A method and system for correcting oil particle data of a heading machine based on multi-source simulation data

Abstract

This invention relates to a method and system for correcting abrasive particle data of tunnel boring machine (TBM) oil based on multi-source simulation data, belonging to the field of intelligent TBM technology. The method includes the following steps: Based on the geometry of the main pipeline of the TBM's main drive oil and the bypass where the abrasive particle monitoring device is located, the main drive pipeline and bypass flow domains are extracted and discretized to obtain a main drive pipeline-bypass flow domain grid; the momentum exchange process between the continuous phase of the oil and the discrete phase of the abrasive particles is calculated bidirectionally to construct a CFD-DPM numerical simulation model, and physicochemical parameters and boundary conditions are set to obtain the numerical simulation solution results; a prediction model is constructed using an optimal surrogate model set, and the initial abrasive particle proportion data is corrected based on the predicted abrasive particle distribution data of the main pipeline and combined with the simulated abrasive particle proportion data. This invention achieves accurate estimation of abrasive particle data of TBM oil by combining the CFD-DPM numerical simulation model with the surrogate model.

Description

Technical Field

[0001] This invention belongs to the field of intelligent tunnel boring machine technology, specifically relating to a method and system for correcting oil abrasive particle data of tunnel boring machines based on multi-source simulation data. Background Technology

[0002] With the networking of urban underground space, the diversification of energy development and storage methods, and the in-depth advancement of transportation construction in western my country, the demand for the construction of deep, large-diameter, and complex geological tunnels is constantly increasing. Among these, tunnel boring machines have been widely used due to their high efficiency and safety advantages.

[0003] The main drive system of a tunnel boring machine (TBM) is the core subsystem that drives the cutterhead to rotate, completing rock cutting and stripping, and bearing the main loads. It consists of components such as a motor, reducer, and main bearing. High burial depth and complex geological conditions cause the TBM main drive system to withstand strong impact loads, and its large diameter causes it to withstand reciprocating alternating loads. This leads to wear, fracture, and adhesion of the main drive components, resulting in main drive failures and even project delays. Therefore, early warning of TBM main drive system failures is necessary, and these failures often manifest as abnormal wear in their early stages. Therefore, abnormal wear monitoring of the main drive system is an effective means of achieving early warning of failures. Based on this, timely handling and maintenance strategies can be developed to reduce the probability of failures and even downtime, which is of great significance for the safe, reliable, and stable operation of the TBM.

[0004] The proportion of abrasive particles of different diameters within the hydraulic fluid is a direct variable reflecting the wear state of components such as hydraulic fluid, gears, and bearings. With the rapid development of materials science, electronics, and computer vision technologies, online abrasive particle monitoring technology based on electromagnetic induction and computer vision is widely used in equipment in aviation, shipbuilding, and gas turbine industries. To accurately estimate the wear state of the main drive system of tunnel boring machines (TBMs), manufacturers and research institutions have deployed online abrasive particle monitoring devices to obtain the proportion of abrasive particles of different diameters in the TBM's main drive hydraulic fluid. However, these monitoring devices are often not installed on the main pipeline but rather through an additional bypass. The diameter difference between the main pipeline and the bypass is significant (generally more than 5 times), resulting in a difference in the proportion and distribution of abrasive particles of different diameters between the main pipeline and the bypass where the monitoring device is located. In other words, the results obtained from hydraulic abrasive particle monitoring cannot directly and accurately reflect the proportion and distribution of abrasive particles of different diameters in the main pipeline. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for correcting the wear particles in tunnel boring machine oil based on multi-source simulation data, so as to realize the accurate estimation of the proportion and distribution of wear particles of different diameters in tunnel boring machine oil, and to provide a reference for estimating the main drive operating status.

[0006] To address the aforementioned technical problems, a method for correcting tunneling machine oil abrasive particle data based on multi-source simulation data is provided, comprising the following steps: Based on the geometry of the main pipeline of the tunneling machine's main drive hydraulic system and the bypass where the hydraulic abrasive monitoring device is located, the main drive pipeline and bypass flow domains are extracted. These flow domains are then discretized to obtain a main drive pipeline-bypass flow domain mesh. Based on this mesh, a bidirectional coupled calculation is performed on the momentum exchange process between the continuous hydraulic phase and the discrete abrasive phase, constructing a CFD-DPM numerical simulation model. The physicochemical parameters and boundary conditions of the CFD-DPM numerical simulation model are set, and the numerical simulation results are obtained. Based on the numerical simulation results, cross-sections are set on the main pipeline and bypass, and the initial abrasive proportion data are obtained using these cross-sections as flow surfaces. Based on the actual operating conditions of the main pipeline, a... The experiment schemes are divided into groups; the CFD-DPM numerical simulation model is used to generate the abrasive particle simulation ratio data corresponding to each group of experimental schemes; a prediction model is constructed through the optimal surrogate model set, taking the measured abrasive particle distribution data of the bypass as input and outputting the predicted abrasive particle distribution data of the main pipeline; based on the predicted abrasive particle distribution data of the main pipeline, the initial abrasive particle ratio data is corrected in combination with the abrasive particle simulation ratio data.

[0007] Based on the correction of abrasive data of tunnel boring machine oil, the oil monitoring data of tunnel boring machine can be accurately corrected, providing data support for the evaluation of the main drive operation status of tunnel boring machine, and improving the construction quality and intelligent level of tunnel boring machine tunneling.

[0008] Preferably, the extraction process of the main drive pipeline and bypass watershed includes: Boolean operations are used to remove minute structures in the main drive oil pipeline and bypass, resulting in initial main pipeline and initial bypass. Screening conditions are then established to filter these initial main pipeline and initial bypass. The screening conditions are: the screening process retains key flow channel regions that significantly affect the oil flow field distribution and abrasive particle transport characteristics; and the lengths of the main pipeline and bypass are equal to their respective diameters. times, .

[0009] This ensures the accuracy of the analysis results.

