Intelligent control system for internal gear pump

Abstract

The application discloses an internal gear pump intelligent control system. A detection module collects working condition parameters such as meshing clearance, four types of external interference parameters, tooth surface fatigue-lubrication parameters containing acoustic emission signals and metal abrasive particle content in real time. An edge computing module calculates volume efficiency based on meshing clearance and lubrication state, obtains an interference comprehensive coefficient by dimensionless weighting of external interference, and obtains fatigue damage degree according to acoustic emission, abrasive particle, clearance and time length; a full-link energy consumption model is constructed with meshing clearance, interference comprehensive coefficient and fatigue damage degree as correction factors, and real-time energy consumption is calculated; optimal operation parameters are solved by dynamically adjusting and optimizing weights with the minimum energy consumption as a target, and a coordinated control signal is generated after PID coefficients are set according to three variables with optimal flow as a given value. An execution module responds to signals to coordinate and adjust motor speed, pipe network valve opening degree and active lubrication parameters. The application realizes full-closed loop intelligent control of mechanical state-external interference-tooth surface fatigue coupling.

Description

Technical Field

[0001] This invention relates to the fields of fluid machinery and automatic control technology, and in particular to an intelligent control system for an internal gear pump. Background Technology

[0002] Internal gear pumps are widely used in industrial applications such as petrochemicals, engineering machinery, metallurgy, mining, and hydraulic transmission due to their compact structure, stable output flow, low noise, and good self-priming performance. They are key power devices in fluid transport systems. As industrial systems develop towards high-precision control, intelligent operation, and energy conservation and emission reduction, higher requirements are placed on the operational stability, energy efficiency, and service life of internal gear pumps.

[0003] During long-term operation, internal gear pumps are not only affected by changes in their internal mechanical structure, but also inevitably by external environmental disturbances and operating condition fluctuations. Key meshing components within the pump body undergo gradual state evolution under continuous load; simultaneously, factors such as power supply conditions, pipeline pressure fluctuations, environmental vibrations, and changes in the physical properties of the medium can all disrupt system operation. Under complex operating conditions, these multiple factors often exhibit a coupled and superimposed state, comprehensively impacting pump operating efficiency, control stability, and lifespan reliability.

[0004] Existing control systems for internal gear pumps often focus on adjusting single operating parameters or optimizing local performance, lacking a comprehensive perception and collaborative processing mechanism for multi-source operating status information. Under complex industrial conditions, when mechanical conditions, external disturbances, and lubrication conditions change dynamically, the control system struggles to simultaneously optimize operating efficiency, energy consumption levels, and structural health, leading to fluctuations in operating accuracy, unstable energy consumption levels, and difficulty in timely assessment of fatigue risks in key components.

[0005] Furthermore, the tooth surface undergoes a microstructural evolution process during long-term meshing, and its damage development exhibits staged and cumulative characteristics. Without an effective mechanism for identifying and coordinating the control of early changes in condition, problems such as decreased efficiency, increased vibration, or reduced reliability may occur in later stages of operation. Simultaneously, there is a dynamic correlation between lubrication status, energy consumption level, and structural wear; if these are not adjusted in a coordinated manner according to operating conditions, it can easily lead to an imbalance in the overall system performance.

[0006] In the current trend of industrial intelligence development, how to achieve multi-parameter comprehensive perception, dynamic modeling and closed-loop optimization control of the operating status of internal gear pumps in complex disturbance environments, so that the system can ensure operating accuracy while taking into account energy efficiency and structural health, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0007] To address the shortcomings of existing technologies, the present invention aims to provide an intelligent control system for an internal meshing gear pump, which enables deep coupled perception and dynamic modeling of mechanical wear state, external interference intensity, and tooth surface fatigue degree, and drives PID coefficient self-tuning and multi-execution link coordinated adjustment with the goal of optimizing energy consumption across the entire link.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an intelligent control system for an internal gear pump, comprising:

[0009] The detection module is used to collect the operating parameters, external interference parameters, and tooth surface fatigue-lubrication parameters of the internal gear pump in real time. The operating parameters include the meshing clearance, the external interference parameters include power grid voltage fluctuations, environmental vibrations, sudden changes in medium viscosity, and pipeline pressure shocks, and the tooth surface fatigue-lubrication parameters include tooth surface meshing acoustic emission signals, metal abrasive content, and lubrication status parameters.

[0010] The edge computing module, connected to the detection module, is configured as follows:

[0011] Real-time volumetric efficiency is calculated based on meshing clearance and lubrication status parameters;

[0012] External interference parameters are dimensionless and weighted fused to generate comprehensive interference coefficients;

[0013] The degree of tooth surface fatigue damage is calculated based on acoustic emission signals, metal abrasive content, meshing clearance, and duration.

[0014] A full-link energy consumption model of pump body, motor, pipeline network and lubrication system is constructed. The meshing clearance, interference comprehensive coefficient and tooth surface fatigue damage degree are used as correction factors. Real-time full-link energy consumption is calculated by combining real-time volumetric efficiency and operating parameters.

[0015] With the goal of minimizing energy consumption across the entire chain, the optimization weights are adjusted based on the interference comprehensive coefficient and the degree of fatigue damage to solve for the optimal operating parameters, which include the optimal pump outlet flow rate, optimal motor speed, optimal pipeline valve opening, and optimal lubrication parameters.

[0016] Using the optimal pump outlet flow rate as a given value, the PID coefficients are tuned according to the meshing clearance, interference comprehensive coefficient, and tooth surface fatigue damage degree;

[0017] A coordinated control signal is generated based on the tuned PID coefficients and the optimal operating parameters.

[0018] An execution module, connected to the edge computing module, is used to respond to the collaborative control signal and adjust the motor speed, the pipeline valve opening, and the lubrication parameters of the active lubrication system according to the optimal motor speed, the optimal pipeline valve opening, and the optimal lubrication parameters.

[0019] The edge computing module also iteratively executes the above operations based on the feedback operating parameters.

[0020] Furthermore, the edge computing module includes:

[0021] The volumetric efficiency calculation unit, connected to the detection module, is configured as follows:

[0022] The detection module acquires real-time meshing clearance, real-time pump body temperature rise, real-time medium viscosity, and real-time lubrication flow rate, as well as a preset initial volumetric efficiency and a preset reference lubrication flow rate.

[0023] The meshing clearance correction amount is generated based on the real-time meshing clearance and the preset meshing clearance influence coefficient; the thermal effect correction amount is generated based on the real-time pump body temperature rise and the preset temperature correction coefficient; the medium viscosity correction amount is generated based on the absolute value of the deviation between the real-time medium viscosity and the rated medium viscosity and the preset viscosity correction coefficient; and the lubrication state correction amount is generated based on the ratio of the real-time lubrication flow rate to the preset reference lubrication flow rate and the preset lubrication effect correction coefficient.

[0024] The real-time volumetric efficiency is generated by subtracting the meshing clearance correction, the thermal effect correction, the medium viscosity correction, and the lubrication state correction from the initial volumetric efficiency.

[0025] Among them, the preset meshing clearance influence coefficient, the preset temperature correction coefficient, the preset viscosity correction coefficient, and the preset lubrication effect correction coefficient are all configurable thresholds calibrated through experiments.

[0026] Furthermore, the edge computing module includes:

[0027] The interference quantization unit, connected to the detection module, is configured as follows:

[0028] The detection module acquires real-time grid voltage, real-time environmental vibration acceleration, real-time medium viscosity, and real-time pipeline pressure impact peak value.

[0029] A voltage interference quantification value is generated based on the ratio of the absolute value of the deviation between the real-time grid voltage and the rated grid voltage to the rated grid voltage.

[0030] Based on the ratio of the real-time environmental vibration acceleration to the preset maximum allowable vibration acceleration threshold, a vibration interference quantification value is generated;

[0031] Based on the ratio of the absolute value of the deviation between the real-time medium viscosity and the rated medium viscosity to the rated medium viscosity, a viscosity interference quantification value is generated.

[0032] Based on the ratio of the real-time pipeline pressure shock peak value to the pump outlet rated pressure, a pressure shock interference quantification value is generated.

