A control system and method for multi-stage pressurized grouting

Abstract

This invention discloses a control system and method for multi-stage pressurized grouting, relating to the field of grouting fire prevention and extinguishing technology. The method includes: acquiring data collected by various sensors during the trial delivery phase, identifying the current working condition scenario, and obtaining baseline control parameters; acquiring current multi-sensor data and next-level associated data, and preprocessing them to obtain current multi-source feature data; obtaining dynamically adjusted control parameters based on the current multi-source feature data and the baseline control parameters; obtaining the original frequency modulation command for the first closed loop and the original valve adjustment command for the second closed loop based on the dynamically adjusted control parameters; and performing coupling analysis and dynamic weight arbitration on the original frequency modulation command and the original valve adjustment command to obtain the final coordinated control command. This invention achieves parameter feedforward adaptive compensation through multi-source features, dynamically adjusts control parameters, and resolves control conflicts using dynamic weight arbitration, exhibiting significant advantages such as rapid response, stable operation, and strong adaptability.

Description

Technical Field

[0001] This invention belongs to the field of grouting fireproofing technology, and specifically provides a control system and method for multi-stage pressurized grouting. Background Technology

[0002] In coal mining, shallow-buried, closely spaced spontaneously combusting coal seams pose a serious risk of spontaneous combustion in goaf areas. Grouting for fire prevention and extinguishing is a core technical means to ensure mining safety. Among these, long-distance underground grouting and pressurization operations are a critical link, requiring the accurate delivery of fire prevention and extinguishing grout to target areas in different types of goaf areas, such as those with long horizontal distances and vertical overburden, using pressurization devices.

[0003] However, the working range of a single pressurized grouting device is limited, making it difficult to cover complex and ever-changing grouting needs. In scenarios where multiple devices need to operate in series or in parallel, the lack of a unified flow and pressure coordination mechanism among the devices can easily lead to overall oscillations and efficiency decline, making it difficult to form a stable and efficient joint grouting capability.

[0004] In the long-distance grouting and pressurization process in coal mines, the core control objective of the pressurization device is to simultaneously stabilize the buffer tank level and outlet pressure to ensure continuous operation and conveying power. However, the existing dual-closed-loop control method has the following shortcomings: On the one hand, the dynamic characteristics of the slurry pumping device are heavily dependent on operating parameters. The slurry density directly affects fluid inertia and pumping efficiency, while pipeline resistance is time-varying and difficult to measure. Pure feedback regulation requires waiting for deviations to occur before starting regulation, which can easily lead to response lag, untimely regulation, or over-regulation. On the other hand, there is strong bidirectional coupling interference between level regulation and pressure regulation, which can easily trigger reverse regulation in the dual closed loop, causing oscillation of the slurry pumping device and weakening control stability. Summary of the Invention

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a control method for multi-stage pressurized grouting, applied to a multi-stage pressurized grouting system composed of multiple multi-stage pressurized grouting devices connected in series in sequence, wherein each stage device is defined as a stage workstation, including: Acquire data collected by various sensors during the trial delivery phase, identify the current operating conditions, and obtain baseline control parameters; Collect multi-sensor data from this level workstation and associated data from the next level workstation, and obtain the current multi-source feature data after preprocessing; Based on the current multi-source characteristic data and baseline control parameters, the dynamically adjusted control parameters are obtained. The control parameters include the liquid level loop proportional gain, the pressure loop adjustment coefficient, and the disturbance state judgment result. Based on the dynamically adjusted control parameters, the original frequency modulation command for the first closed loop and the original valve control command for the second closed loop are obtained. The original frequency modulation command and the original valve control command are coupled and dynamically weighted to obtain the final coordinated control command.

[0006] Furthermore, based on the current multi-source feature data and baseline control parameters, dynamically adjusted control parameters are obtained, including: Based on the current multi-source feature data, the inertial change of the slurry is predicted by feedforward, and the proportional gain of the liquid level loop is adaptively adjusted. Joint analysis of pressure fluctuation characteristics in current multi-source feature data is performed to generate disturbance state determination results; Based on the current multi-source characteristic data and benchmark control parameters, the valve status and slurry resistance status are evaluated to obtain the pressure loop adjustment coefficient.

[0007] Furthermore, based on the current multi-source feature data, feedforward prediction of slurry inertial changes is performed, and adaptive adjustment is used to obtain the liquid level loop proportional gain, including: Based on the density deviation in current multi-source feature data and historical multi-source feature data, a density deviation sequence is constructed and the density change trend is calculated. Based on the density change trend, the inertial prediction factor is obtained; The gain correction amount is obtained by querying the preset mapping table based on the inertial prediction factor; The current proportional gain of the level loop is calculated based on the gain correction amount and the reference proportional gain of the level loop in the reference control parameters.

[0008] Furthermore, based on the current multi-source characteristic data and baseline control parameters, the valve status and slurry resistance status are evaluated to obtain the pressure loop adjustment coefficient, including: Based on the valve opening feedback from the current multi-source feature data, and combined with the valve nonlinearity interval division threshold of the benchmark control parameters, the current valve opening interval is obtained. Based on the slurry density deviation in the current multi-source feature data, the slurry resistance change state is obtained; Based on the valve's current opening range, the slurry resistance change status, and the pressure loop reference adjustment coefficient in the reference control parameters, the current pressure loop adjustment coefficient is obtained.

[0009] Furthermore, based on the dynamically adjusted control parameters, the original frequency modulation command for the first closed loop and the original valve control command for the second closed loop are obtained, including: The first closed loop is a closed loop with the core objective of stabilizing the liquid level in the buffer tank, and its output is the original frequency modulation command for adjusting the speed of the slurry pump. The second closed loop is a closed loop with the core objective of stabilizing the pipeline outlet pressure, and its output is the original valve control command to adjust the valve opening.

[0010] Furthermore, obtaining the original frequency modulation command for the first closed loop includes: Based on the current proportional gain of the level loop in the dynamically adjusted control parameters, and combined with the level deviation between the buffer tank level and the set level value, the proportional term output is calculated. Based on the liquid level deviation, the anti-saturation integral term is obtained; The original frequency modulation command for the first closed loop is obtained by superimposing the proportional term output with the integral term.

[0011] Furthermore, obtaining the original control valve command for the second closed loop includes: Obtain the disturbance state determination result in the dynamically adjusted control parameters. If the determination result is an instantaneous disturbance, the original valve control command is 0. If the determination result is the actual pressure deviation, then the current adjustment coefficient of the pressure loop in the dynamically adjusted control parameters is used to calculate the value of the original valve control command.

[0012] Furthermore, coupling analysis and dynamic weight arbitration are performed on the original frequency modulation command and the original valve control command to obtain the final coordinated control command, including: Based on the original frequency modulation command and the original valve control command, a strong coupling conflict flag is obtained; Dynamic weight arbitration is performed based on the strong coupling conflict flag to obtain the liquid level weight factor and pressure weight factor. Based on the original frequency modulation command, the original valve control command, the liquid level weighting factor, and the pressure weighting factor, the final coordinated control command is obtained.