[0010] Preferably, a tetrahedral mesh structure is used for discretization, and a boundary layer mesh is added in the corresponding regions of the main pipeline and the bypass, with local mesh refinement performed in the connection region between the main pipeline and the bypass; the main drive pipeline-bypass watershed mesh satisfies: The mesh was tested for orthogonality quality, with a minimum orthogonality quality of 0.2 or higher; a boundary layer mesh was used in the near-wall region, and the dimensionless distance y⁺ from the wall was controlled within the range of 30 to 300.

[0011] It can ensure both computational accuracy and computational efficiency.

[0012] Preferably, the construction process of the CFD-DPM numerical simulation model is as follows: A two-way coupled computational model of the continuous oil phase and the discrete abrasive phase was built based on the Euler–Lagrange method. The continuous oil phase adopted… Turbulence calculations were performed using the equation model, with a pressure-based solver and a pressure-velocity coupled algorithm. Lagrangian trajectories of discrete phase abrasive particles were tracked based on the steady-state flow field. The material properties of the two-phase flow were set according to the actual physicochemical parameters of the main drive oil and abrasive particles of the tunnel boring machine. The abrasive particles were set to be spherical, and their particle size distribution was described using a distribution function. The CFD-DPM numerical simulation model was then constructed.

[0013] It provides data support for evaluating the main drive operation status of tunnel boring machines, thereby improving the construction quality and intelligent level of tunnel boring machine tunneling.

[0014] Preferably, the solution process of the CFD-DPM numerical simulation model is as follows: The main pipeline inlet is set as a velocity inlet, and corresponding inlet velocity, turbulence intensity, and turbulent viscosity are applied according to the operating conditions. The bypass inlet is connected to the main pipeline and inherits its flow parameters. Both the main pipeline outlet and the bypass outlet are set as pressure outlets. The pipeline wall is set as a no-slip stationary wall, and the particles and the wall adopt a reflective boundary condition. The elastic collision behavior between the abrasive particles and the wall is described by the Grant restitution coefficient model. A surface jet source is set at the main pipeline inlet section, and the particles are randomly generated at the inlet section. The particle size follows the actual operating condition distribution, and the initial jet velocity matches the direction and magnitude of the local velocity of the continuous phase. The Lagrangian method is used for particle trajectory tracking. The maximum number of tracking steps, time step, and coupling iteration interval are determined according to the range of the main driving main pipeline and bypass flow domain. The momentum equation, turbulent kinetic energy, and dissipation rate adopt the first-order upwind scheme. The residual convergence standard value of the oil continuous phase and the abrasive discrete phase is set, and the mass flow rate and particle number of the main pipeline and bypass sections are monitored at the same time. When the fluctuation amplitude of the mass flow rate and particle number is less than the preset value, it is judged as convergence, and the numerical calculation results and iteration curves of the CFD-DPM model are obtained.

[0015] In order to obtain the proportion of abrasive particles of different diameters in the main channel and the bypass channel.

[0016] Preferably, both the initial abrasive particle proportion data and the simulated abrasive particle proportion data are two-dimensional data, each including: The abrasive ratio data of the main pipeline and bypass corresponding to the diameter sub-interval, the abrasive distribution data of the main pipeline and the bypass in the full diameter interval; the diameter sub-interval is a sub-interval obtained by discretely dividing the full diameter interval, and the full diameter interval is a complete interval consisting of the minimum diameter to the maximum diameter of the abrasive particles of the main drive oil of the tunnel boring machine.

[0017] This facilitates the design of the main drive oil lubrication parameters for the tunnel boring machine.

[0018] Preferably, the process for determining the full diameter range is as follows: Based on the initial proportion data of abrasive particles, the minimum and maximum diameters of the abrasive particles in the main drive oil are determined, and the range from the minimum to the maximum diameter is defined as the full diameter range of the abrasive particles in the main drive oil. Correspondingly, the diameter sub-range is a number of abrasive particle diameter ranges obtained by discretizing the full diameter range in an equal or non-equal interval manner, and the number of diameter sub-ranges obtained is not less than 15.

[0019] This provides a reference for estimating the state of the main drive hydraulic fluid in tunnel boring machines.

[0020] Preferably, the process for determining the optimal proxy model set is as follows: A set of candidate surrogate models is pre-defined, which includes at least: multinomial regression model, radial basis function network model, kriging model, and support vector regression model; the accuracy index value of each abrasive grain diameter sub-interval is obtained by evaluating the accuracy of the models in the candidate surrogate model set. For each abrasive grain diameter sub-range, select the accuracy index value. The highest-performing model is taken as the optimal surrogate model for that sub-interval, and the set of optimal surrogate models is the prediction model; wherein, the model accuracy index value The calculation formula is as follows: , In the formula, Indicates the first Group experiment in the first Measured values ​​for each diameter sub-interval; , , m The number of diameter subintervals; Models within the set of candidate proxy models Regarding the first Group experiment in the first Estimates for each diameter sub-interval; This is the average of the estimated values ​​of all models within the candidate proxy model set.

[0021] The closer the R-value is to 1.00, the higher the accuracy of the model.

[0022] Preferably, the correction process for the initial abrasive particle ratio data is as follows: The abrasive distribution data of different diameter sub-intervals of the bypass measured by the abrasive sensor, the actual flow velocity of the main drive hydraulic pipeline of the tunnel boring machine, and the actual abrasive flow rate are input into the optimal surrogate model corresponding to each diameter sub-interval, and the abrasive distribution prediction data of different diameter sub-intervals of the main pipeline are output. Based on the abrasive distribution prediction data, the initial abrasive proportion data obtained based on the numerical simulation solution is corrected to obtain the actual data of the abrasive proportion of different diameter sub-intervals of the main pipeline, thus completing the correction of the initial abrasive proportion data.

[0023] This significantly reduces the measurement error of the proportion of abrasive grains with different diameters, proving the innovation and feasibility of the invention.