[0033] Obtain preset voltage interference weighting coefficients, vibration interference weighting coefficients, viscosity interference weighting coefficients, and pressure shock interference weighting coefficients; then perform a weighted summation of the quantized values ​​of voltage interference, vibration interference, viscosity interference, and pressure shock interference to generate a comprehensive interference coefficient.

[0034] Among them, the preset maximum allowable vibration acceleration threshold, the preset voltage interference weight coefficient, the preset vibration interference weight coefficient, the preset viscosity interference weight coefficient, and the preset pressure impact interference weight coefficient are all configurable thresholds calibrated through experiments, and the sum of each weight coefficient is a fixed threshold.

[0035] Furthermore, the edge computing module includes:

[0036] The fatigue prediction unit, connected to the detection module, is configured as follows:

[0037] The detection module acquires real-time meshing acoustic emission signals, real-time metal abrasive content, real-time meshing clearance, and cumulative running time.

[0038] The real-time meshing acoustic emission signal is subjected to low-pass filtering below a preset filtering frequency threshold, and the peak characteristics of the filtered signal are extracted.

[0039] Obtain the preset acoustic emission peak threshold, the preset maximum allowable metal abrasive content threshold, the preset maximum allowable meshing clearance threshold, and the preset tooth surface design fatigue life threshold;

[0040] An acoustic emission contribution factor is generated based on the ratio of the peak characteristic to the preset acoustic emission peak threshold; an abrasive contribution factor is generated based on the ratio of the real-time metal abrasive content to the preset maximum allowable metal abrasive content threshold; a clearance contribution factor is generated based on the ratio of the real-time meshing clearance to the preset maximum allowable meshing clearance threshold; and a duration contribution factor is generated based on the ratio of the cumulative running time to the preset tooth surface design fatigue life threshold.

[0041] Obtain preset acoustic emission weighting coefficients, abrasive particle weighting coefficients, gap weighting coefficients, and duration weighting coefficients. Then, perform a weighted summation on the acoustic emission contribution factor, the abrasive particle contribution factor, the gap contribution factor, and the duration contribution factor to generate the degree of tooth surface fatigue damage.

[0042] Among them, the preset filter frequency threshold, the preset acoustic emission peak threshold, the preset maximum allowable metal abrasive content threshold, the preset maximum allowable meshing clearance threshold, the preset tooth surface design fatigue life threshold, and each of the weight coefficients are all configurable thresholds calibrated through experiments, and the sum of each of the weight coefficients is a fixed threshold.

[0043] Furthermore, the edge computing module includes:

[0044] The energy consumption modeling unit, connected to the volumetric efficiency calculation unit, the interference quantification unit, the fatigue prediction unit, and the detection module, is configured as follows:

[0045] The detection module acquires pump outlet pressure, pump outlet flow rate, motor speed, frequency converter frequency, pipeline length, pipeline diameter, medium density, real-time lubrication pressure, and real-time lubrication flow rate.

[0046] Based on the pump outlet pressure, the pump outlet flow rate, and the real-time volumetric efficiency, combined with a preset pump body mechanical efficiency threshold, the energy consumption of the pump body sub-stage is calculated.

[0047] Based on the energy consumption of the pump body sub-component, the motor speed, the frequency conversion frequency and the preset motor efficiency mapping relationship, combined with the voltage interference quantification value and the preset voltage interference correction coefficient, the energy consumption of the motor sub-component is calculated.

[0048] Based on the pipeline length, the pipeline diameter, the medium density, the pump outlet flow rate, and the preset pipeline friction coefficient threshold, combined with the pressure shock interference quantification value and the preset pressure shock correction coefficient, the energy consumption of the pipeline sub-link is calculated.

[0049] Based on the real-time lubrication pressure, the real-time lubrication flow rate, and the preset lubrication system efficiency threshold, calculate the energy consumption of the lubrication system sub-processes.

[0050] The energy consumption of the pump body sub-component, the energy consumption of the motor sub-component, the energy consumption of the pipeline sub-component, and the energy consumption of the lubrication system sub-component are summed to generate the basic energy consumption value of the entire link;

[0051] Among them, the preset pump body mechanical efficiency threshold, the preset voltage interference correction coefficient, the preset pipeline friction coefficient threshold, the preset pressure shock correction coefficient, the preset lubrication system efficiency threshold, and the preset motor efficiency mapping relationship are all configurable thresholds calibrated through experiments.

[0052] Furthermore, the edge computing module includes:

[0053] The energy consumption correction unit, connected to the energy consumption modeling unit, the interference quantization unit, and the fatigue prediction unit respectively, is configured as follows:

[0054] The energy consumption modeling unit generates the full-link basic energy consumption value, the interference quantification unit generates the interference comprehensive coefficient, and the fatigue prediction unit generates the tooth surface fatigue damage degree.

[0055] Obtain the preset interference correction coefficient threshold and the preset fatigue correction coefficient threshold;

[0056] A first correction factor is generated based on the interference comprehensive coefficient and the preset interference correction coefficient threshold, and a second correction factor is generated based on the degree of tooth surface fatigue damage and the preset fatigue correction coefficient threshold.

[0057] The full-link basic energy consumption value is multiplied by the first correction factor and the second correction factor respectively to generate the real-time full-link energy consumption;

[0058] The preset interference correction coefficient threshold and the preset fatigue correction coefficient threshold are both configurable thresholds calibrated through experiments.

[0059] Furthermore, the edge computing module includes:

[0060] The optimization solution unit is connected to the energy consumption modeling unit, the interference quantification unit, and the fatigue prediction unit, respectively; the optimization solution unit includes:

[0061] The weight adjustment subunit is configured as follows:

[0062] Obtain the interference comprehensive coefficient generated by the interference quantization unit and the degree of tooth surface fatigue damage generated by the fatigue prediction unit;

[0063] The system has a first interference level threshold, a second interference level threshold, a first fatigue level threshold, a second fatigue level threshold, and at least four sets of optimization weight allocation strategies. Each set of strategies includes at least energy consumption optimization weight, control accuracy weight, and lubrication optimization weight.

[0064] The interference comprehensive coefficient is compared with the first interference level threshold and the second interference level threshold, and the tooth surface fatigue damage degree is compared with the first fatigue level threshold and the second fatigue level threshold.

[0065] When the degree of fatigue damage to the tooth surface is greater than the second fatigue level threshold, a first optimization weight allocation strategy is selected, wherein the value of the lubrication optimization weight in the first optimization weight allocation strategy is greater than the values ​​of the energy consumption optimization weight and the control accuracy weight.

[0066] When the interference comprehensive coefficient is greater than the second interference level threshold, a second optimization weight allocation strategy is selected, wherein the value of the control accuracy weight in the second optimization weight allocation strategy is greater than the values ​​of the energy consumption optimization weight and the lubrication optimization weight.

[0067] When the degree of tooth surface fatigue damage is less than or equal to the first fatigue level threshold and the interference comprehensive coefficient is less than or equal to the first interference level threshold, a third optimization weight allocation strategy is selected. In the third optimization weight allocation strategy, the value of the energy consumption optimization weight is greater than the value of the control accuracy weight and the lubrication optimization weight.

[0068] When the degree of tooth surface fatigue damage is between the first fatigue level threshold and the second fatigue level threshold, and the interference comprehensive coefficient is between the first interference level threshold and the second interference level threshold, a fourth optimization weight allocation strategy is selected, wherein the energy consumption optimization weight, the control accuracy weight, and the lubrication optimization weight are all equal in the fourth optimization weight allocation strategy.

[0069] Wherein, the first interference level threshold, the second interference level threshold, the first fatigue level threshold, the second fatigue level threshold, and the weight values ​​in each of the optimized weight allocation strategies are all configurable thresholds calibrated through experiments.

[0070] Furthermore, the optimization solution unit also includes:

[0071] The parameter optimization subunit, connected to the weight adjustment subunit, the energy consumption modeling unit, and the detection module, is configured as follows:

[0072] The system acquires the real-time end-to-end energy consumption model generated by the energy consumption modeling unit, the current operating condition parameters collected by the detection module, and the current optimized weight allocation strategy output by the weight adjustment subunit.