[0013] Furthermore, multi-sensor data from this level of workstation and associated data from the next level of workstation are collected and preprocessed to obtain current multi-source feature data, including: Based on the multi-sensor data collected from the current workstation and the associated data from the next workstation during the continuous delivery phase, effective verification and bad value removal are performed. Multidimensional feature extraction was performed on the processed multi-sensor data and associated data to obtain pipeline outlet pressure fluctuation characteristics, slurry state characteristics, pipeline resistance characteristics, and upper and lower level matching characteristics. By packaging the pipeline outlet pressure fluctuation characteristics, slurry state characteristics, pipeline resistance characteristics, and upper and lower level matching characteristics, the current multi-source characteristic data is obtained.

[0014] A control system for multi-stage pressurized grouting, used to implement a control method for multi-stage pressurized grouting, is deployed in each workstation of a multi-stage pressurized grouting system consisting of multiple multi-stage pressurized grouting devices connected in series, comprising: Initialize the scene recognition module: used to acquire data collected by various sensors during the trial delivery phase, identify the current working condition scene, and obtain the baseline control parameters; Data feature extraction module: used to collect multi-sensor data from the current workstation and related data from the next level workstation, and obtain the current multi-source feature data after preprocessing; Dynamic control parameter adjustment module: used to obtain dynamically adjusted control parameters based on current multi-source characteristic data and baseline control parameters. The control parameters include liquid level loop proportional gain, pressure loop adjustment coefficient and disturbance state judgment result; Original instruction generation module: used to obtain the original frequency modulation instruction for the first closed loop and the original valve control instruction for the second closed loop after dynamic adjustment of control parameters. Coupling and Coordination Command Generation Module: This module performs coupling analysis and dynamic weight arbitration on the original frequency modulation command and the original valve control command to obtain the final coordinated control command.

[0015] The beneficial effects of using this invention are: By identifying the overall operating conditions during the trial delivery phase to provide a control benchmark, and then using multi-source feature data that integrates the current and next-level states, the system can predict changes in slurry inertia and identify disturbance types. This drives feedforward adaptive compensation of control parameters. When dealing with disturbances during operation, the system resolves conflicts between the liquid level and pressure control loops through dynamic weight arbitration, generating the final coordinated control command. This system can adapt to complex operating conditions such as changes in slurry density, valve nonlinearity, and time-varying pipeline resistance, significantly improving the continuity, operational stability, and overall reliability of grouting operations. This solves the problems of response lag, device oscillation, and control instability caused by the strong nonlinearity and strong coupling of the dynamic characteristics of pressurized grouting equipment in existing technologies. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the steps of a control method for multi-stage pressurized grouting according to the present invention. Figure 2 This is a flowchart illustrating the steps of dynamically adjusting the proportional gain of the liquid level loop in this invention. Figure 3 This is a flowchart illustrating the steps of dynamically adjusting the pressure ring adjustment coefficient according to the present invention. Figure 4 This is a flowchart illustrating the steps of coupling analysis and dynamic weight arbitration in this invention. Detailed Implementation

[0017] The present invention will now be described in detail with reference to the accompanying drawings.

[0018] Example 1

[0019] Reference Figures 1-4 A control system for multi-stage pressurized grouting, wherein the multi-stage pressurized grouting equipment consists of multiple pressurized grouting devices connected in series to form a multi-stage pressurized grouting system; The multi-stage pressurized grouting system includes an initialization scene recognition module, a data feature extraction module, a dynamic control parameter adjustment module, and an instruction generation module; Among them, the initialization scene recognition module is used to acquire data collected by each sensor during the trial delivery phase, identify the current working condition scene, and obtain the baseline control parameters; The data feature extraction module is used to collect multi-sensor data from the current workstation and related data from the next-level workstation, and obtains the current multi-source feature data after preprocessing. The dynamic control parameter adjustment module is used to obtain dynamically adjusted control parameters based on the current multi-source characteristic data and the baseline control parameters. The control parameters include the liquid level loop proportional gain, the pressure loop adjustment coefficient, and the disturbance state judgment result. The instruction generation module includes a raw instruction generation module and a coupled collaborative instruction generation module; The original instruction generation module is used to obtain the original frequency modulation instruction for the first closed loop and the original valve control instruction for the second closed loop based on the dynamically adjusted control parameters. The coupling and coordination instruction generation module is used to perform coupling analysis and dynamic weight arbitration on the original frequency modulation instruction and the original valve control instruction to obtain the final coordinated control instruction.

[0020] Specifically, multiple pressurized grouting devices are deployed in series along the grout delivery pipeline, and the pressurized grouting devices are defined as workstations, that is, multi-level workstations are deployed continuously.

[0021] Example 2

[0022] A control method for multi-stage pressurized grouting, employing a control system for multi-stage pressurized grouting as described in Example 1, includes the following specific steps: Step S1: Acquire the data collected by each sensor during the trial delivery phase, identify the current working condition scenario, and obtain the baseline control parameters; Collect multi-sensor data from this level workstation and associated data from the next level workstation, and obtain the current multi-source feature data after preprocessing; Based on the current multi-source feature data and the baseline control parameters, the dynamically adjusted control parameters are obtained. The control parameters include the liquid level loop proportional gain, the pressure loop adjustment coefficient, and the disturbance state determination result. Based on the dynamically adjusted control parameters, the original frequency modulation command for the first closed loop and the original valve control command for the second closed loop are obtained. The original frequency modulation command and the original valve control command are coupled and dynamically weighted to obtain the final coordinated control command. Several multi-stage pressurized grouting units with identical functions are connected end-to-end in a series along the grout delivery pipeline. Each stage of the multi-stage pressurized grouting unit is called a primary workstation. This series structure allows the outlet of the upstream workstation to be directly connected to the inlet of the downstream workstation. The grout flows sequentially through each workstation and is finally delivered to the target area. Each primary workstation is a fully functional independent pressurization unit with complete grout receiving, temporary storage, pressurization, and output capabilities.

[0023] Each multi-stage pressurized grouting device is centered around an intelligent programmable control cabinet. It uses a battery flow meter to collect the input and output flow rates of the multi-stage pressurized grouting device, a pressure sensor to collect the inlet and outlet pressures of the pipeline, an immersion-type liquid level sensor to collect the liquid level of the buffer tank, and a slurry density sensor to collect the slurry density. The control commands are executed through a slurry pump, pipeline valves, intelligent regulating valves, and frequency converters.

[0024] Furthermore, it interacts with other multi-stage pressurized grouting devices via bus or wireless communication to form a multi-stage control network, enabling coordinated operation of multiple multi-stage pressurized grouting devices. This network supports state synchronization and coordinated operation among multiple multi-stage pressurized grouting devices, forming an integrated pressurized system that ensures stable pressure and flow during the grouting process under long-distance and complex working conditions.