[0024] On the other hand, a tunneling machine oil abrasive data correction system based on multi-source simulation data is also provided, for implementing the aforementioned tunneling machine oil abrasive data correction method based on multi-source simulation data, including: The first module is configured to extract the main drive pipeline and bypass flow domain based on the geometry of the main drive pipeline of the tunneling machine and the bypass where the oil wear monitoring device is located, and to discretize the main drive pipeline and bypass flow domain to obtain the main drive pipeline-bypass flow domain grid. The second module is configured to perform bidirectional coupled calculations on the momentum exchange process between the continuous oil phase and the discrete abrasive phase based on the main drive pipeline-bypass watershed grid, and construct a CFD-DPM numerical simulation model; set the physicochemical parameters and boundary conditions of the CFD-DPM numerical simulation model, and obtain the numerical simulation solution results. The third module is configured to set cross-sections on the main pipeline and bypass based on the numerical simulation results, and obtain initial abrasive particle proportion data using these cross-sections as flow surfaces; based on the actual operating conditions of the main pipeline, it forms... Experimental schemes were designed; the CFD-DPM numerical simulation model was used to generate the abrasive particle simulation ratio data for each experimental scheme. The fourth module is configured to construct a prediction model using the optimal surrogate model set, taking the measured abrasive particle distribution data of the bypass as input and outputting the predicted abrasive particle distribution data of the main pipeline; based on the predicted abrasive particle distribution data of the main pipeline, the initial abrasive particle proportion data is corrected by combining the simulated abrasive particle proportion data.

[0025] By constructing a corresponding surrogate model to characterize the mapping relationship between the proportion of abrasive particles of different diameters in the main pipeline and the bypass, the correction and estimation of the proportion of abrasive particles of different diameters in the main pipeline can be realized based on the abrasive particle measurement data of the bypass.

[0026] The beneficial effects of this invention are: 1. By constructing a CFD-DPM numerical simulation model of the main drive hydraulic pipeline and bypass of a tunnel boring machine (TBM), the mapping relationship of the proportion of abrasive particles of different diameters in the main pipeline and bypass is obtained, providing a reference for estimating the state of the main drive hydraulic fluid of the TBM. Based on the surrogate model and the correction of abrasive particle data of the TBM hydraulic fluid using CFD-DPM, the hydraulic fluid monitoring data of the TBM can be accurately corrected, providing data support for the evaluation of the main drive operating status of the TBM and improving the construction quality and intelligent level of tunnel boring machine tunneling.

[0027] 2. Experimental design based on the Latin hypersolution method was carried out. The proportion of abrasive particles of different diameters in the main drive hydraulic pipeline and bypass under different working conditions was obtained by CFD-DPM numerical simulation model of the main drive hydraulic pipeline and bypass under different working conditions. The proportion of abrasive particles of different diameters in the main drive hydraulic pipeline was accurately estimated by a surrogate model based on given working conditions and hydraulic monitoring data (bypass). Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating the oil abrasive data correction method of the present invention; Figure 2 This is a schematic diagram of the installation method of the main drive oil monitoring device of the present invention; Figure 3 This is a schematic diagram of the main view of the oil abrasive monitoring device of the present invention; Figure 4 This is a schematic diagram of the right side of the oil abrasive particle monitoring device of the present invention; Figure 5 This is a schematic diagram of the main drive main pipeline-bypass watershed of the present invention; Figure 6 This is a schematic diagram of the main drive main pipeline-bypass watershed grid of the present invention; Figure 7 This is a schematic diagram of the iterative curves and solution diagram of the CFD-DPM numerical simulation of the main drive pipeline and bypass in this invention; Figure 8 This is a schematic diagram showing the proportion of abrasive particles of different diameters in the main pipeline and bypass obtained by CFD-DPM numerical simulation in this invention. Figure 9 This is a schematic diagram of the experimental design results of the CFD-DPM numerical simulation of the main drive pipeline and bypass in this invention; Figure 10 This is a schematic diagram comparing the accuracy of the main drive main circuit and the bypass CFD-DPM proxy model of the present invention; Figure 11 This is a schematic diagram comparing the results of oil abrasive monitoring before and after correction according to the present invention; Figure 12 This is a schematic diagram of the architecture of the oil abrasive data correction system of the present invention; Figure 2The components include: 1. TBM main drive oil bypass; 2. TBM main drive oil main pipeline; 3. Oil abrasion monitoring device. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be described below with reference to the accompanying drawings. The embodiments described below are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] Example 1 This invention provides a method for correcting the wear particles of tunneling machine oil based on multi-source simulation data. The multi-source simulation data is realized through a CFD-DPM numerical simulation model and a surrogate model. The CFD-DPM numerical simulation model refers to the computational fluid dynamics-discrete phase model, which is a numerical simulation method that combines computational fluid dynamics with a discrete phase model. The surrogate model refers to the fact that most engineering design problems require simulation experiments to evaluate the objective function and constraint function when different design parameters are used.

[0031] By constructing a CFD-DPM numerical simulation model of the main drive hydraulic pipeline and bypass of a tunnel boring machine, the mapping relationship of the proportion of abrasive particles of different diameters in the main pipeline and bypass is obtained. This solves the problem of lack of basis for inferring the actual situation of abrasive particles in the main pipeline by monitoring abrasive particles in the hydraulic system, and provides a reference for estimating the state of the main drive hydraulic system of a tunnel boring machine.