[0073] With the goal of minimizing energy consumption across the entire chain, and with pump outlet flow, motor speed, pipeline valve opening, and lubrication parameters of the active lubrication system as variables to be optimized, a weighted multi-objective optimization function is constructed based on the current optimization weight allocation strategy.

[0074] The gradient descent method is used to iteratively optimize the variable to be optimized. Each variable is updated along the negative gradient direction with a preset learning rate threshold until the difference in the total energy consumption of two adjacent iterations is less than a preset convergence threshold or the number of iterations reaches a preset maximum number of iterations threshold.

[0075] The pump outlet flow rate, motor speed, pipeline valve opening, and lubrication parameters at the end of the iteration are determined as the optimal pump outlet flow rate, optimal motor speed, optimal pipeline valve opening, and optimal lubrication parameters, respectively.

[0076] The preset learning rate threshold, the preset convergence threshold, and the preset maximum number of iterations threshold are all configurable thresholds calibrated through experiments.

[0077] Furthermore, the edge computing module includes:

[0078] A PID collaborative control unit is connected to the optimization solution unit, the detection module, the interference quantization unit, and the fatigue prediction unit, respectively; the PID collaborative control unit includes:

[0079] The PID tuning subunit is configured as follows:

[0080] The optimal pump outlet flow rate determined by the parameter optimization subunit is obtained as the control setpoint, and the real-time meshing clearance collected by the detection module, the interference comprehensive coefficient generated by the interference quantification unit, and the degree of tooth surface fatigue damage generated by the fatigue prediction unit are obtained.

[0081] The system includes preset thresholds for PID initial proportional coefficient, PID initial integral coefficient, PID initial derivative coefficient, design meshing clearance, clearance correction coefficient, first disturbance correction coefficient, second disturbance correction coefficient, third disturbance correction coefficient, first fatigue correction coefficient, second fatigue correction coefficient, and third fatigue correction coefficient.

[0082] Based on the deviation between the real-time meshing clearance and the designed meshing clearance threshold, the clearance correction coefficient threshold, and the PID initial proportional coefficient threshold, combined with the product of the interference comprehensive coefficient and the first interference correction coefficient threshold, and the product of the tooth surface fatigue damage degree and the first fatigue correction coefficient threshold, the tuned proportional coefficient is generated.

[0083] Based on the deviation between the real-time meshing clearance and the designed meshing clearance threshold, the clearance correction coefficient threshold, and the PID initial integral coefficient threshold, combined with the product of the interference comprehensive coefficient and the second interference correction coefficient threshold, and the product of the tooth surface fatigue damage degree and the second fatigue correction coefficient threshold, the tuned integral coefficient is generated.

[0084] Based on the deviation between the real-time meshing clearance and the designed meshing clearance threshold, the clearance correction coefficient threshold, and the PID initial differential coefficient threshold, combined with the product of the interference comprehensive coefficient and the third interference correction coefficient threshold, and the product of the tooth surface fatigue damage degree and the third fatigue correction coefficient threshold, the tuned differential coefficient is generated.

[0085] The collaborative signal generation subunit, connected to both the PID tuning subunit and the optimization solution unit, is configured as follows:

[0086] The adjusted proportional coefficient, the adjusted integral coefficient, and the adjusted differential coefficient are obtained, as well as the optimal motor speed, optimal pipeline valve opening, and optimal lubrication parameters determined by the parameter optimization subunit.

[0087] Based on the deviation between the optimal pump outlet flow rate and the real-time pump outlet flow rate, PID calculation is performed using the adjusted proportional coefficient, the adjusted integral coefficient, and the adjusted derivative coefficient to generate a pump flow rate adjustment signal.

[0088] Based on the optimal motor speed, optimal pipeline valve opening, and optimal lubrication parameters, motor speed adjustment signals, valve opening adjustment signals, and lubrication parameter adjustment signals are generated respectively.

[0089] The pump flow rate adjustment signal, motor speed adjustment signal, valve opening adjustment signal, and lubrication parameter adjustment signal are synchronously output to the execution module.

[0090] Among them, the initial proportional coefficient threshold of the PID, the initial integral coefficient threshold of the PID, the initial derivative coefficient threshold of the PID, the design meshing clearance threshold, the clearance correction coefficient threshold, the first interference correction coefficient threshold, the second interference correction coefficient threshold, the third interference correction coefficient threshold, the first fatigue correction coefficient threshold, the second fatigue correction coefficient threshold, and the third fatigue correction coefficient threshold are all configurable thresholds calibrated through experiments.

[0091] Furthermore, the execution module includes:

[0092] The motor drive execution unit, connected to both the cooperative signal generation subunit and the drive motor, is configured as follows:

[0093] Acquire the motor speed regulation signal and the pump flow regulation signal;

[0094] The built-in pump-specific variable frequency drive responds to the motor speed adjustment signal to adjust the real-time speed of the drive motor to the optimal motor speed, and responds to the pump flow adjustment signal and the calibrated PID coefficient to perform variable frequency closed-loop regulation of the drive motor, so that the deviation between the pump outlet flow and the optimal pump outlet flow is maintained within the preset flow deviation threshold.

[0095] The pipeline regulation execution unit, connected to both the coordinated signal generation subunit and the electric regulating valve of the outlet pipeline, is configured as follows:

[0096] Obtain the valve opening adjustment signal;

[0097] In response to the valve opening adjustment signal, the real-time opening of the electric regulating valve is adjusted to the optimal pipeline valve opening, so that the pipeline resistance characteristics are matched with the optimal motor speed and the optimal pump outlet flow rate;

[0098] The active lubrication execution unit, connected to both the cooperative signal generation subunit and the active lubrication system, is configured as follows:

[0099] The lubrication parameter adjustment signal is acquired, and the lubrication parameter adjustment signal includes at least the optimal lubrication flow rate adjustment command, the optimal lubrication pressure adjustment command, and the optimal lubrication temperature adjustment command.

[0100] The active lubrication actuator includes:

[0101] The variable frequency drive for the lubrication pump adjusts the speed of the lubrication pump to a speed corresponding to the optimal lubrication flow rate in response to the optimal lubrication flow rate adjustment command;

[0102] A miniature electromagnetic proportional valve, in response to the optimal lubrication pressure adjustment command, adjusts the lubrication circuit pressure to the optimal lubrication pressure;

[0103] An electric three-way regulating valve, in response to the optimal lubrication temperature regulation command, adjusts the opening of the lubrication and cooling branch to the opening corresponding to the optimal lubrication temperature;

[0104] The active lubrication execution unit also feeds back the adjusted real-time lubrication flow rate, real-time lubrication pressure and real-time lubrication temperature to the edge computing module through the detection module. The edge computing module is used to determine whether the deviation between each real-time lubrication parameter and the corresponding optimal lubrication parameter exceeds the preset lubrication adjustment deviation threshold.

[0105] The preset flow deviation threshold and the preset lubrication adjustment deviation threshold are both configurable thresholds calibrated through experiments.

[0106] The beneficial effects of this invention are:

[0107] This invention achieves coordinated closed-loop adjustment of operating parameters by fusing and sensing multiple sources of operating parameters, external disturbance parameters, and tooth surface fatigue-lubrication parameters, and combining edge computing to construct a full-link energy consumption model and dynamic optimization control mechanism. This has the following beneficial effects:

[0108] 1. Significantly improved control precision and anti-interference capability:

[0109] By using meshing clearance, disturbance comprehensive coefficient, and tooth surface fatigue damage degree as control parameters and model correction factors, the operating model can be corrected in real time, making the pump outlet flow control more stable, reducing control deviations caused by changes in mechanical state and external disturbances, and improving the stability and robustness of the system under complex working conditions.

[0110] 2. Achieve coordinated optimization of energy consumption across the entire energy chain:

[0111] By constructing a full-link energy consumption model of the pump body, motor, pipeline network and lubrication system, and using mechanical state and disturbance degree as correction variables for dynamic optimization, the energy consumption of each link is no longer adjusted in isolation, achieving optimal control of overall energy consumption and improving the overall energy efficiency of the system.