[0025] By identifying global operating conditions through trial delivery and providing a collaborative benchmark, predicting slurry inertial changes based on multi-source features and identifying disturbance types, the system achieves feedforward adaptive compensation of control parameters. Under disturbance conditions, it utilizes dynamic weight arbitration to intelligently resolve conflicts between the liquid level and pressure control loops, obtaining the final collaborative control command. It can adapt to changes in slurry density, valve nonlinearity, and time-varying pipeline resistance, significantly improving the continuity, stability, and reliability of the grouting process in complex downhole environments. It solves the problems of response lag, device oscillation, and insufficient stability caused by the nonlinear dynamic characteristics and strong coupling of pressurized grouting equipment in existing technologies.

[0026] The steps for acquiring data from various sensors during the trial delivery phase, identifying the current operating condition, and obtaining baseline control parameters include:

[0027] The pressure-distance ratio was obtained based on the data collected by each sensor during the trial delivery phase. Identify the current working condition scenario based on the pressure-distance ratio; Based on the current operating conditions, the baseline control parameters are obtained.

[0028] In this embodiment, the trial delivery phase refers to the following: Before formal grouting, the current workstation receives a synchronization readiness command from the previous workstation, initiates its readiness settings, and issues a synchronization readiness command to the next lower workstation to confirm its readiness. A low-flow-rate trial run is then conducted, for example, delivering grout to the grouting pipeline at a flow rate of 10-20 cubic meters per hour. This ensures that the low flow rate avoids pressure shocks to the pipeline while accurately capturing the resistance characteristics of different scenarios. The 10-20 cubic meters per hour figure is an example; specific data can be determined based on actual scenario experiments.

[0029] Simultaneously collect the test flow rate and the test pipeline outlet pressure P test Key data for this level of workstation, such as the level of the buffer tank and the density of the slurry during trial delivery.

[0030] Based on the outlet pressure P of the test pipeline test The calculated pressure-distance ratio K=P is based on the pipeline transport distance L from the workstation at this level. test / L, analyze the fluctuation characteristics of flow rate and pressure based on K value; When K < the first scene threshold, it is determined to be a horizontal transmission scene; When K ≥ the second scene threshold, it is determined to be a vertical transmission scene; When the threshold of the first scenario ≤ K < the threshold of the second scenario, it is determined to be a fracture grouting scenario. The thresholds of the first and second scenarios are determined by experiments on the multi-stage pressurized grouting device under different working conditions, and this embodiment does not limit them.

[0031] Based on the identified scenario S, the corresponding baseline control parameters are retrieved from the parameter library, including: liquid level setpoint H. set Pressure target range [P] min ,P max ], liquid level loop reference proportional gain, pressure loop reference adjustment coefficient k p P0, valve nonlinearity interval division threshold, liquid level safe interval [H] low H high ]; where P min P is the minimum value of the pressure target. max H represents the maximum value of the pressure target. low H is the minimum value for safe liquid level. high The maximum safe level is set. Then, a start command is sent to activate the downhole slurry pump, open the main delivery pipeline gate, and simultaneously adjust the frequency converter to ensure the slurry flow rate steadily increases from the test flow rate to the target flow rate, officially entering the continuous slurry delivery phase.

[0032] By transmitting synchronization readiness commands at each level, it ensures that each workstation is in a controllable and ready state. Key data is collected during small-scale trial runs to identify working scenarios and automatically match and call scenario-based benchmark control parameters. This ensures that each workstation has a control benchmark that matches its actual working conditions before formal grouting, thereby significantly improving the stability of equipment startup, the consistency of response, and the reliability of multi-level collaboration. This lays a key technical foundation for subsequent continuous and stable grouting operations.

[0033] The steps include collecting multi-sensor data from the current workstation and associated data from the next-level workstation, and preprocessing it to obtain the current multi-source feature data. Based on the multi-sensor data collected from the current workstation and the associated data from the next workstation during the continuous delivery phase, effective verification and bad value removal are performed. Multidimensional feature extraction was performed on the processed multi-sensor data and associated data to obtain pipeline outlet pressure fluctuation characteristics, slurry state characteristics, pipeline resistance characteristics, and upper and lower level matching characteristics. By packaging the pipeline outlet pressure fluctuation characteristics, slurry state characteristics, pipeline resistance characteristics, and upper and lower level matching characteristics, the current multi-source characteristic data is obtained.

[0034] Entering the continuous slurry delivery phase, at the beginning of each control cycle, the controller of each workstation performs the following operations: Simultaneously collect multi-sensor data from this workstation, including slurry density ρ, buffer tank level H, and pipeline inlet pressure P. in With export pressure P out Input and output flow rate Q in and Q out And the feedback V of the regulating valve opening fb Inverter output frequency feedback value F fb Simultaneously, it acquires the next-level related data and obtains the pipeline outlet pressure P of the next-level workstation through a multi-level control network. out2 and output flow Q out2 .

[0035] All the data obtained is marked with the same timestamp to ensure the time consistency of the data.

[0036] The system performs valid item verification and bad value removal on the collected data. For example, it determines whether each data is within the physical reasonable range, interpolates and replaces abnormal data with the valid value of the previous control cycle, records sensor faults, and sends warnings.

[0037] Then, multi-dimensional features are calculated to extract the pipeline outlet pressure fluctuation characteristics. For a given short-term analysis window, which corresponds to N control cycle data, the maximum value P of the pipeline outlet pressure within that window is used to find the maximum value. out,maxand minimum value P out,min Calculate the absolute fluctuation range A abs =P out,max -P out,min Then calculate the relative fluctuation amplitude A. rel =A abs / P avg ; Among them, P avg This represents the average outlet pressure of the pipeline within this window.

[0038] The least squares method was used to perform linear fitting on the pipeline outlet pressure and time series within the window, and the first-order linear fitting slope S of the pipeline outlet pressure within the window was obtained. A Compare the slope of the linear fit with the trend threshold. If S A If the value is greater than the positive trend threshold, then the pressure change is an upward trend; If S A If the pressure change is less than the negative trend threshold, it is considered a downward trend; otherwise, it is considered a stable fluctuation. The trend threshold is a positive number, and the specific value can be determined by experimentation based on the actual application scenario. This embodiment does not limit this value.

[0039] Calculate the root mean square error R between the pipeline outlet pressure and its linear fitting line within this window. mse R mse A small value indicates a clear trend. mse A large value indicates drastic fluctuations and a weak trend.

[0040] Extract the slurry state characteristics and calculate the current density deviation Δρ = ρ - ρ ref ; In the formula, ρ ref The initial test slurry density or the slurry density of the previous control cycle.