[0032] A flange is added to the main drive pipeline of the tunnel boring machine to lead out a bypass for installing an oil abrasive monitoring device 3, such as... Figures 2-4 As shown, the oil abrasive monitoring device 3 indirectly obtains abrasive information from the TBM main drive oil main pipeline 2 by monitoring the abrasive particles in the TBM main drive oil bypass 1. Considering the oil flow field and the differences in drag force between abrasive particles of different diameters, large-diameter abrasive particles are unlikely to enter the smaller-diameter bypass. Therefore, it is necessary to correct the monitoring results of the TBM main drive oil bypass 1 to accurately estimate the actual situation of the abrasive particles in the TBM main drive oil main pipeline 2. Figure 1 As shown, it includes the following steps: Based on the geometric structure of the main pipeline of the tunneling machine's main drive oil and the bypass where the oil wear monitoring device 3 is located, the main drive pipeline and bypass flow domains are extracted. The extraction method can employ Boolean operations, which are a logical deduction method using digital symbols. The main drive pipeline and bypass flow domains are then discretized to obtain a main drive pipeline-bypass flow domain mesh. In this embodiment, a tetrahedral unstructured network is used for discretization. A tetrahedral unstructured network is a network structure composed of irregular tetrahedrons as basic voxels, dividing the space into a series of adjacent but non-overlapping tetrahedrons.

[0033] Based on the main drive pipeline-bypass flow domain grid, a bidirectional coupled calculation is performed on the momentum exchange process between the continuous oil phase and the discrete abrasive phase to construct a CFD-DPM numerical simulation model. In this embodiment, the Euler-Lagrange method is used to perform a bidirectional coupled calculation on the momentum exchange process between the oil and abrasive particles. The Euler-Lagrange method is a core motion description and numerical solution method in the fields of continuum mechanics, fluid mechanics, multiphase flow, and particle simulation. The background field of the fluid / continuous medium is described by the Euler method (fixed coordinate system, focusing on the change of physical quantities at spatial points), and the discrete phase / deformable interface / particle phase is described by the Lagrange method (following the motion of material clusters / particles, focusing on the change of physical quantities of individuals). The mass, momentum, and energy exchange between the two phases / two fields are realized through coupling conditions, balancing computational efficiency and motion tracking accuracy.

[0034] The physical and chemical parameters and boundary conditions of the CFD-DPM numerical simulation model are set to obtain the numerical simulation solution results. In this embodiment, the parameters of the CFD-DPM numerical simulation model are set, including the two-phase flow material properties of oil and abrasive particles, boundary conditions and discrete phase parameters, etc. By setting the physical and chemical parameters of the main driving oil and abrasive particles, as well as the outlet boundary conditions, the solution results of the CFD-DPM numerical simulation model are obtained.

[0035] Based on the numerical simulation results, cross-sections are set on the main pipeline and bypass, and the initial abrasive particle proportion data are obtained using these cross-sections as flow surfaces; based on the actual operating conditions of the main pipeline, a... Experimental scheme. In this embodiment, cross-sections are set in the main pipeline and bypass. By statistically analyzing the number of abrasive particles of different diameters flowing through the cross-sections, the initial proportion data of abrasive particles of different diameters in the main pipeline and bypass are obtained, such as... Figure 8 As shown, Latin hypercube is used for experimental design to generate multiple CFD-DPM numerical simulation experimental schemes. Generally, n experimental schemes are generated, and it is recommended that n be 20 or higher. Latin hypercube is a multidimensional, hierarchical, uniform, and non-overlapping experimental design method used to generate sample points uniformly and efficiently over multiple input variable values. It is particularly suitable for numerical simulations and surrogate modeling with limited sample size and the need to cover the entire operating range.

[0036] A predictive model is constructed using an optimal surrogate model set. Measured abrasive particle distribution data from the bypass is used as input, and predicted abrasive particle distribution data for the main pipeline is output. The mapping relationship between the input and output parameters is fitted using the surrogate model, enabling rapid and accurate prediction of the abrasive particle proportion in the main pipeline under different operating conditions. Based on the predicted abrasive particle distribution data for the main pipeline, the initial abrasive particle proportion data is corrected by combining simulated abrasive particle proportion data, thereby improving the accuracy and reliability of abrasive particle monitoring.

[0037] By constructing a CFD-DPM numerical simulation model of the main drive hydraulic pipeline 2 and the bypass of the TBM, the mapping relationship of the proportion of abrasive particles of different diameters in the main pipeline and the bypass is obtained. Based on the surrogate model and the correction of the abrasive particle data of the tunnel boring machine hydraulic system using CFD-DPM, the hydraulic monitoring data of the tunnel boring machine can be accurately corrected, providing data support for the evaluation of the main drive operation status of the tunnel boring machine, and improving the construction quality and intelligent level of tunnel boring machine tunneling.

[0038] Furthermore, the extraction process of the main drive pipeline and bypass watershed includes: Boolean operations are used to remove minute structures in the main drive oil pipeline and bypass to obtain the initial main pipeline and initial bypass. Screening conditions are then constructed to screen the initial main pipeline and initial bypass. In this embodiment, when extracting the flow domain of the main drive oil pipeline and bypass, the geometric model of the main drive oil pipeline and the bypass where the oil wear monitoring device 3 is located is simplified. The flow domain is extracted through Boolean operations to remove minute structures and obtain the initial main pipeline and initial bypass.

[0039] The screening criteria are as follows: the screening process retains key flow channel regions that significantly affect the distribution of the oil flow field and the transport characteristics of abrasive particles, and the lengths of the main pipeline and the bypass pipeline are equal to their respective diameters. times, In this embodiment, the screening process retains key flow channel regions that significantly affect the distribution of the oil flow field and the transport characteristics of abrasive particles. It should be noted that because the bypass length of the main pipeline is much larger than the bypass diameter, it can lead to singular mesh generation and unstable CFD-DPM model calculations. The main pipeline and bypass flow domains of the main driving oil are further simplified, retaining only the abrasive particles and oil inlet portions. The resulting flow domain is as follows: Figure 5 As shown. The retained portion of the flow area is determined based on the geometry of the main and bypass pipelines of the main driving fluid. Generally, to ensure the accuracy of the analysis results, the length of the main pipeline and the bypass pipeline are 1 / 3 of their respective diameters. times, Based on the selection criteria, the initial main pipeline and the initial bypass are selected.