[0112] 3. Achieve predictive control of tooth surface fatigue:

[0113] A fatigue damage assessment mechanism is constructed based on acoustic emission signals, metal abrasive content, and meshing clearance changes. This allows the tooth surface condition to participate in the optimization of operating parameters and lubrication adjustment process, achieving synergy between structural protection and operational control, and improving the reliability and service life of key components.

[0114] 4. Form a multi-module collaborative closed-loop control system:

[0115] The detection module, edge computing module, and execution module form an iterative optimization closed loop, which can continuously update the optimal control strategy based on feedback operating parameters, so that the system maintains a better operating state at different stages, improving overall operating efficiency and long-term stability.

[0116] 5. A comprehensive optimization effect that balances energy saving, accuracy, and lifespan:

[0117] By deeply coupling mechanical state, disturbance factors and energy consumption model, a dynamic balance is achieved between control precision, energy utilization efficiency and structural health, meeting the requirements of high precision and long-term stable operation in industrial scenarios. Attached Figure Description

[0118] Figure 1 This is a schematic diagram of the intelligent control system for the internal meshing gear pump in this invention;

[0119] Figure 2 This is a schematic diagram of the structure of the optimization solution unit in this invention;

[0120] Figure 3 This is a schematic diagram of the PID collaborative control unit in this invention.

[0121] Reference numerals: 1. Detection module; 2. Edge computing module; 21. Volumetric efficiency calculation unit; 22. Interference quantification unit; 23. Fatigue prediction unit; 24. Energy consumption modeling unit; 25. Energy consumption correction unit; 26. Optimization solution unit; 261. Weight adjustment subunit; 262. Parameter optimization subunit; 27. PID collaborative control unit; 271. PID tuning subunit; 272. Collaborative signal generation subunit; 3. Execution module; 31. Motor drive execution unit; 32. Pipeline regulation execution unit; 33. Active lubrication execution unit. Detailed Implementation

[0122] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Identical components are denoted by the same reference numerals. It should be noted that the terms "front," "rear," "left," "right," "upper," and "lower" used in the following description refer to directions in the accompanying drawings, and the terms "bottom surface," "top surface," "inner," and "outer" refer to directions toward or away from the geometric center of a specific component, respectively.

[0123] like Figure 1 As shown, this is the first embodiment of the present invention. This first embodiment provides an intelligent control system for an internal meshing gear pump, which can realize multi-source fusion analysis of mechanical state, external interference and tooth surface fatigue state, and perform collaborative closed-loop control based on the whole-link energy consumption optimization target, thereby significantly improving control accuracy, reducing energy consumption and extending equipment life.

[0124] I. Overall System Composition;

[0125] This system consists of four parts: a detection module 1, an edge computing module 2, an execution module 3, and a communication module. The detection module 1 is used to collect real-time operating parameters, external interference parameters, and tooth surface fatigue-lubrication parameters of the internal gear pump. The edge computing module 2 is connected to the detection module 1 and performs data fusion, model calculation, and decision optimization. The execution module 3 is connected to the edge computing module 2 and responds to coordinated control signals to adjust multiple aspects of the pump body, motor, pipeline network, and lubrication system. The communication module enables bidirectional real-time data transmission between the modules.

[0126] II. Detection Module 1;

[0127] The detection module 1 includes a working condition detection unit, an external interference detection unit, and a tooth surface fatigue-lubrication detection unit. The selection and arrangement of sensors for each unit are as follows.

[0128] The operating condition detection unit includes:

[0129] The miniature displacement sensor, selected from the KEYENCE GT2 series, has a range of 0 to 0.2 mm and an accuracy of ≥0.1 μm. It is placed on the end cover of the pump body at the top or root of the meshing teeth to collect real-time meshing clearance.

[0130] The pressure sensor is an SSTL series diffused silicon sensor with a range of 0 to 10 MPa and an accuracy of ±0.5%FS. It is installed at the pump inlet and outlet and key nodes of the pipeline network.

[0131] The flow sensor selected is the LDG series electromagnetic flow meter, with a measuring range of 0 to 100 m. 3 / h, accuracy ±0.3%FS, located at the pump outlet and main pipeline; temperature sensor, PT100 platinum resistance, range -20 to 150℃, accuracy ±0.1℃, located on the pump body and motor housing; torque sensor, JN338 series torque sensor, range 0 to 500 N·m, accuracy ±0.2%FS, located at the connection between the pump shaft and the motor.

[0132] The external interference detection unit includes:

[0133] The power grid voltage sensor uses a JDZ10 series voltage transformer with a range of 0 to 450V and an accuracy of ±0.2%FS, and is placed at the motor power supply end.

[0134] The environmental vibration sensor selected is a YZC series piezoelectric vibration sensor with a measurement range of 0 to 10 m / s. 2 Accuracy ±0.5%FS, located on the pump body base;

[0135] The online viscosity sensor is the NDJ-1C series online rotational viscometer, with a range of 1 to 1000 mPa·s and an accuracy of ±1%FS, and a reusable media detection pipeline.

[0136] The pipeline impact sensor is a CYG1001 series pressure impact sensor with a range of 0 to 15 MPa and a response time of ≤1 ms. It is installed in the pipeline near the pump outlet.

[0137] The tooth surface fatigue-lubrication detection unit includes:

[0138] The acoustic emission sensor is the AE110 series industrial-grade acoustic emission sensor, with a frequency response range of 10kHz to 1MHz and an accuracy of ≥1μV. It is placed on the meshing tooth surface of the pump body end cover.

[0139] The oil contamination sensor is a PQ200 series sensor with a range of 0 to 1000 ppm and an accuracy of ±5 ppm, and is installed at the outlet of the lubrication circuit.

[0140] The lubrication pressure sensor is from the Miniature series, with a range of 0 to 2 MPa and an accuracy of ±0.3%FS.

[0141] The lubrication flow sensor uses the MF57 series miniature flow sensor, with a range of 0 to 10 L / min and an accuracy of ±0.5%FS, and is arranged in the active lubrication circuit; the lubricating oil temperature sensor reuses the PT100, and a new temperature measuring branch is arranged at the lubricating oil tank outlet.

[0142] All detection units are connected to the Profinet industrial bus, eliminating the need for additional communication architecture and allowing for direct adaptation to existing internal gear pumps.

[0143] III. Edge Computing Module 2;

[0144] Edge computing module 2 uses a Siemens S7-1500 series PLC, integrating core algorithms such as volumetric efficiency calculation, interference quantification, fatigue prediction, energy consumption modeling, optimization solution, and PID collaborative control. Edge computing module 2 is configured to: calculate real-time volumetric efficiency based on meshing clearance and lubrication state parameters; dimensionlessly convert external interference parameters and generate a weighted fusion interference comprehensive coefficient; calculate the degree of tooth surface fatigue damage based on acoustic emission signals, metal abrasive content, meshing clearance, and duration; construct a full-link energy consumption model of the pump body, motor, pipeline network, and lubrication system, using meshing clearance, interference comprehensive coefficient, and tooth surface fatigue damage degree as correction factors, and calculate real-time full-link energy consumption based on real-time volumetric efficiency and operating parameters; and, with the goal of minimizing full-link energy consumption, adjust and optimize weights based on the interference comprehensive coefficient and fatigue damage degree to solve for optimal operating parameters, including optimal pump outlet flow rate, optimal motor speed, optimal pipeline valve opening, and optimal lubrication parameters.

[0145] Using the optimal pump outlet flow rate as a given value, the PID coefficients are tuned based on the meshing clearance, interference comprehensive coefficient, and tooth surface fatigue damage degree; a coordinated control signal is generated based on the tuned PID coefficients and optimal operating parameters.

[0146] IV. Execution Module 3;

[0147] The execution module 3 is connected to the edge computing module 2. It is used to respond to the collaborative control signal and adjust the motor speed, pipeline valve opening and lubrication parameters of the active lubrication system according to the optimal motor speed, optimal pipeline valve opening and optimal lubrication parameters. The execution module 3 and the edge computing module 2 are linked through the Profinet bus, which can adapt to the power supply and control logic of the original system without destructive modification.