[0041] Extract pipeline resistance characteristics and calculate the current flow deviation ΔQ=Q in -Q out It is used to trigger alarms for leakage or decreased pump efficiency, but does not participate in the control cycle. The transient pipeline resistance R = (P) is estimated. out -P in ) / (Q out 2 +ε); In the formula, Q out 2 The square of the output flow rate is used to reflect the relationship between the square of the flow velocity and the resistance. ε is a very small positive number used to prevent errors when it approaches 0. It can reflect the time-varying parameter of the current pipeline flow resistance. Increasing its value will lead to a higher outlet pressure required for the same flow rate.

[0042] Extract the matching features between the upper and lower levels, and calculate the matching flow deviation ΔQ between the output flow of the current workstation and the output flow of the next lower level workstation. match =Q out -Q out2 Calculate the matching pressure deviation ΔP between the outlet pressure of the pipeline at this level and the outlet pressure of the pipeline at the next level. match =P out -P out2 .

[0043] All extracted features are packaged into structured data to obtain the current multi-source feature data.

[0044] By synchronously collecting and fusing multi-sensor data from this workstation with related data from the next level, a multi-source feature data system was constructed, encompassing pressure fluctuations, slurry status, pipeline resistance, and matching relationships between upper and lower levels. This enables each control command to perceive its own operating conditions and grasp downstream pressure and flow status. Based on this multi-source feature data, comprehensive and consistent data support is provided for subsequent dynamic adjustment of control parameters, feedforward compensation, and collaborative arbitration, fundamentally enhancing the overall perception capability, response accuracy, and reliability of collaborative control under complex time-varying conditions.

[0045] The steps for obtaining dynamically adjusted control parameters based on current multi-source feature data and baseline control parameters include:

[0046] Based on the current multi-source feature data, the inertial change of the slurry is predicted by feedforward, and the proportional gain of the liquid level loop is adaptively adjusted. Joint analysis of pressure fluctuation characteristics in current multi-source feature data is performed to generate disturbance state determination results; Based on the current multi-source characteristic data and benchmark control parameters, the valve status and slurry resistance status are evaluated to obtain the pressure loop adjustment coefficient.

[0047] Dynamic adjustment of control parameters is the core of achieving adaptive control in the grouting process. Based on multi-source characteristic data and baseline control parameters, the control parameters of the level loop and pressure loop are collaboratively optimized through three processing paths to adapt to changes in grout density, valve nonlinearity, and time-varying pipeline resistance. Simultaneously, interference signals in the operating conditions are identified to avoid ineffective adjustments. In other words, through deep matching of current multi-source characteristic data with baseline control parameters, the current proportional gain of the level loop, the interference state, and the current adjustment coefficient of the pressure loop are dynamically calibrated. These three parameters work together to form control parameters adapted to real-time operating conditions, laying a solid adaptive foundation for generating accurate and stable final control commands.

[0048] The steps for obtaining the current proportional gain of the level loop based on current multi-source characteristic data and baseline control parameters include: Based on the density deviation in current multi-source feature data and historical multi-source feature data, a density deviation sequence is constructed and the density change trend is calculated. Based on the density change trend, the inertial prediction factor is obtained; The gain correction amount is obtained by querying the preset mapping table based on the inertial prediction factor; The current proportional gain of the level loop is calculated based on the gain correction amount and the reference proportional gain of the level loop in the reference control parameters.

[0049] like Figure 2 As shown, the density deviations in the current multi-source feature data and the historical multi-source feature data of the previous n control cycles are obtained to form a density deviation sequence {Δρ(t),Δρ(t-1),...,Δρ(tn)}, where t is the current control cycle; The density variation trend is quantified based on the density deviation sequence to obtain the long-term slope, which is obtained by performing linear regression on the density deviation sequence to obtain the slope S. long .

[0050] Then, a mapping relationship between density and inertia is established.

[0051] As is well known, an increase in slurry density means an increase in mass inertia and flow resistance. The increase in mass inertia means that a greater pressure is required for the same flow rate change. Therefore, the predicted trend of slurry density change will directly translate into a prediction of the control response speed requirement.

[0052] In this embodiment, an inertial prediction factor I is set. pred I pred =I base ×[1+k1×Δρ+k2×S long ]; I base As a reference inertia; Its corresponding slurry density ρ ref The inertia is determined by k1 and k2, which are weighting coefficients determined based on scenario experiments.

[0053] A mapping table between the inertial prediction factor and the gain correction was constructed. This mapping table was determined through simulation and field tests. When the inertial prediction factor is larger, the level loop needs a more agile response to counteract the inertial effect and avoid level runaway. Therefore, the proportional gain needs to be increased appropriately.

[0054] Several threshold values ​​are set to divide the inertial prediction factor intervals, corresponding to different gain correction amounts Δk. p ; When I pred <The first threshold indicates that the slurry becomes thinner and the inertia decreases. The corresponding gain correction is a negative correction, which reduces the gain to prevent overshoot. When the first partitioning threshold ≤ I pred <The second threshold indicates normal inertia, corresponding to zero correction, maintaining the reference proportional gain of the liquid level loop; When I pred ≥ The second threshold indicates that the slurry has thickened and its inertia has increased, which corresponds to a positive correction, increasing the gain to improve the response speed; The first division threshold, the second division threshold, and the different gain correction amounts corresponding to each interval are determined based on specific multi-stage pressurization equipment tests.

[0055] The calculated inertia prediction factor is used to obtain Δk by looking up the mapping table. p Finally, the current proportional gain k of the liquid level loop is obtained. p H=k p H0×(1+Δk p ).

[0056] By constructing a time-series sequence of density deviation and calculating its changing trend, an innovative online prediction and quantification of slurry flow inertia was achieved, and the proportional gain of the level loop was dynamically adjusted accordingly. The change in slurry density is transformed into a quantitative requirement for the response speed of the control system, adaptively determining the gain correction amount. This feedforward gain adjustment can enhance the control response in advance to stabilize the level when the slurry thickens and its inertia increases, and appropriately reduce the gain to prevent overshoot when the slurry thins. This significantly improves the adaptability and disturbance rejection of the level control loop to time-varying slurry conditions, fundamentally overcoming the inherent defects of traditional fixed-parameter controllers in responding to density changes with lag or instability.