[0040] Furthermore, a tetrahedral structured mesh is used for discretization, and a boundary layer mesh is added to the corresponding areas of the main pipeline and the bypass. Local mesh refinement is also performed in the connection area between the main pipeline and the bypass. In this embodiment, a tetrahedral unstructured mesh is used for discretization, and local refinement is performed in the connection part between the main pipeline and the bypass to improve the capture accuracy of the flow separation area and particle transport characteristics.

[0041] The main drive main pipeline - bypass basin grid satisfies the following: The grid passes the orthogonal quality inspection, and the minimum orthogonal quality is 0.2 or above; The boundary layer grid is used in the near-wall region, and the dimensionless wall distance y⁺ is controlled within the range of 30 to 300. In this embodiment, by inspecting the grid quality through orthogonal quality, a minimum orthogonal quality of 0.2 or above is considered qualified and meets the requirements of numerical simulation; Orthogonal Quality is one of the core indicators used to evaluate grid quality in numerical simulations such as Computational Fluid Dynamics (CFD). It measures the orthogonality between the normal vector of the grid cell face and the vector connecting the cell centroids, with a value range of 0 to 1, where 1 indicates the optimal grid quality and 0 indicates the worst grid quality. In the near-wall regions of the extracted main pipeline and bypass, the boundary layer grid is used for processing, and the dimensionless wall distance is controlled within the range of 30 to 300; In the extracted main pipeline and bypass, boundary layers are added, and the number of boundary layers is set to 5. The total number of grid cells in the computational domain is determined through grid independence verification to obtain the grid number when the computational structure tends to be stable, which is used as the final grid scheme. As Figure 6 shown, when the grid number is 68987, the calculation results tend to be stable, which can balance computational efficiency while ensuring computational accuracy, and is used as the final grid scheme.

[0042] Furthermore, the construction process of the CFD-DPM numerical simulation model is as follows: Based on the Euler–Lagrange method, a two-way coupling calculation model of the continuous phase of the oil fluid and the discrete phase of the abrasive particles is established, ignoring the collision effect between particles and the reverse influence of the volume fraction on the continuous phase, and only considering the momentum exchange between the particles and the fluid; The continuous phase of the oil fluid uses the equation model for turbulent flow calculation, selects the pressure-based solver, adopts the pressure–velocity coupling algorithm, and uses the pressure-based solver for numerical solution. The pressure–velocity coupling algorithm selects the Phase-Coupled SIMPLE method, and Lagrangian trajectory tracking of the discrete phase abrasive particles is performed on the basis of the steady-state flow field. The Phase-Coupled SIMPLE method is a numerical algorithm used to solve the pressure–velocity coupling in multiphase flow problems (especially Eulerian multiphase flow), and it is an extension of the classical SIMPLE (Semi-Implicit Method for Pressure-Linked Equations) algorithm in multiphase flow scenarios; The Lagrangian method is one of the methods to describe fluid motion. Taking the motion of a certain fluid particle as the research object, observing the change law of its motion parameters when this particle moves from one point to another in the flow field, and integrating the motions of numerous fluid particles to obtain the motion law of all fluid particles within a certain space.

[0043] Based on the actual physicochemical parameters of the main drive oil and abrasive particles of the tunnel boring machine, the material properties of the two-phase flow are set. The abrasive particles are set to spherical shape, and their particle size distribution is described using a distribution function. The particle size distribution is set according to the actual working conditions, thus completing the construction of the CFD-DPM numerical simulation model. In this embodiment, the material properties of the two-phase flow are set as follows: the density of the continuous phase oil is... Dynamic viscosity All parameters used are constant physical properties. The density of discrete phase abrasive particles is... The particle shape is assumed to be spherical, and the flow rate is set to... The particle size distribution is set according to the actual working conditions. Mesh, corresponding to a particle size range of The distribution is described using the Rosin–Rammler distribution function, a classic two-parameter model for describing particle / droplet size distribution. Particulate materials are defined as inert, and chemical reactions, fragmentation, and phase transitions are not considered.

[0044] Furthermore, the solution process of the CFD-DPM numerical simulation model is as follows: Set the main pipeline inlet as a velocity inlet, and apply the corresponding inlet velocity, turbulence intensity, and turbulence viscosity according to the operating conditions. The turbulence intensity is taken as... turbulent viscosity is The bypass inlet is connected to the main pipeline and inherits its flow parameters. Both the main pipeline outlet and the bypass outlet are set as pressure outlets, with the outlet pressure taken as standard atmospheric pressure. The pipe wall is set as a non-slip, stationary wall, and the particles and the wall are subjected to reflective boundary conditions. The elastic collision behavior between the abrasive particles and the wall is described by the Grant-Tabakoff coefficient of restitution model. The Grant-Tabakoff coefficient of restitution model (often referred to as the Grant model) is the most classic and widely used particle-wall collision coefficient of restitution model in CFD discrete phase (DPM) simulation. It is specifically used to describe the normal / tangential velocity loss when particles impact the wall at different angles.

[0045] A surface jet source is set at the inlet section of the main pipeline. Particles are randomly generated at the inlet section, and the particle size follows the actual working condition distribution. The initial jet velocity matches the direction and magnitude of the local flow velocity in the continuous phase to ensure that the particles enter the computational domain with the flow. The particle mass flow rate is set according to the wear working conditions.

[0046] Particle trajectory tracking is performed using the Lagrangian method. The maximum number of tracking steps, time step size, and coupling iteration interval are determined based on the extent of the main drive pipeline and the bypass watershed. In this embodiment, the Lagrangian method is used for particle trajectory tracking, with the maximum number of tracking steps set to 300 and the time step size set to [missing information]. The coupling iteration interval between the continuous phase and the discrete phase is set to 10 steps to ensure the stability and accuracy of the particle trajectory calculation; the particle is set to escape at the exit boundary and to reflect when entering the wall region.