[0148] Edge computing module 2 also iteratively executes the above operations based on the feedback operating parameters.

[0149] In this embodiment, communication utilizes the Profinet industrial bus to achieve bidirectional real-time data transmission between the detection module 1, edge computing module 2, and execution module 3, with a transmission delay of ≤10ms, ensuring closed-loop control synchronization. An Ethernet interface is reserved for connection to a host computer for data monitoring, parameter calibration, and troubleshooting.

[0150] V. Closed-loop control process;

[0151] After the system is powered on, the following steps are performed to achieve full closed-loop control.

[0152] Step 1: System initialization. Edge computing module 2 loads initial PID parameters, coefficients, motor efficiency MAP, tooth surface design fatigue life, and other parameters; detection module 1, execution module 3, and communication module complete self-tests; the active lubrication system starts running according to rated lubrication parameters.

[0153] Step 2: Multi-parameter synchronous acquisition. Detection module 1 acquires operating parameters (meshing clearance, pump inlet and outlet pressure, pump outlet flow rate, pump body temperature rise, motor temperature, pump shaft torque), interference parameters (grid voltage, vibration acceleration, medium viscosity, pipeline impact peak value), and fatigue-lubrication parameters (acoustic emission signal, metal abrasive content, lubrication pressure, lubrication flow rate, lubrication oil temperature) in real time at a frequency of 10Hz, and transmits them to edge computing module 2 via the Profinet bus; at the same time, edge computing module 2 calculates the cumulative runtime.

[0154] Step 3: Interference Quantification and Fatigue Prediction. Edge computing module 2 executes the interference quantification algorithm to calculate the quantized values ​​of voltage interference, vibration interference, viscosity interference, pressure shock interference, and interference comprehensive coefficient, and determines the interference intensity;

[0155] The peak value of the acoustic emission signal is extracted by low-pass filtering. The fatigue damage degree is calculated by combining the metal abrasive content, meshing clearance and cumulative running time, and the fatigue level is determined. If the fatigue damage degree is >0.7, an early warning signal is issued through the host computer and the data is recorded.

[0156] Step 4: Active Lubrication Parameter Optimization and Adjustment. Based on the degree of fatigue damage and the comprehensive interference coefficient, the optimal lubrication parameters are calculated using an active lubrication collaborative adjustment algorithm. Control signals are output to the lubrication pump frequency converter, the micro electromagnetic proportional valve, and the electric three-way regulating valve to adjust the lubrication flow rate, pressure, and temperature to the optimal values, respectively. The adjusted lubrication parameters are collected in real time and fed back to the edge computing module 2. If the deviation is > ±5%, this step is repeated. At the same time, the adjusted lubrication parameters are synchronized to the volumetric efficiency model and the energy consumption model.

[0157] Step 5: Full-link energy consumption modeling and optimal parameter solution. Calculate real-time volumetric efficiency based on real-time meshing clearance, pump body temperature rise, viscosity interference quantification, and optimal lubrication parameters; calculate real-time full-link energy consumption based on real-time volumetric efficiency, operating parameters, interference comprehensive coefficient, and fatigue damage level; adjust and optimize weights according to the interference comprehensive coefficient and fatigue damage level, and use the gradient descent method to solve for optimal pump outlet flow, optimal motor speed, optimal pipeline valve opening, and optimal lubrication parameters.

[0158] Step 6: PID Coefficient Tuning and Multi-Segment Coordinated Adjustment. Using the optimal pump outlet flow rate as a given value, and combining real-time meshing clearance, interference comprehensive coefficient, and fatigue damage degree, calculate the tuned PID coefficients (P, I, D) according to the tuning formula. Synchronously output the tuned PID control signal to the frequency converter to adjust the motor speed to the optimal motor speed, and to the electric regulating valve to adjust the valve opening to the optimal pipeline valve opening, while maintaining lubrication parameters at their optimal values, ensuring the pump outlet flow rate accurately tracks the optimal pump outlet flow rate.

[0159] Step 7: Feedback Correction and Closed-Loop Iteration. Detection module 1 collects and feeds back the adjusted parameters to edge computing module 2 in real time. Edge computing module 2 compares the feedback values ​​with the optimal parameters. If the flow deviation is >±2%, energy consumption deviation is >±2.5%, or lubrication parameter deviation is >±5%, then steps 3 to 6 are repeated; otherwise, the current control parameters are maintained until the operating conditions change, triggering the next round of adjustment. The closed-loop iteration cycle is synchronized with the acquisition frequency at 0.1s.

[0160] VI. Specific Industrial Application Examples;

[0161] The verification was conducted in the fluid transport system of a petrochemical enterprise, using an internal gear pump model IDB-50 with designed meshing clearance, a designed tooth surface fatigue life of 10,000 hours, and transported lubricating oil with a rated viscosity of 100 mPa·s, a rated outlet pressure of 3 MPa, and a rated flow rate of 20 m³ / s. 3 / h, grid rated voltage 380V, pipeline length 15m, pipe diameter 0.1m.

[0162] The system is configured according to the above scheme: the detection module 1 is equipped with the listed sensors; the edge computing module 2 is a Siemens S7-1500 and loaded with the aforementioned preset parameters; the execution module 3 is equipped with an ABB ACS580 frequency converter, a VQ electric ball valve, an EVP proportional valve, a Delta V fatigue damage level 007 frequency converter, and a T941H three-way valve; communication uses the Profinet bus.

[0163] Under simulated high-intensity interference (interference comprehensive coefficient = 0.22) and medium fatigue (fatigue damage degree = 0.5) conditions, continuous testing for 24 hours showed that the pump outlet flow control accuracy was ±1.8%, which is better than the ±7.2% of the traditional system. Under the same process requirements, the total energy consumption of the system was 12.5kW, a 32.4% reduction compared to the 18.5kW of the traditional system, with lubrication energy consumption reduced by 40%. After 8000 hours of continuous operation, the system predicted tooth surface fatigue (fatigue damage degree = 0.72) 620 hours in advance and initiated enhanced lubrication, extending the time to macroscopic cracks on the tooth surface to 14200 hours, extending the service life by 42%, and reducing the unplanned downtime rate to 4.8%. Under grid voltage fluctuations of ±10% and environmental vibrations of 2m / s, the system performed well. 2Under the combined disturbances of viscosity mutation of 80 mPa·s and pressure shock of 4.5 MPa, the control accuracy fluctuation is ≤ ±0.3%, while the deviation of traditional systems rises to ±10.5% or even causes short-term shutdowns.

[0164] The experimental data above show that the system achieves deep coupling control of mechanical state, external disturbance and tooth surface fatigue, and achieves significant synergistic optimization in terms of control accuracy, energy saving effect, equipment life and anti-interference ability.

[0165] VII. Overall Technical Effects:

[0166] The intelligent control system for the internal gear pump in this embodiment achieves the following comprehensive technical effects by constructing a fully closed-loop intelligent control architecture that includes "mechanical state perception, external interference detection, tooth surface fatigue prediction, full-link energy consumption modeling, PID coefficient self-tuning, and multi-stage coordinated adjustment":

[0167] By deeply coupling and coordinating the control of three variables—mechanical state, external disturbance, and tooth surface fatigue—the technical challenges of traditional systems, such as decreased control accuracy with wear, incomplete energy consumption optimization, difficulty in predicting fatigue risks, and poor multi-module coordination, have been solved. This has enabled the coordinated optimization of precise control, end-to-end energy saving, and active tooth surface protection.

[0168] Example 2, the second embodiment of the present invention, provides an intelligent control system for an internal gear pump. Its edge computing module 2 integrates a multi-unit collaborative processing architecture, achieving deep coupled control of mechanical wear, external interference, and gear surface fatigue through a series of quantization models and optimization algorithms. The system hardware configuration and communication method, identical to those in Example 1, will not be repeated here. The following focuses on the specific algorithms, parameter definitions, and technical contributions of each component of the edge computing module 2.