[0057] The steps for obtaining the interference state determination result based on the current multi-source feature data include: Based on the characteristics of the pipeline outlet pressure fluctuation at this level of workstation, it is determined whether interference exists, and the interference status judgment result is obtained. A feature comprehensive judgment rule is pre-set, including: Rule 1: When A abs >A th1 、│S A │>S th1 R mse <R th1 If the fluctuation is large and the change is drastic, but the random fluctuation deviates from the fitted line, it indicates that this is a peak or pulse with a clear trend that occurs and ends quickly. Typical examples include water hammer and solid particles that block and then break through. Rule 2, when A abs <A th2 、│S A │<S th2 R mse >R th2If the fluctuation is small and has no obvious trend, but the random fluctuation is violent, it is determined to be instantaneous interference. This rule corresponds to small fluctuation amplitude and no obvious trend, but the random fluctuation is violent, indicating that it is high-frequency sensor noise or small turbulence in the fluid. Rule 3, when A th2 ≤A abs ≤A th1 S A The value is continuously positive or negative for more than 3 periods, and R mse <R th2 If the fluctuation is moderate, it is determined to be a true pressure deviation. This rule corresponds to a moderate fluctuation range with a clear and continuous upward or downward trend, and the fluctuation is relatively regular around the trend line. This indicates that the pressure drift is caused by a real and slow change in load, such as gradual blockage of pipelines or slow changes in upstream flow. Rule 4: When the fluctuation characteristics do not meet any of the above explicit rules, the system enters a delayed verification state and continues to observe the pressure changes for the next 3 control cycles. If the pressure automatically returns to the vicinity of the target range during this period, it is determined to be a short-term fluctuation and is shielded; if the pressure deviation persists or expands, it is determined to be a true deviation.

[0058] It should be noted that A th1 and A th2 These are the thresholds for large fluctuations and small fluctuations, S th1 and S th2 These are the thresholds for large slope changes and small slope changes, respectively, R th1 and R th2 These are the low random fluctuation threshold and the high random fluctuation threshold, respectively. Their specific values ​​are determined based on the specific multi-stage pressurized grouting equipment and on-site working condition tests. 3 is example data and is not limited in this embodiment.

[0059] Rules 1 through 3 are checked sequentially. If a match is found between rule 1 and rule 2, the shielding adjustment flag is determined to be true. If a match is found between rule 3 and rule 4, the shielding adjustment flag is determined to be false. If none of these match, the delayed verification process of rule 4 is initiated. Until verification is complete, the shielding adjustment flag state from the previous control cycle is temporarily maintained. Therefore, the interference state determination result is represented by the shielding adjustment flag. If the shielding adjustment flag is true, the corresponding interference state determination result is instantaneous interference; if the shielding adjustment flag is false, the corresponding interference state determination result is actual pressure deviation.

[0060] By setting up a multi-rule comprehensive judgment logic based on pressure fluctuation amplitude, trend slope, and random fluctuation, the system can identify and classify pipeline outlet pressure fluctuations, accurately distinguish between instantaneous interferences such as water hammer and momentary blockage and high-frequency noise such as sensor noise, and identify the actual pressure deviation caused by pipeline blockage and real load changes. This effectively avoids valve malfunctions and frequent oscillations caused by treating all types of fluctuations the same, significantly improves the accuracy and stability of pressure control, and enhances the anti-interference capability in complex fluid transportation processes.

[0061] The steps for obtaining the current adjustment coefficient of the pressure loop based on current multi-source characteristic data and baseline control parameters include: Based on the valve opening feedback from the current multi-source feature data, and combined with the valve nonlinearity interval division threshold of the benchmark control parameters, the current valve opening interval is obtained. Based on the slurry density deviation in the current multi-source feature data, the slurry resistance change state is obtained; Based on the valve's current opening range, the slurry resistance change status, and the pressure loop reference adjustment coefficient in the reference control parameters, the current pressure loop adjustment coefficient is obtained.

[0062] like Figure 3 As shown, based on the factory calibration and fine-tuning in actual scenarios, the valve's nonlinear range is divided into low-opening, medium-opening, and high-opening ranges. The low-opening range can be set to 0%-30%, where valve adjustment is flexible but flow is low; therefore, a large reference adjustment coefficient is needed to produce a significant flow change with small adjustments. The medium-opening range can be set to 30%-70%, where adjustment characteristics are stable, and a conventional reference adjustment coefficient k is used. p P0; The high opening range can be set to 70%-100%. In this range, the adjustment sensitivity decreases, and the reference adjustment coefficient decreases to prevent over-adjustment. Here, 30% and 70% are example data for the valve's non-linear range division thresholds; this embodiment does not limit these values. Based on the valve opening feedback V... fb Determine the current opening range of the valve.

[0063] Next, slurry resistance correction is performed, using density deviation as a variable for slurry resistance to obtain the slurry resistance change state. A standard resistance is set, such as defining the standard resistance as 0. If density deviation Δρ > 0, it means that the flow rate is smaller at the same opening degree, or that a higher pressure is required to achieve the same flow rate. In this case, within the same opening degree range, the adjustment coefficient should be appropriately increased to make the valve adjustment action more powerful in order to counteract the increased resistance. If density deviation Δρ < 0, the adjustment coefficient should be appropriately decreased to prevent over-adjustment.

[0064] Then, based on the current valve opening range and the change in slurry resistance, the current adjustment coefficient k of the pressure loop is obtained. pFor example, P constructs a correlation table of valve opening degree, slurry resistance, and adjustment parameters, combining three nonlinear opening intervals with three density deviations into nine modes. Each mode corresponds to an adjustment coefficient and a maximum single-step adjustment amount ΔV. max This adjustment coefficient is based on the benchmark adjustment coefficient k. p The floating adjustment is performed on P0, and the maximum single-step adjustment is a positive number. The specific adjustment coefficient and the maximum single-step adjustment need to be determined based on the test of the multi-stage pressurized grouting equipment. This embodiment does not limit them here.

[0065] Finally, by consulting the correlation table based on the current valve opening range and the change in slurry resistance, the corresponding adjustment coefficient can be obtained, thus yielding the current adjustment coefficient of the pressure loop.

[0066] By comprehensively considering the resistance changes reflected by the valve opening range and the slurry density deviation, the pressure loop adjustment coefficient is dynamically determined, effectively solving the constraints on control performance caused by the inherent nonlinearity of the valve and the time-varying operating conditions.

[0067] This embodiment matches differentiated adjustment parameters based on the different adjustment characteristics of the valve in the low, medium, and high opening ranges, and dynamically corrects the parameters by incorporating the influence of slurry density changes on flow resistance, ensuring that the valve's actuation force always matches the current operating conditions. This significantly improves the accuracy and stability of pressure control, avoiding control instability or response lag caused by abrupt changes in valve adjustment characteristics or slurry state changes, thus achieving robust and efficient pressure loop control in complex multi-stage grouting environments.

[0068] The steps for obtaining the original frequency modulation command for the first closed loop and the original valve control command for the second closed loop based on the dynamically adjusted control parameters include:

[0069] The first closed loop is a closed loop with the core objective of stabilizing the liquid level in the buffer tank, and its output is the original frequency modulation command for adjusting the speed of the slurry pump. The second closed loop is a closed loop with the core objective of stabilizing the pipeline outlet pressure, and its output is the original valve control command to adjust the valve opening.