[0047] The momentum equation, turbulent kinetic energy, and dissipation rate are expressed using a first-order upwind scheme. Residual convergence criteria for the continuous oil phase and the discrete abrasive phase are set, while the mass flow rate and particle count at the main pipeline and bypass sections are monitored. Convergence is determined when the fluctuations in mass flow rate and particle count are less than preset values, yielding the numerical calculation results and iterative curves of the CFD-DPM model. In this example, the momentum equation, turbulent kinetic energy, and dissipation rate are all expressed using a first-order upwind scheme, and the pressure term uses the PRESTO! format. The residual convergence criteria for the continuous oil phase and the discrete abrasive phase are respectively... (Continuous phase) and (Discrete phase), simultaneously monitoring the mass flow rate and particle count at the main pipeline and bypass cross-sections; when the fluctuation range of mass flow rate and particle count is less than When convergence is determined, the numerical calculation results and iteration curves of the CFD-DPM model are obtained, such as... Figure 7 (b) and Figure 7 As shown in (a) of the diagram.

[0048] Furthermore, both the initial abrasive particle proportion data and the simulated abrasive particle proportion data are two-dimensional data, and both include: Abrasive particle ratio data for main and bypass circuits corresponding to diameter sub-intervals, abrasive particle distribution data for main circuits and bypass circuits across the entire diameter interval; The diameter sub-interval is a sub-interval obtained by discretizing the full diameter interval, and the full diameter interval is a complete interval consisting of the minimum diameter to the maximum diameter of the abrasive particles of the main drive oil of the tunnel boring machine.

[0049] Furthermore, the process for determining the full diameter range is as follows: Based on the initial abrasive particle proportion data, the minimum and maximum diameters of the main drive oil abrasive particles are determined, and the range from the minimum to the maximum diameter is defined as the full diameter range of the main drive oil abrasive particles. In this embodiment, as... Figure 9 As shown, based on the operating conditions of the tunnel boring machine's main drive hydraulic system, the maximum hydraulic flow velocity can be determined by referring to the main drive hydraulic lubrication design parameters. The minimum value is The maximum abrasive flow rate is The minimum value is The upper bound of the experimental design scheme is then... The lower bound is .

[0050] Correspondingly, the diameter sub-intervals are several abrasive grain diameter intervals obtained by discretizing the full diameter interval in an equal or non-equal spacing manner, and the number of diameter sub-intervals obtained is not less than 15. In this embodiment, for the main drive oil abrasive grains with a minimum diameter of 20um and a maximum diameter of 200um, the abrasive grain diameter is divided into 18 intervals: 20-30um, 31-40um, 41-50um, 51-60um, 61-70um, 71-80um, 81-90um, 91-100um, 101-110um, 111-120um, 121-130um, 131-140um, 141-150um, 151-160um, 161-170um, 171-180um, 181-190um, and 191-200um. The abrasive grain proportion of each diameter interval is statistically analyzed to form a corresponding dataset.

[0051] Furthermore, the process for determining the optimal agent model set is as follows: A set of alternative surrogate models is preset, which includes at least the following: multinomial regression model (PRS), radial basis network model (RBF), kriging model (KRG), and support vector regression model (SVR). These four are the four most mainstream types of surrogate models, specifically used to fit a fast and approximate function that can replace simulation from multi-source simulation data / experimental data for prediction, optimization, sensitivity analysis, and uncertainty analysis. Multinomial Regression (PRS) is a classic surrogate model based on a global multinomial function, fitting the functional relationship between input variables and output response using the least squares method. Radial Basis Network (RBF) is a local interpolation surrogate model using radially symmetric functions as basis functions, achieving a nonlinear mapping between input and output through the linear weighted superposition of multiple local basis functions. Kriging (KRG) is an interpolation surrogate model based on spatial correlation, constructing an input-output mapping relationship through a combination of a global trend term and a local stochastic process term. Support Vector Regression (SVR) is a regression model based on statistical learning theory and structural risk minimization, mapping low-dimensional nonlinear problems to a high-dimensional feature space using kernel functions to construct the optimal regression function.

[0052] By evaluating the accuracy of the models within the candidate proxy model set, the accuracy index values ​​for each abrasive grain diameter sub-range are obtained. ; The closer the value is to 1.00, the higher the accuracy of the model.

[0053] For each abrasive grain diameter sub-range, select the accuracy index value. The highest-performing model is taken as the optimal surrogate model for that sub-interval, and the set of optimal surrogate models is the prediction model. In this embodiment, for each diameter interval dataset, the dataset is randomly divided into 5 parts, and one part is selected as the test data in turn, while the other data are used as the training data. The optimal surrogate model is selected through multi-fold cross-validation.

[0054] Among them, the model accuracy index value The calculation formula is as follows: , In the formula, Indicates the first Group experiment in the first Measured values ​​for each diameter sub-interval; , , m The number of diameter subintervals; Models within the set of candidate proxy models Regarding the first Group experiment in the first Estimates for each diameter sub-interval; This is the average of the estimated values ​​of all models within the candidate proxy model set.

[0055] Taking 61-70µm as an example, the accuracy of different surrogate models is as follows: Figure 10 As shown, the Kriging method model has the highest accuracy and is therefore used as the final proxy model. Similarly, the proxy models for the proportion of abrasive particles in the main pipeline for different diameter ranges are shown in Table 1.

[0056] Table 1. Selection Results of CFD-DPM Agent Models for Main Driveway and Bypass Circuits Furthermore, the correction process for the initial abrasive particle ratio data is as follows: The abrasive distribution data of different diameter sub-intervals of the bypass measured by the abrasive sensor, the actual flow velocity of the main drive oil pipeline of the tunnel boring machine, and the actual abrasive flow rate are input into the optimal surrogate model corresponding to each diameter sub-interval, and the abrasive distribution prediction data of different diameter sub-intervals of the main pipeline are output. Using the predicted abrasive particle distribution data as a basis for correction, the initial abrasive particle proportion data obtained from the numerical simulation solution is corrected to obtain the actual data of abrasive particle proportion in different diameter sub-intervals of the main pipeline, thus completing the correction of the initial abrasive particle proportion data.