[0169] I. Volumetric Efficiency Calculation Unit 21;

[0170] The volumetric efficiency calculation unit 21 is connected to the detection module 1 and is used to calculate the volumetric efficiency of the pump body in real time based on mechanical wear and changes in operating conditions, providing a basis for subsequent energy consumption modeling.

[0171] This unit acquires the following real-time parameters collected by the detection module: meshing clearance. (Unit: μm) Pump body temperature rise (Unit: °C) Medium viscosity (Unit: mPa·s), Lubrication flow rate (Unit: L / min), and simultaneously obtain pre-stored parameters: initial volumetric efficiency. (Dimensionless, value from 0.95 to 0.98, determined by factory testing of the pump body), preset reference lubrication flow rate. (Unit: L / min, determined by the lubrication system design value, ranging from 0.5 to 2 L / min).

[0172] The unit incorporates four experimentally calibrated correction factors: meshing clearance influence factor. (Dimensionless, ranging from 0.02 to 0.05), Temperature correction factor (Dimensionless, 0.001 to 0.003), Medium viscosity correction factor (Dimensionless, 0.005 to 0.01), Lubrication effect correction factor (Dimensionless, 0.003 to 0.008). The coefficients were obtained by fitting experimental data under different wear conditions, temperature conditions, viscosity fluctuations, and lubrication conditions.

[0173] Based on the above parameters, the unit calculates the real-time volumetric efficiency using the following formula. :

[0174] ;

[0175] In the formula: This represents the absolute value of the viscosity deviation of the medium. Rated medium viscosity (in mPa·s, provided by the medium supplier); This represents the ratio of the lubrication flow rate to the reference value.

[0176] This formula, for the first time, incorporates meshing clearance as a core wear indicator into volumetric efficiency calculation. It also integrates corrections for thermal effects, viscosity fluctuations, and lubrication conditions, resolving the technical problem of decreased calculation accuracy over time caused by traditional volumetric efficiency models that only consider inherent pump parameters and neglect mechanical wear and dynamic changes in operating conditions. The real-time volumetric efficiency calculated using this formula reflects the pump's current true mechanical state, laying a data foundation for accurate calculation of energy consumption across the entire supply chain.

[0177] II. Interference Quantization Unit 22;

[0178] The interference quantization unit 22, connected to the detection module 1, is used to quantize four types of external interference parameters into dimensionless indices that can be used for control.

[0179] This unit acquires the real-time grid voltage collected by the detection module 1. (Unit: V) Environmental vibration acceleration (unit m / s) 2 ), medium viscosity (Unit: mPa·s) Peak pressure surge in pipeline network (Unit: MPa).

[0180] Simultaneously acquire pre-stored parameters: rated grid voltage (Unit: V) Rated medium viscosity (Unit: mPa·s) Rated pressure at pump outlet (Unit: MPa), Preset maximum allowable vibration acceleration threshold (unit m / s) 2 The value ranges from 1 to 3 m / s 2 (This was determined through experiments).

[0181] This unit performs dimensionless transformation on each interference quantity according to the following formula:

[0182] Voltage interference quantization value: The value ranges from 0 to 0.2;

[0183] Vibration disturbance quantification value: The value ranges from 0 to 0.3;

[0184] Viscosity interference quantification value: The value ranges from 0 to 0.4;

[0185] Quantification of pressure shock interference: The value ranges from 0 to 0.3.

[0186] This unit obtains the preset interference weighting coefficient: voltage weighting. Vibration weight Viscosity weight Pressure weight The sum of the four is 1. In this embodiment, we take... =0.2、 =0.2、 =0.3、 =0.3, this allocation is based on the sensitivity test calibration of the impact of various disturbance factors on pump operation in the industrial field.

[0187] Interference Combination Coefficient We obtain the result by weighted summation:

[0188] ;

[0189] The value ranges from 0 to 0.32, with larger values ​​indicating stronger external interference. According to... The value can be used to classify interference levels: weak interference ( ≤0.1), medium interference (0.1< ≤0.2), strong interference ( >0.2).

[0190] This unit unifies and integrates four types of disturbances with different physical dimensions into a single comprehensive index, enabling the system to quantitatively assess the degree of disturbance from the external environment to its operation. This index serves as a core variable in subsequent energy consumption model correction, weight allocation optimization, and PID coefficient tuning, solving the problem that traditional control systems treat external disturbances as unmeasurable noise and cannot actively adapt, thus significantly improving the system's robustness under complex operating conditions.

[0191] III. Fatigue Prediction Unit 23;

[0192] The fatigue prediction unit 23 is connected to the detection module 1 and is used to integrate multi-source information to achieve early quantitative prediction of tooth surface fatigue damage.

[0193] This unit acquires the real-time meshing acoustic emission signal collected by the detection module 1. (Unit: μV), Metal abrasive content (unit: ppm), real-time meshing clearance δ (unit: μm), and cumulative running time (Unit: h). Simultaneously, obtain the following pre-stored thresholds: preset filter frequency threshold (100kHz in this embodiment), and preset acoustic emission peak threshold. (Unit: μV, range: 100 to 500 μV, calibrated through crack initiation experiments), preset maximum allowable metal abrasive content threshold. (Unit: ppm, industry standard generally uses 200ppm), preset maximum permissible meshing clearance threshold. (Unit: μm, based on design gap) This embodiment is twice as large as the previous one. =50μm, therefore =100μm), preset tooth surface design fatigue life threshold (Unit: h, provided by the pump manufacturer, typically 8000 to 12000 h).

[0194] This unit first analyzes the acoustic emission signal. Perform low-pass filtering (100kHz) and extract the peak characteristics of the filtered signal. (Unit: μV) and pulse count (optional, used for auxiliary analysis; this embodiment mainly uses peak value). Then calculate each contribution factor using the following formula:

[0195] Acoustic emission contribution factor = ;

[0196] Abrasive contribution factor = ;

[0197] Gap contribution factor = ;

[0198] Duration contribution factor = .

[0199] This unit acquires preset weighting coefficients: acoustic emission weights. Abrasive weight Gap weight Duration weight The sum of the four is 1. In this embodiment, we take... =0.4、 =0.3、 =0.2、 =0.1, this allocation is based on the experimental calibration of the sensitivity of each parameter to fatigue damage (acoustic emission is most sensitive to microcracks and has the highest weight).

[0200] Tooth surface fatigue damage degree We obtain the result by weighted summation:

[0201] ;

[0202] The value ranges from 0 to 1, with higher values ​​indicating more severe fatigue damage. According to... Fatigue levels are classified by value: weak fatigue ( ≤0.3), moderate fatigue (0.3< ≤0.7), strong fatigue ( >0.7).

[0203] This unit, for the first time, deeply integrates four types of heterogeneous data: acoustic emission (direct characterization of microcracks), metal abrasive particles (indirect characterization of wear products), meshing clearance (cumulative structural changes), and runtime (cumulative over time). This enables the prediction of tooth surface fatigue 500 to 1000 hours in advance. This indicator is not only used for early warning but also serves as a key variable to drive the dynamic adjustment of subsequent lubrication parameters and the correction of energy consumption models. It solves the technical problem that traditional methods can only perform post-event detection or offline analysis and cannot incorporate fatigue risk into real-time control.

[0204] IV. Energy Consumption Modeling Unit 24;

[0205] The energy consumption modeling unit 24 is connected to the volumetric efficiency calculation unit 21, the interference quantification unit 22, the fatigue prediction unit 23 and the detection module 1 respectively, and is used to establish a full-link energy consumption model including four sub-links: pump body, motor, pipeline network and lubrication.

[0206] This unit acquires the pump outlet pressure collected by the detection module. (Unit: MPa) Pump outlet flow rate (unit: m) 3 / h), motor speed (Unit: rpm) Variable frequency (Unit: Hz), Pipeline length (Unit: m) Pipeline diameter (Unit: m), Medium density (Unit: kg / m³) 3 Real-time lubrication pressure (Unit: MPa), Real-time lubrication flow rate (Unit: L / min).