[0070] The dynamically adapted control parameters are transformed into specific control commands for the actuators of the workstation at this level. The first closed loop calculates and generates the original frequency modulation command to adjust the slurry pump speed based on the dynamically adjusted proportional gain of the level loop and the level deviation. This command directly addresses the level regulation requirement, actively managing the output flow of this level by changing the pump frequency, which is the foundation for ensuring the stability of the buffer tank level and the continuity of the system's liquid supply. The second closed loop calculates and generates the original valve adjustment command to adjust the valve opening based on the dynamically adjusted pressure loop adjustment coefficient, the disturbance state judgment result, and the pressure deviation. This command directly addresses the pressure regulation requirement, adjusting the pipeline resistance of this level by changing the valve opening, which is the foundation for ensuring the stability of the outlet pressure and the delivery power. This step ensures the specificity of each closed-loop control and reserves adjustment space for the subsequent coupling and coordination of original commands, making it a key link connecting the dynamic optimization of control parameters and the final coordinated control.

[0071] The steps for obtaining the original frequency modulation command for the first closed loop include: Based on the current proportional gain of the level loop in the dynamically adjusted control parameters, and combined with the level deviation between the buffer tank level and the set level value, the proportional term output is calculated. Based on the liquid level deviation, the anti-saturation integral term is obtained; The original frequency modulation command for the first closed loop is obtained by superimposing the proportional term output with the integral term.

[0072] First, calculate and update the level deviation. This is based on the buffer tank level H and the level setpoint H0. set Calculate the liquid level deviation e in the current control cycle. H ,e H =H set -H+α1×ΔP match ; In the formula, α1 is the liquid level compensation coefficient, used to convert the flow matching requirement into the liquid level adjustment amount, that is, the liquid level setting correction value corresponding to the unit flow deviation, which is determined based on the test of the multi-stage pressurized grouting device; the liquid level deviation of the current control cycle is updated to the historical liquid level deviation sequence to prepare for the next cycle; Calculate the proportional term P of the dynamic gain pout P pout =k p H×e H The larger the density deviation, the greater the adjustment force, and the gain has already been feedforward optimized based on density prediction. Then, the anti-saturation integral term is calculated. An integral gain k is set. i H is usually a fixed value, or it can be finely adjusted according to the operating conditions, so that the integral increment ΔI=k can be calculated. i H×e H ×T; In the formula, T is the control period.

[0073] The anti-saturation logic is executed. This step prevents the integral term from becoming too large. Specifically, it checks whether the final output command of the previous control cycle has reached the physical limit of the actuator, i.e., whether it is saturated. If it is not saturated, the integral term is accumulated normally. sum (n)=I sum (n -1)+ΔI; In the formula, I sum (n) represents the current integration term; I sum (n - 1) represents the integral term of the previous control cycle; If the circuit is saturated and the deviation direction is the same as the saturation direction, then integration stops; if the circuit is saturated but the deviation direction is opposite to the saturation direction, then reverse integration is allowed. The initial frequency modulation command for the first closed loop is generated. The initial frequency modulation command Δf raw =P pout +I sum (n); Δf raw >0 indicates that the inverter output frequency needs to be increased; Δf raw <0 indicates that the inverter output frequency needs to be reduced.

[0074] By incorporating the matching flow deviation into the level deviation calculation in the form of a level compensation coefficient, source-level coordinated control of flow balance across multiple workstations is achieved. This method, while responding to level deviations at its own workstation, actively and dynamically corrects the control target based on the flow matching status with lower-level workstations. Combining a dynamic proportional gain optimized by slurry density prediction with an anti-saturation integral term, the method can quickly and smoothly generate the original frequency modulation command. This not only effectively stabilizes the level in the local buffer tank but also coordinates the flow supply relationship between multiple levels, thereby significantly improving the response speed, stability, and overall flow coordination capability of the multi-level workstations.

[0075] The steps for obtaining the original control valve command in the second closed loop include: Obtain the disturbance state determination result in the dynamically adjusted control parameters. If the determination result is an instantaneous disturbance, the original valve control command is 0. If the determination result is the actual pressure deviation, then the value of the original valve control command is calculated based on the current adjustment coefficient of the pressure loop in the dynamically adjusted control parameters.

[0076] Based on the shielding adjustment obtained from the interference determination, if the shielding adjustment is true, skip the calculation and directly set the original valve control command ΔV. raw =0; If the shielding adjustment is false, then continue with the subsequent pressure regulation calculation steps. First, calculate the current pressure deviation and the pressure control target value.

[0077] Calculate the target pressure control value P target P target =(P max +P min ) / 2+α2×ΔP match ; In the formula, α2 is the pressure compensation coefficient, which was determined through experiments.

[0078] Calculate the current pressure deviation e p e p =P target -P out ; If e p >0, the current pressure is lower than the target value, and the valve needs to be opened to increase the flow and pressure; If e p <0, the current pressure is higher than the target value, and the valve needs to be closed to reduce flow and pressure.

[0079] Calculate the original valve control command ΔV raw ΔV raw =k p P+e p This ensures that under different operating conditions and with different slurries, the same pressure deviation can produce a valve action of appropriate magnitude.

[0080] In addition, to prevent excessive instantaneous valve movement due to a sudden increase in pressure deviation, which could cause severe system oscillations or water hammer, the ΔV... raw Amplitude limiting is implemented to ensure that slightly larger movements are allowed in the valve high opening range or when slurry resistance is high; and the movement range is limited in the sensitive range or when resistance is low.

[0081] When ΔV raw >ΔV max Then, ΔV is redefined. raw =ΔV max ; When ΔV raw <-ΔV max Then, ΔV is redefined. raw =-ΔV max ; ΔV raw >0 indicates that the valve opening needs to be increased; ΔV raw <0 indicates that the valve opening needs to be reduced.

[0082] By incorporating the matching pressure deviation into the dynamic correction of the pressure target value and selectively implementing adjustments based on the interference judgment results, stable and robust operation of the pressure control loop is achieved. Instantaneous interference is effectively filtered out, preventing malfunctions of valves caused by invalid fluctuations such as water hammer and noise. When dealing with real pressure deviations, the pressure control loop's anti-interference capability under complex operating conditions is enhanced by integrating feedforward correction of downstream pressure and dynamically optimizing adjustment parameters based on valve opening and slurry resistance. This prevents over-adjustment and oscillation while ensuring a rapid and stable response to real load changes.

[0083] The steps for performing coupling analysis and dynamic weight arbitration on the original frequency modulation command and the original valve control command to obtain the final coordinated control command include:

[0084] Based on the original frequency modulation command and the original valve control command, a strong coupling conflict flag is obtained; Dynamic weight arbitration is performed based on the strong coupling conflict flag to obtain the liquid level weight factor and pressure weight factor. Based on the original frequency modulation command, the original valve control command, the liquid level weighting factor, and the pressure weighting factor, the final coordinated control command is obtained.

[0085] like Figure 4 As shown, the coupling effect of frequency variation on pressure is first evaluated, specifically including: Calculate the change in flow rate caused by a unit change in frequency, where k is the unit change in flow rate. fp =Q rated / (f rated ×η); In the formula, Q rated This is the pump's rated flow rate; f rated The rated frequency; η is the current pump efficiency estimate, which can be estimated based on the load rate.

