actual requirements.

[0048] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by one of ordinary skill in the art without departing from the spirit and technical principles disclosed herein are intended to be covered by the claims of the present invention.

Claims

1. The intelligent testing system for wear of filling slurry conveying pipeline based on big data is characterized by: include: a pipeline testing module, wherein the pipeline testing module generates a pipeline signal; a database module, the database module including a multi-dimensional pipeline database and a pre-processing module, the database module receiving and storing the pipeline signal transmitted by the pipeline testing module, and constructing the multi-dimensional pipeline database by integrating historical pipeline signal data sets and external pipeline operation data acquired through the network; the pre-processing module receiving the pipeline signal and performing data pre-processing to form a characteristic signal; A data analysis module, wherein the data analysis module constructs a pipeline wear simulation model based on a multi-dimensional pipeline database, the characteristic signal is input into the pipeline wear simulation model to perform dynamic wear analysis and generate a simulation prediction signal and a control parameter signal, the control parameter signal is transmitted to the pipeline testing module to adjust the slurry delivery parameters, and the pipeline testing module generates a slurry control signal; A feedback optimization module receives and compares the slurry control signal and the simulation prediction signal, extracts the difference characteristic value between the two, analyzes and generates a feedback optimization signal, and inputs it to the data analysis module and the database module; when the feedback optimization module compares the slurry control signal and the simulation prediction signal, after the control signal acts on the actual system, the data acquisition unit continuously monitors the key parameters of the pipe wall to form a real-time feedback signal. After the system aligns the feedback signal with the simulation prediction signal in time and space, it applies DMD technology to extract the dominant time and space modes of the two, and evaluates the control effect by calculating the mode amplitude correlation coefficient. When the correlation coefficient is lower than the preset threshold, the difference feature depth is activated. Analysis process, when performing the difference feature deep analysis process, after reconstructing the phase space of the slurry control signal with differences, calculate the maximum Lyapunov exponent to judge the chaotic characteristics of the system, and extract the key modal components based on singular value decomposition to form the difference eigenvalues. When performing the difference eigenvalue analysis, introduce a causal reasoning framework, construct a Bayesian network with slurry parameters as dependent variables and wear characteristics as result variables, estimate the conditional probability distribution through the Markov chain Monte Carlo method, identify the control parameters whose contribution to wear differences exceeds the preset value, generate the feedback optimization signal containing the parameter adjustment priority list, and update the parameter weights of the pipeline wear simulation model through the incremental learning mechanism.

2. The intelligent testing system for filler slurry conveying pipeline wear based on big data according to claim 1 is characterized in that: The pipeline testing module includes a test pipeline unit, a slurry circulation unit and a data acquisition unit. The slurry circulation unit controls the slurry to be delivered to the test pipeline unit. The data acquisition unit is arranged in the test pipeline unit. The data acquisition unit collects various data when the slurry flows through the test pipeline unit and forms the pipeline signal.

3. The intelligent testing system for filler slurry conveying pipeline wear based on big data according to claim 1 is characterized in that: When the database module integrates historical pipeline signal data sets with external pipeline operation data, it includes a data fusion stage, a data cleaning stage, and a data repair stage. The data fusion stage uses a distributed stream processing engine to perform real-time access and format alignment on the multi-source heterogeneous pipeline signals. The pipeline signals include but are not limited to pressure waveforms, vibration modal spectra, and pipe wall thickness measurement data under different geological conditions, pipe diameter specifications, and conveying media. The data cleaning stage uses an adaptive threshold filtering algorithm to eliminate pulse noise interference and separates trend items and random fluctuation components in the pipeline signals based on time series decomposition technology. The data repair stage uses a spatiotemporal correlation modeling method to construct a tensor filling model with pipe segment position as the spatial dimension and operation cycle as the time dimension, and uses the alternating least squares method to optimize the objective function to complete data repair. The processed standardized data establishes a spatial index according to the pipe segment topology structure, and a time series database is used to compress and store high-frequency sampling data to form the database module.