[0207] Simultaneously acquire pre-stored parameters: preset pump body mechanical efficiency threshold. (Dimensionless, values ​​from 0.9 to 0.95), Preset motor efficiency mapping relationship (Efficiency tables of the motor at different speeds and frequencies were obtained through experimental calibration), and the threshold value of the friction coefficient along the pipeline network was preset. (Dimensionless, value from 0.01 to 0.03), preset lubrication system efficiency threshold (Dimensionless, value from 0.85 to 0.95), preset voltage interference correction coefficient (0.5 is used in this embodiment), preset pressure impact correction coefficient (In this embodiment, 0.4 is used).

[0208] The energy consumption of each sub-stage in this unit is calculated using the following formula:

[0209] ① Pump energy consumption: ;

[0210] In the formula The real-time volumetric efficiency is output by the volumetric efficiency calculation unit 21.

[0211] ② Motor energy consumption: ;

[0212] In the formula The correction term is the voltage interference quantization value output by interference quantization unit 22. Used to correct the impact of grid voltage fluctuations on motor efficiency.

[0213] ③ Pipeline network energy consumption: ;

[0214] In the formula The cross-sectional area of ​​the pipeline network, ; This is the quantified value of pressure shock interference; correction term. Used to compensate for the additional energy consumption caused by pressure shocks.

[0215] ④ Lubrication energy consumption: ;

[0216] Summing the four terms yields the total energy consumption value across the entire chain:

[0217] ;

[0218] This unit is the first to reduce the meshing gap (through...) Indirect manifestation), external interference (through) , The energy consumption model incorporates both direct correction and lubrication status (independent sub-factors) into its energy consumption model, enabling energy consumption calculations to reflect the combined effects of mechanical wear, environmental disturbances, and lubrication conditions in real time. This solves the technical problem that traditional energy consumption models suffer from decreased optimization accuracy over time due to neglecting multi-source coupling factors.

[0219] V. Energy Consumption Correction Unit 25;

[0220] The energy consumption correction unit 25 is connected to the energy consumption modeling unit 24, the interference quantification unit 22 and the fatigue prediction unit 23 respectively, and is used to introduce a global correction factor to further improve the accuracy of energy consumption calculation.

[0221] The unit acquires the end-to-end basic energy consumption value output by energy consumption modeling unit 24. and the interference comprehensive coefficient output by the interference quantization unit 22 The degree of tooth surface fatigue damage output by fatigue prediction unit 23 Simultaneously, obtain the pre-stored correction coefficients: preset interference correction coefficient threshold. (Dimensionless, values ​​range from 0.8 to 1.2, calibrated through energy consumption deviation experiments under different disturbance intensities), preset fatigue correction coefficient threshold. (Dimensionless, values ​​range from 0.9 to 1.1, calibrated through energy consumption deviation experiments at different fatigue stages).

[0222] This unit generates real-time end-to-end energy consumption according to the following formula. :

[0223] ;

[0224] In the formula, It serves as the first correction factor, used to offset the deviation of energy consumption calculation caused by external interference; This is the second correction factor, used to offset the bias in energy consumption calculation caused by tooth surface fatigue. The correction logic is: the stronger the interference and the more severe the fatigue, the smaller the correction factor, thus making the calculated energy consumption closer to the actual value.

[0225] Experiments show that, after introducing dual correction, the accuracy of energy consumption calculation can be controlled within ±2.5%. This unit solves the calculation deviation problem that may occur in the basic energy consumption model under strong disturbance or strong fatigue conditions, and provides a high-confidence objective function for subsequent optimization solutions.

[0226] VI. Optimize the solution of unit 26;

[0227] The optimization solution unit 26 is connected to the energy consumption modeling unit 24, the disturbance quantification unit 22, and the fatigue prediction unit 23, respectively, as shown in the reference. Figure 2 It includes a weight adjustment subunit 261 and a parameter optimization subunit 262.

[0228] (a) Weight adjustment subunit 261;

[0229] This sub-unit is configured to dynamically adjust the weight allocation of the optimization target based on the current disturbance intensity and fatigue level.

[0230] The subunit acquires the interference synthesis coefficient output by the interference quantization unit 22. The degree of tooth surface fatigue damage output by fatigue prediction unit 23 The first interference level threshold is pre-stored. =0.1, second interference level threshold =0.2, first fatigue level threshold =0.3, second fatigue level threshold =0.7, and at least four sets of optimization weight allocation strategies, each strategy containing energy consumption optimization weights. Control precision weight Lubrication optimization weight (The sum of the three is 1). The specific selection rules are as follows:

[0231] when When the fatigue level is >0.7 (severe fatigue), select the first strategy: =0.4, =0.3, =0.3, prioritizing fatigue protection of tooth surfaces;

[0232] when When the interference level is >0.2 (strong interference), the second strategy is selected: =0.4, =0.3, =0.3, prioritizing control accuracy;

[0233] when ≤0.3 and When the value is ≤0.1 (weak fatigue and weak interference), the third strategy is selected: =0.5, =0.3, =0.2, prioritizing optimal energy consumption;

[0234] When 0.3 < ≤0.7 and 0.1< When the value is ≤0.2 (medium operating condition), select the fourth strategy: = = =1 / 3, the three are balanced.

[0235] The aforementioned level thresholds and the weights of each strategy have all been calibrated through industrial experiments and are configurable parameters.

[0236] This sub-unit achieves adaptive adjustment of the optimization objective, enabling the system to dynamically balance energy saving, control accuracy, and fatigue protection at different operating stages, thus solving the technical problem that fixed-weight optimization cannot take into account the coordination of multiple objectives.

[0237] (ii) Parameter optimization subunit 262;

[0238] This sub-unit is connected to the weight adjustment sub-unit 261, the energy consumption modeling unit 24, and the detection module 1 respectively, and is used to solve for the optimal operating parameters under the current working conditions.

[0239] The sub-unit acquires the real-time end-to-end energy consumption model generated by the energy consumption modeling unit 24 (i.e., The calculation formula), the current operating parameters (such as pressure, flow rate, etc.) collected by the detection module 1, and the current optimized weight allocation strategy (energy consumption optimization weight) output by the weight adjustment subunit 261. Control precision weight Lubrication optimization weight ). Using pump outlet flow rate Q, motor speed n, and pipeline valve opening degree... Lubrication flow rate Lubrication pressure Lubrication temperature For the variables to be optimized, a weighted multi-objective optimization function is constructed based on the weights:

[0240] ;

[0241] in , , As a normalized baseline value, For flow control deviation, This represents the deviation of lubrication parameters from ideal values. In practical solutions, it can be directly expressed as... The objective is to minimize the gradient, and other objectives (such as control accuracy requirements) are reflected in the constraints. The weights are used to adjust the direction of gradient descent.

[0242] The unit uses gradient descent for iterative optimization: a preset learning rate threshold is set. (In this embodiment, the value is 0.01), preset convergence threshold (Set to 0.001kW), preset maximum iteration threshold (Take 100). Calculate the total energy consumption of the entire process under the current variable value in each iteration. The partial derivatives of the variable with respect to each variable are used to update the variable values ​​along the negative gradient direction. When two adjacent iterations... The difference is less than Or the number of iterations reaches The process terminates at a certain point, and the variable value at that point is determined as the optimal pump outlet flow rate. Optimal motor speed Optimal pipeline valve opening Optimal lubrication flow rate Optimal lubrication pressure Optimal lubrication temperature .

[0243] This subunit enables rapid calculation of the optimal combination of operating parameters for energy consumption under complex working conditions. The optimization process takes into account both control accuracy and lubrication requirements, providing target values ​​for subsequent PID tuning and coordinated adjustment.

[0244] VII. PID Co-control Unit 27;

[0245] The PID collaborative control unit 27 is connected to the optimization solution unit 26, the detection module 1, the interference quantization unit 22, and the fatigue prediction unit 23, respectively, as shown in the reference. Figure 3 It includes a PID tuning subunit 271 and a cooperative signal generation subunit 272.