[0086] Calculate the influence coefficient of flow rate change on pressure, influence coefficient k. qp =2×R×Q out A comprehensive assessment of the coupling pressure change ΔP est =k fp ×k qp ×Δf raw ; Δf raw This will cause a change in flow rate, which will affect the current pipe resistance R and the current output flow rate Q. out This will lead to changes in pressure.

[0087] Determine the direction of the influence of the original frequency modulation command and the original valve control command on the flow rate; If Δf raw If the value is greater than 0, the frequency will increase, thus increasing the output flow. If ΔV raw If the value is greater than 0, the valve will open to reduce pipeline resistance and increase the output flow rate. If Δf raw If the value is less than 0, the frequency will be reduced, thus decreasing the output flow. If ΔV raw If the value is less than 0, the valve will be closed, increasing pipeline resistance and reducing output flow. Based on this, the necessary and sufficient condition for a conflict is sign(Δf) raw )≠sign(ΔV raw ); The sign function (·) is used to indicate the sign of a number.

[0088] Calculate the absolute value of the change in coupling pressure and set a conflict threshold P. conf This threshold indicates that if the estimated change in coupling pressure is less than this value, the impact of the conflict is considered minor and can be ignored. The specific value is determined based on experiments.

[0089] In summary, when sign(Δf) raw )≠sign(ΔV raw And |ΔP est │>P conf Set strong coupling conflict to true, otherwise set it to false.

[0090] This embodiment can predict the magnitude of pressure disturbance caused by the frequency regulation action of the liquid level loop, and based on the direction of the influence of the original valve control command on the flow rate, identify conflicting operating conditions with significant negative coupling, identify potential internal oscillation risks in advance, and provide key and reliable decision-making basis for subsequent execution of dynamic weight arbitration.

[0091] When the strong coupling conflict is true, dynamic weight arbitration is performed to obtain the liquid level weight factor and pressure weight factor to quantify the urgency of the real-time control deviation.

[0092] Based on the current liquid level deviation e H and the safe liquid level range [H] low H high ] Calculate the safety margin M of the liquid level H That is, e H If the value is greater than 0, calculate the margin from the upper limit; otherwise, calculate the margin from the lower limit. In the formula, the safety margin of the liquid level is M. H Always positive; Calculate the liquid level weighting factor W H =│e H │ / (M H +ε); In the formula, ε is a decimal number to prevent division by zero; This step indicates that the greater the level deviation and the smaller the level safety margin, the higher the value of this factor and the greater the urgency of level control.

[0093] Based on the current pressure deviation e P and pressure target range [P] min ,P max ] Calculate the pressure safety margin M P That is, e P >0, calculate the margin from the upper limit; Otherwise, calculate the margin from the lower limit, and similarly calculate the pressure weighting factor W. P =│e P │ / (M P +ε).

[0094] Then, normalization is performed so that W H +W P =1.

[0095] By calculating the degree to which the liquid level and pressure deviate from their safety boundaries, a weighting factor reflecting the urgency of their respective controls is determined. This ensures that when the liquid level is on the verge of being out of control or the pressure is close to its limit, the stability of the most urgent loop is prioritized. This not only resolves control conflicts but also significantly improves the overall safety of multivariable control under complex operating conditions.

[0096] According to the strong coupling conflict flag, if the strong coupling conflict flag is false, it indicates that there is no significant conflict or slight coupling effect between the original frequency modulation command and the original valve control command, and a pass-through strategy is adopted, that is, the final frequency modulation command Δf final =Δf raw Final valve control command ΔV final =ΔV raw This process yields the final coordinated control command. If the strong coupling conflict flag is true, it indicates a significant conflict between the original frequency modulation command and the original valve control command. In this mode, the original frequency modulation command and the original valve control command are synthesized and compensated based on dynamically calculated weighting factors to generate the final frequency modulation command and valve control command.

[0097] The original frequency modulation command of the liquid level loop is scaled according to the liquid level weighting factor, actively subtracting a portion of the coupling effects that are detrimental to pressure caused by its own actions; Calculate the final frequency modulation command Δf final =W H ×Δf raw -β×ΔP est ; In the formula, This is the coupling compensation coefficient, used to balance the effect of counteracting coupling interference with the retention of the main control action of the liquid level loop.

[0098] The original valve control command in the pressure loop is scaled by a pressure weighting factor. Since coupling effects are primarily handled in the level loop command, no further compensation is needed. The final valve control command ΔV is then calculated. final =W P ×ΔV raw The final frequency modulation command and the final valve control command together constitute the final coordinated control command.

[0099] Preferably, a global check can be performed before the final coordinated control command is sent to the corresponding execution device to calculate the frequency and valve opening after the final coordinated control command is executed, and to verify whether it is within the physical limits of the execution device, so that the final coordinated control command is safe and effective.

[0100] It can maintain a fast direct response when there is no conflict in the commands, and when a strong coupling conflict is detected, it makes a reasonable compromise between the commands of both parties based on the dynamic arbitration weight. It also innovatively embeds the prediction compensation for the pressure disturbance caused by itself into the liquid level loop command, avoiding the antagonistic oscillation between the two loops. While ensuring the core functions of each control loop, it achieves the global optimization and dynamic stability of the overall operation of the multi-level workstation, and significantly improves the control and coordination capabilities under complex coupling conditions.

[0101] It also provides a step-by-step instruction execution and cycle iteration process, sending the final coordinated control instruction to the corresponding execution device, i.e., the final frequency adjustment instruction is sent to the frequency converter, and the final valve adjustment instruction is sent to the intelligent regulating valve for execution. Simultaneously, it stores current liquid level deviation, pressure deviation, etc., for the next control cycle, and updates historical multi-source feature data for predicting slurry density change trends. The collected data, multi-source feature data, and final coordinated control instruction for the current control cycle are uploaded to the monitoring platform through a multi-level control network for remote monitoring and historical backtracking. After completion, it waits for the next control cycle.

[0102] When the grouting volume reaches the required level or a stop command is received, the multi-stage workstations execute a sequential shutdown procedure. The sequence is as follows: from the last-stage workstation to the first-stage workstation, the frequency is reduced sequentially until the pump stops. After monitoring the liquid level of each stage's buffer tank and emptying it, the corresponding pipeline valves are closed, and finally, the power supply is cut off. This process ensures that the grout in the pipeline is discharged in an orderly manner, preventing sedimentation and blockage.

[0103] It also includes monitoring the sensor data and equipment status of multi-level workstations, judging anomalies based on preset operating rules, performance prediction models or multivariate correlation analysis models, and identifying equipment anomalies, performance degradation and operational risks.