[0057] In this embodiment, based on the abrasive monitoring data of the tunnel boring machine, after correction using the established optimal surrogate model, the main drive pipeline and bypass oil of the tunnel boring machine are extracted, and the actual abrasive diameter distribution is examined under a microscope. The results before and after correction are as follows: Figure 11As shown, it can be seen that the method proposed in this embodiment significantly reduces the measurement error of the proportion of abrasive grains with different diameters, proving the innovation and feasibility of the present invention.

[0058] Example 2 This embodiment provides a tunneling machine oil abrasive data correction system based on multi-source simulation data, used to implement the tunneling machine oil abrasive data correction method based on multi-source simulation data described in Embodiment 1, such as... Figure 12 As shown, it includes: The first module is configured to extract the main drive pipeline and bypass flow domain based on the geometry of the main drive pipeline of the tunneling machine and the bypass where the oil wear monitoring device 3 is located, and to discretize the main drive pipeline and bypass flow domain to obtain the main drive pipeline-bypass flow domain grid.

[0059] Tetrahedral unstructured meshes were used for discretization. Boundary layers were added to the main pipeline and bypass, and the local mesh was refined in the connection area between the main pipeline and bypass to obtain the watershed mesh of the tunnel boring machine's main drive main pipeline and bypass.

[0060] The second module is configured to perform bidirectional coupled calculations on the momentum exchange process between the continuous oil phase and the discrete abrasive phase based on the main drive pipeline-bypass watershed grid, and construct a CFD-DPM numerical simulation model; set the physicochemical parameters and boundary conditions of the CFD-DPM numerical simulation model, and obtain the numerical simulation solution results.

[0061] It is used to solve the CFD-DPM model of the main drive pipeline and bypass of the tunnel boring machine, and to obtain data such as the distribution of abrasive particles of different sizes under ideal conditions.

[0062] The third module is configured to set cross-sections on the main pipeline and bypass based on the numerical simulation results, and obtain initial abrasive particle proportion data using these cross-sections as flow surfaces; based on the actual operating conditions of the main pipeline, it forms... Experimental schemes were developed; the CFD-DPM numerical simulation model was used to generate abrasive particle simulation ratio data for each experimental scheme.

[0063] By setting the directional cross-sections of the main pipeline and the bypass pipeline, the number of abrasive particles of different diameters flowing through the cross-sections is counted, and the proportion and distribution of abrasive particle diameters in the main pipeline and the bypass pipeline are obtained.

[0064] The fourth module is configured to construct a prediction model using the optimal surrogate model set, taking the measured abrasive particle distribution data of the bypass as input and outputting the predicted abrasive particle distribution data of the main pipeline; based on the predicted abrasive particle distribution data of the main pipeline, the initial abrasive particle proportion data is corrected by combining the simulated abrasive particle proportion data.

[0065] Based on the surrogate model and CFD-DPM, the oil abrasive data correction of tunnel boring machines can accurately correct the oil monitoring data of tunnel boring machines, provide data support for the evaluation of the main drive operation status of tunnel boring machines, and improve the construction quality and intelligent level of tunnel boring machine tunneling.

[0066] This invention is not limited to the specific technical solutions described in the above embodiments. Besides the above embodiments, this invention may have other implementation methods. For those skilled in the art, any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this invention should be included within the protection scope of this invention.

Claims

1. A method for modifying oil abrasive particle data of a heading machine based on multi-source simulation data, characterized in that, The method comprises the following steps: According to the geometry of the main drive oil main pipeline and the bypass where the oil particle monitoring device (3) is located, the main drive main pipeline and the bypass basin are extracted and discretized to obtain the main drive main pipeline-bypass basin grid; Based on the main drive main pipeline-bypass basin grid, the momentum exchange process between the oil continuous phase and the particle dispersed phase is calculated bidirectionally to construct a CFD-DPM numerical simulation model; the physical and chemical parameters and boundary conditions of the CFD-DPM numerical simulation model are set to obtain the numerical simulation solution; According to the numerical simulation solving result, a section is arranged on the main pipeline and the bypass, initial proportion data of abrasive particles are obtained with the section as a flow surface; according to the actual operation condition of the main pipeline, a group experiment scheme is formed The CFD-DPM numerical simulation model is used to generate the simulation proportion data of abrasive particles corresponding to each group experiment scheme. A prediction model is constructed by means of an optimal proxy model set, the bypass measured particle distribution data are taken as input, and the main pipeline predicted particle distribution data are output; based on the main pipeline predicted particle distribution data, the initial particle proportion data are corrected in combination with the particle simulation proportion data.

2. The method according to claim 1, wherein, The extraction process of the main drive main pipeline and the bypass basin comprises: The initial main pipeline and the initial bypass are obtained by removing the small structures in the main drive oil main pipeline and the bypass by means of Boolean operation, and the screening conditions are constructed to screen the initial main pipeline and the initial bypass; The screening condition is that the screening process retains a key flow channel area which has a significant influence on the oil flow field distribution and abrasive particle transport characteristics, and the length of the main pipeline and the length of the bypass are each times the corresponding diameter.

3. The method of claim 1, wherein, Tetrahedral structure grid is adopted for discretization, boundary layer grid is added to the areas corresponding to the main pipeline and the bypass, and local grid is encrypted in the connection area of the main pipeline and the bypass; The main drive main pipeline-bypass basin grid satisfies: The grid passes the orthogonal quality inspection, and the minimum orthogonal quality is 0.2 or higher; The boundary layer grid is adopted in the near-wall area, and the wall surface non-dimensional distance y+ is controlled within the range of 30-300.

4. The method of claim 1, wherein, The construction process of the CFD-DPM numerical simulation model comprises: Based on Euler-Lagrange method, a two-way coupling model of oil continuous phase and abrasive particles dispersed phase was established. The oil continuous phase was calculated by Reynolds-Averaged Navier-Stokes (RANS) model, and the dispersed phase was calculated by Lagrangian method. The pressure-based solver and pressure-velocity coupling algorithm were used to calculate the turbulent flow of oil continuous phase. The Lagrangian trajectory tracking of dispersed phase was based on the steady flow field. The two-phase flow material properties are set according to the actual physical and chemical parameters of the tunnel boring machine main drive oil and particles, the particles are spherical, the particle size distribution is described by using a distribution function, and the CFD-DPM numerical simulation model is constructed.