4. The intelligent testing system for filler slurry conveying pipeline wear based on big data according to claim 1 is characterized in that: When the preprocessing module receives the pipeline signal and performs data preprocessing to form a characteristic signal, it first performs multi-scale feature extraction on the pipeline signal, uses wavelet packet transform to decompose the time domain signal of the pipeline signal into different frequency band subspaces, and calculates the energy entropy of each subband as the frequency domain feature quantity. At the same time, the intrinsic mode function component of the signal is obtained through empirical mode decomposition, and its instantaneous amplitude-frequency characteristics are extracted to form a time-frequency joint feature vector; after normalizing the feature vector, the principal component analysis algorithm is used to reduce the feature dimension, retain the feature components whose cumulative contribution rate exceeds the preset threshold, and form the characteristic signal. The generated feature signal is pushed to the data analysis module in real time through the message queue.

5. The intelligent testing system for filler slurry conveying pipeline wear based on big data according to claim 1 is characterized in that: The pipeline wear simulation model is constructed using an integrated learning method to improve prediction accuracy by stacking multiple base models. The pipeline wear simulation model includes a first-layer base model, a second-layer base model, and a third-layer base model. The first-layer base model contains a fluid dynamics equation solver based on physical mechanisms to calculate the slurry flow field distribution and particle collision energy; the second-layer base model uses a gradient boosting decision tree to capture the nonlinear relationship in historical data; the third-layer base model is a three-dimensional convolutional neural network to process the microscopic wear morphology features in the pipeline cross-section detection image, and finally dynamically adjusts the weight coefficients of each base model through a meta-learning framework to form the pipeline wear simulation model.

6. The intelligent testing system for filler slurry conveying pipeline wear based on big data according to claim 1 is characterized in that: When performing the dynamic wear analysis, a sliding time window mechanism is used to divide the continuous pipeline signal into analysis segments of fixed length according to the time-varying characteristics of the slurry conveying parameters. After the characteristic signal in each time window is input into the simulation model, the output contains the predicted value of the pipe wall thickness loss rate for the next three operating cycles, and forms the simulation prediction signal.

7. The intelligent testing system for filler slurry conveying pipeline wear based on big data according to claim 1 is characterized in that: The slurry control signal is generated by a multi-objective optimization algorithm. The defined objective function variables include but are not limited to three optimization dimensions: minimizing wear rate, maximizing conveying efficiency, and minimizing energy consumption cost. The optimal solution set is found through the Pareto frontier search algorithm, and the parameter combination with the highest comprehensive score is selected based on fuzzy decision theory. The fuzzy membership function dynamically adjusts the shape parameters according to the real-time working conditions and forms the slurry control signal.

8. An intelligent test method for wear of a filling slurry conveying pipeline based on big data using any one of claims 1 to 7, characterized in that: include: S100: real-time collection of pipeline signals of slurry flow in the delivery pipeline; S2 00: receiving and storing the pipeline signal, building a multi-dimensional pipeline database by integrating historical pipeline signal data sets and external pipeline operation data obtained through the network, and performing data preprocessing on the pipeline signal to generate a characteristic signal; S300: constructing a pipeline wear simulation model based on the multi-dimensional pipeline database, inputting the characteristic signal into the simulation model to perform dynamic wear analysis, and generating a simulation prediction signal including predicted wear parameters and a slurry control signal representing optimal operating parameters; S400: Transmitting the slurry control signal to an execution unit of a slurry delivery system to adjust slurry parameters; S5 00: Acquire the adjusted pipeline operation feedback signal in real time, compare the characteristics of the feedback signal with the simulation prediction signal, extract the difference characteristic value between the two, and generate an optimization instruction including a parameter correction strategy; S600: Synchronously feeding back the optimization instruction to the pipeline wear simulation model and the multi-dimensional pipeline database, driving the simulation model to iteratively update parameters.

Images