[0104] Once an anomaly is detected, an alert is issued at the workstation where the anomaly is located and on the remote monitoring platform, displaying the location and type of the anomaly. Based on the severity of the anomaly, safety responses are automatically executed, such as emergency pump shutdown and valve closure, to maximize the safety of equipment and personnel.

[0105] The above content is only a preferred embodiment of the present invention. For those skilled in the art, many changes can be made in the specific implementation and application scope based on the concept of the present invention. As long as these changes do not depart from the concept of the present invention, they all fall within the protection scope of the present invention.

Claims

1. A control method for multi-stage pressurized grouting, characterized in that, This is applied to a multi-stage pressurized grouting system consisting of multiple multi-stage pressurized grouting devices connected in series, wherein each stage device is defined as a first-stage workstation, including: Acquire data collected by various sensors during the trial delivery phase, identify the current operating conditions, and obtain baseline control parameters; Collect multi-sensor data from this level workstation and associated data from the next level workstation, and obtain the current multi-source feature data after preprocessing; Based on the current multi-source characteristic data and baseline control parameters, the dynamically adjusted control parameters are obtained. The control parameters include the liquid level loop proportional gain, the pressure loop adjustment coefficient, and the disturbance state judgment result. Based on the dynamically adjusted control parameters, the original frequency modulation command for the first closed loop and the original valve control command for the second closed loop are obtained. The original frequency modulation command and the original valve control command are coupled and dynamically weighted to obtain the final coordinated control command.

2. The control method for multi-stage pressurized grouting according to claim 1, characterized in that, Based on the current multi-source feature data and baseline control parameters, the dynamically adjusted control parameters are obtained, including: Based on the current multi-source feature data, the inertial change of the slurry is predicted by feedforward, and the proportional gain of the liquid level loop is adaptively adjusted. Joint analysis of pressure fluctuation characteristics in current multi-source feature data is performed to generate disturbance state determination results; Based on the current multi-source characteristic data and benchmark control parameters, the valve status and slurry resistance status are evaluated to obtain the pressure loop adjustment coefficient.

3. The control method for multi-stage pressurized grouting according to claim 2, characterized in that, Based on current multi-source feature data, feedforward prediction of slurry inertial changes is performed, and adaptive adjustment is used to obtain the proportional gain of the liquid level loop, including: Based on the density deviation in current multi-source feature data and historical multi-source feature data, a density deviation sequence is constructed and the density change trend is calculated. Based on the density change trend, the inertial prediction factor is obtained; The gain correction amount is obtained by querying the preset mapping table based on the inertial prediction factor; The liquid level loop proportional gain is calculated based on the gain correction amount and the liquid level loop reference proportional gain in the reference control parameters.

4. The control method for multi-stage pressurized grouting according to claim 2, characterized in that, Based on current multi-source characteristic data and baseline control parameters, the valve status and slurry resistance status are evaluated to obtain the pressure loop regulation coefficient, including: Based on the valve opening feedback from the current multi-source feature data, and combined with the valve nonlinearity interval division threshold of the benchmark control parameters, the current valve opening interval is obtained. Based on the slurry density deviation in the current multi-source feature data, the slurry resistance change state is obtained; Based on the valve's current opening range, the slurry resistance change status, and the pressure loop reference adjustment coefficient in the reference control parameters, the current pressure loop adjustment coefficient is obtained.

5. The control method for multi-stage pressurized grouting according to claim 1, characterized in that, The first closed loop is a closed loop with the core objective of stabilizing the liquid level in the buffer tank, and its output is the original frequency modulation command for adjusting the speed of the slurry pump. The second closed loop is a closed loop with the core objective of stabilizing the pipeline outlet pressure, and its output is the original valve control command to adjust the valve opening.

6. The control method for multi-stage pressurized grouting according to claim 5, characterized in that, The acquisition of the original frequency modulation command for the first closed loop includes: Based on the current proportional gain of the level loop in the dynamically adjusted control parameters, and combined with the level deviation between the buffer tank level and the set level value, the proportional term output is calculated. Based on the liquid level deviation, the anti-saturation integral term is obtained; The proportional term output is superimposed with the integral term to obtain the original frequency modulation command for the first closed loop.

7. A control method for multi-stage pressurized grouting according to claim 5, characterized in that, The acquisition of the original valve control command in the second closed loop includes: Obtain the disturbance state determination result in the dynamically adjusted control parameters. If the determination result is an instantaneous disturbance, the original valve control command is 0. If the deviation is determined to be a true pressure deviation, the value of the original valve control command is calculated based on the current adjustment coefficient of the pressure loop in the dynamically adjusted control parameters.

8. The control method for multi-stage pressurized grouting according to claim 1, characterized in that, The original frequency modulation command and the original valve control command are coupled and dynamically weighted to obtain the final coordinated control command, including: Based on the original frequency modulation command and the original valve control command, a strong coupling conflict flag is obtained; Dynamic weight arbitration is performed based on the strong coupling conflict flag to obtain the liquid level weight factor and pressure weight factor. Based on the original frequency modulation command, the original valve control command, the liquid level weighting factor, and the pressure weighting factor, the final coordinated control command is obtained.

9. The control method for multi-stage pressurized grouting according to claim 1, characterized in that, Collect multi-sensor data from this workstation and associated data from the next-level workstation, and preprocess the data to obtain the current multi-source feature data, including: Based on the multi-sensor data collected from the current workstation and the associated data from the next workstation during the continuous delivery phase, effective verification and bad value removal are performed. Multidimensional feature extraction was performed on the processed multi-sensor data and associated data to obtain pipeline outlet pressure fluctuation characteristics, slurry state characteristics, pipeline resistance characteristics, and upper and lower level matching characteristics. By packaging the pipeline outlet pressure fluctuation characteristics, slurry state characteristics, pipeline resistance characteristics, and upper and lower level matching characteristics, the current multi-source characteristic data is obtained.

10. A control system for multi-stage pressurized grouting, used to implement the control method for multi-stage pressurized grouting as described in any one of claims 1-9, characterized in that, Deployed in each workstation of a multi-stage pressurized grouting system, which consists of multiple multi-stage pressurized grouting devices connected in series, including: Initialize the scene recognition module: used to acquire data collected by various sensors during the trial delivery phase, identify the current working condition scene, and obtain the baseline control parameters; Data feature extraction module: used to collect multi-sensor data from the current workstation and related data from the next level workstation, and obtain the current multi-source feature data after preprocessing; Dynamic control parameter adjustment module: used to obtain dynamically adjusted control parameters based on current multi-source characteristic data and baseline control parameters. The control parameters include liquid level loop proportional gain, pressure loop adjustment coefficient and disturbance state judgment result. Original instruction generation module: used to obtain the original frequency modulation instruction for the first closed loop and the original valve control instruction for the second closed loop based on the dynamically adjusted control parameters. Coupling and Coordination Command Generation Module: This module performs coupling analysis and dynamic weight arbitration on the original frequency modulation command and the original valve control command to obtain the final coordinated control command.

Images