5. The method of claim 1, wherein, The solving process of the CFD-DPM numerical simulation model comprises: The main pipeline inlet is set as a velocity inlet, the corresponding inlet flow rate, turbulence intensity and turbulence viscosity are applied according to the working condition, the bypass inlet is communicated with the main pipeline and inherits the flow parameters, the main pipeline outlet and the bypass outlet are both set as pressure outlets, the pipeline wall is set as a non-slip static wall, the particle and the wall adopt a reflection boundary condition, the elastic collision behavior of the particle and the wall is described by using a Grant restitution coefficient model; A face injection source is set at the main pipeline inlet cross section, the particles are randomly generated at the inlet cross section, the particle size obeys the actual working condition distribution, and the injection initial velocity matches the direction and size of the local flow rate of the continuous phase; The Lagrangian method is adopted for particle trajectory tracking, and the maximum tracking step number, time step and coupling iteration interval are determined according to the range of the main drive main pipeline and the bypass basin; The momentum equation, turbulent kinetic energy and dissipation rate adopt a first-order upwind scheme, the residual convergence standard value of the oil continuous phase and the particle dispersed phase is set, and the mass flow rate and the particle number of the main pipeline and the bypass cross section are monitored; when the fluctuation amplitude of the mass flow rate and the particle number is less than a preset value, it is determined that the convergence is achieved, and the CFD-DPM model numerical calculation result and the iteration curve are obtained.

6. The method of claim 1, wherein, Both the initial abrasive particle proportion data and the simulated abrasive particle proportion data are two-dimensional data, and both include: Abrasive particle ratio data for main and bypass circuits corresponding to diameter sub-intervals, abrasive particle distribution data for main circuits and bypass circuits across the entire diameter interval; The diameter sub-interval is a sub-interval obtained by discretizing the full diameter interval, and the full diameter interval is a complete interval consisting of the minimum diameter to the maximum diameter of the abrasive particles of the main drive oil of the tunnel boring machine.

7. The method of claim 6, wherein, The process for determining the full diameter range is as follows: Based on the initial proportion data of abrasive particles, the minimum and maximum diameters of the abrasive particles in the main drive oil are determined, and the range from the minimum to the maximum diameter is defined as the full diameter range of the abrasive particles in the main drive oil. Correspondingly, the diameter sub-interval is a number of abrasive grain diameter intervals obtained by discretizing the full diameter interval in an equal or non-equal interval manner, and the number of diameter sub-intervals obtained is not less than 15.

8. The method of claim 6, wherein, The process for determining the optimal proxy model set is as follows: A set of candidate proxy models is preset, which includes at least: multinomial regression model, radial basis network model, kriging model and support vector regression model; Through the precision evaluation of the models in the set of alternative agent models, the precision index values of each sub-interval of abrasive grain diameter are obtained ; Select precision index value for each abrasive grain diameter subinterval The highest model is the optimal proxy model corresponding to the subinterval, and the optimal proxy model set obtained by collection is the prediction model; The model precision index value The calculation formula of the model precision index value is as follows: ， In the formula, represents the first group of experiments in the first diameter sub-interval of the measured value; , , m is the number of diameter sub-intervals; models within the alternative set of agent models for the first group of experiments in the first estimated values for the first diameter subinterval; is the average of the estimated values for all models within the alternative set of proxy models.

9. The method of claim 1, wherein, The correction process for the initial abrasive particle ratio data is as follows: The abrasive distribution data of different diameter sub-intervals of the bypass measured by the abrasive sensor, the actual flow velocity of the main drive oil pipeline of the tunnel boring machine, and the actual abrasive flow rate are input into the optimal surrogate model corresponding to each diameter sub-interval, and the abrasive distribution prediction data of different diameter sub-intervals of the main pipeline are output. Using the predicted abrasive particle distribution data as a basis for correction, the initial abrasive particle proportion data obtained from the numerical simulation solution is corrected to obtain the actual data of abrasive particle proportion in different diameter sub-intervals of the main pipeline, thus completing the correction of the initial abrasive particle proportion data.

10. A multi-source simulation data based oil abrasive particle data correction system for roadheader, configured to implement the multi-source simulation data based oil abrasive particle data correction method for roadheader according to any one of claims 1 to 9, characterized in that, include: The first module is set to extract the main drive pipeline and bypass flow domain according to the geometry of the main drive pipeline of the tunneling machine and the bypass where the oil wear monitoring device (3) is located, and to discretize the main drive pipeline and bypass flow domain to obtain the main drive pipeline-bypass flow domain grid. The second module is configured to perform bidirectional coupled calculations on the momentum exchange process between the continuous oil phase and the discrete abrasive phase based on the main drive pipeline-bypass watershed grid, and construct a CFD-DPM numerical simulation model; set the physicochemical parameters and boundary conditions of the CFD-DPM numerical simulation model, and obtain the numerical simulation solution results. The third module is configured to set cross sections on the main pipeline and the bypass based on the numerical simulation results, and obtain the initial proportion data of abrasive particles using the cross sections as the flow surface. According to the actual operation condition of the main pipeline, the formation Group experiment scheme; using the CFD-DPM numerical simulation model to generate the simulation proportion data of abrasive particles corresponding to each group of experiment schemes; The fourth module is configured to construct a prediction model using the optimal surrogate model set, taking the measured abrasive particle distribution data of the bypass as input and outputting the predicted abrasive particle distribution data of the main pipeline; based on the predicted abrasive particle distribution data of the main pipeline, the initial abrasive particle proportion data is corrected by combining the simulated abrasive particle proportion data.

Images