[0246] (a) PID tuning subunit 271;

[0247] This subunit is configured to dynamically tune the proportional, integral, and derivative coefficients of the PID controller based on mechanical wear, external disturbances, and tooth surface fatigue.

[0248] This subunit acquires the optimal pump outlet flow rate determined by parameter optimization subunit 262. This serves as the control setpoint, while simultaneously acquiring the real-time meshing clearance collected by detection module 1. Interference quantization unit 22 output interference synthesis coefficient The degree of tooth surface fatigue damage output by fatigue prediction unit 23 The following threshold values, calibrated experimentally, are pre-stored:

[0249] PID initial coefficients: =3、 =0.5、 =0.05 (value range adjustable);

[0250] Design meshing gap threshold =50μm;

[0251] Gap correction factor =0.2 (dimensionless);

[0252] First interference correction coefficient =0.3, Second Interference Correction Coefficient =0.2, Third Interference Correction Coefficient =0.25;

[0253] First fatigue correction factor =0.2, Second fatigue correction factor =0.1, Third fatigue correction factor =0.15.

[0254] The tuning formula is as follows:

[0255] ;

[0256] ;

[0257] ;

[0258] The functions of each correction term in the formula are as follows:

[0259] Meshing clearance item: This reflects the deviation of the current clearance from the design value. When the clearance decreases (wear intensifies), this item is positive, making... , Increase to improve response speed Reduce the value to prevent integral saturation; conversely, increase the value to prevent integral saturation.

[0260] Interference items: The bigger , Increase to enhance anti-interference capability Reduce to avoid over-adjustment.

[0261] Fatigue item: The larger, , To address the uncertainties caused by the deterioration of the tooth surface condition, I is moderately reduced.

[0262] This sub-unit is the first to simultaneously incorporate three variables—meshing clearance, external disturbance, and tooth surface fatigue—into the PID coefficient tuning, enabling the controller to adapt to changes in mechanical wear, environmental disturbances, and structural health. This solves the problem of decreased control accuracy throughout the entire life cycle caused by traditional PID coefficients being fixed or adjusted only according to a single operating condition.

[0263] (ii) Cooperative signal generation subunit 272;

[0264] This subunit is connected to the PID tuning subunit 271 and the optimization solution unit 26 respectively, and is used to generate multi-channel coordinated control signals.

[0265] The subunit obtains the tuned coefficients output by the PID tuning subunit 271. And the optimal motor speed determined by the parameter optimization subunit. Optimal pipeline valve opening Optimal lubrication parameters (optimal lubrication flow rate) Optimal lubrication pressure Optimal lubrication temperature Simultaneously, the real-time pump outlet flow rate collected by the detection module is acquired. Calculate flow deviation .

[0266] Based on deviation The PID coefficients are then used to perform PID calculations to generate a pump flow regulation signal. :

[0267] ;

[0268] In addition, the open-loop feedforward signal is generated directly based on the optimal value:

[0269] Motor speed adjustment signal: corresponds to the optimal motor speed ;

[0270] Valve opening adjustment signal: corresponds to the optimal valve opening in the pipeline network. ;

[0271] Lubrication parameter adjustment signals include lubrication flow rate commands, lubrication pressure commands, and lubrication temperature commands, each corresponding to the optimal lubrication flow rate. Optimal lubrication pressure Optimal lubrication temperature .

[0272] The aforementioned four signals are synchronously output to execution module 3, where they are responded to by the frequency converter driver, electric regulating valve, lubrication pump frequency converter driver, miniature electromagnetic proportional valve, and electric three-way regulating valve, respectively, to achieve coordinated regulation of motor speed, pipeline valve opening, lubrication flow rate, lubrication pressure, and lubrication temperature. During the regulation process, detection module 1 continuously feeds back actual values, and edge computing module 2 determines whether to trigger the next iteration based on the deviation, forming a closed-loop control.

[0273] This unit ensures accurate tracking of the pump outlet flow rate by combining PID closed-loop and optimal feedforward. At the same time, it keeps the lubrication parameters in a state of optimal energy consumption and fatigue protection, achieving synchronous coordination of multiple execution links.

[0274] VIII. Overview of Collaborative Work Among Units;

[0275] The above-mentioned units operate collaboratively in edge computing module 2 according to the following logic:

[0276] The volumetric efficiency calculation unit 21, the interference quantification unit 22, and the fatigue prediction unit 23 process the raw data collected by the detection module 1 in parallel and output the data respectively. , , ;

[0277] Energy consumption modeling unit 24 utilizes , , and calculation of operating parameters ;

[0278] 25 pairs of energy consumption correction units Correction ;

[0279] The weight adjustment sub-unit of the optimization solution unit 26 is based on , Determine the current optimization weights, and optimize the parameters in the sub-unit. The goal is to find the optimal set of operating parameters with the objective of minimizing the minimum.

[0280] The PID tuning subunit 271 of the PID co-control unit 27 is based on , , The PID coefficients are tuned, and the coordinated signal generation subunit combines the optimal parameters to generate multiple control signals.

[0281] The execution module 3 responds to the signal to complete the adjustment, the detection module 1 feeds back the actual value, and the edge computing module 2 judges the deviation and decides whether to iterate again.

[0282] The entire process cycles in 0.1s cycles, achieving a fully closed-loop intelligent control that includes "mechanical state perception - external interference detection - tooth surface fatigue prediction - full-link energy consumption modeling - PID coefficient self-tuning - multi-link coordinated adjustment".

[0283] IX. Technical effects of Example 2;

[0284] This embodiment achieves the following technical effects through the integration of the above-mentioned modular design and algorithm:

[0285] The volumetric efficiency calculation unit 21 incorporates mechanical wear, thermal effects, medium viscosity, and lubrication status into the calculation, so that the volumetric efficiency reflects the real state of the pump body in real time, providing an accurate basis for energy consumption modeling.

[0286] Interference quantization unit 22 unifies four types of heterogeneous interference into quantification indicators, enabling the system to have the ability to actively perceive the environment.

[0287] The fatigue prediction unit 23 integrates multi-source information to identify tooth surface fatigue in advance, realize early warning and drive active lubrication.

[0288] The energy consumption modeling and correction unit constructs a full-link energy consumption model and introduces dual correction to ensure that the energy consumption calculation accuracy is within ±2.5%.

[0289] The optimization solution unit 26 dynamically adjusts the optimization weights according to the working conditions to achieve an adaptive balance between energy saving, control accuracy and fatigue protection.

[0290] The PID co-control unit 27 couples and tunes the three variables, enabling the controller to maintain high-precision control throughout its entire lifecycle.

[0291] In this embodiment, the interconnected units work together to solve the technical problems of traditional control systems, such as decreased control accuracy due to wear, incomplete energy consumption optimization, difficulty in predicting fatigue risks, and poor multi-module coordination. This achieves precise control of the internal gear pump under complex working conditions, and coordinated optimization of end-to-end energy saving and active protection against tooth surface fatigue.

[0292] Example 3 is the third embodiment of the present invention. This embodiment provides an intelligent control system for an internal gear pump, focusing on the specific structure of the execution module 3 and its working mechanism in response to the collaborative control signals of the edge computing module 2. The overall system structure, configuration of the detection module 1, and algorithm of the edge computing module 2, which are the same as in Examples 1 and 2, will not be described again here. The following focuses on the structure, function, and collaborative adjustment process of each unit of the execution module 3.

[0293] I. Overall Structure of Execution Module 3;

[0294] The execution module 3 is connected to the collaborative signal generation subunit 272 of the edge computing module 2. It is used to receive multi-channel collaborative control signals and accordingly synchronize the drive motor of the internal gear pump, the outlet pipeline valve, and the active lubrication system. The execution module 3 consists of three parts: a motor drive execution unit 31, a pipeline regulation execution unit 32, and an active lubrication execution unit 33. Each unit uses mature commercial industrial components and can be directly linked with the edge computing module 2, adapting to the power supply and control logic of the original system without destructive modification.

[0295] II. Motor drive execution unit 31;

[0296] The motor drive execution unit 31 is connected to the cooperative signal generation subunit 272 and the drive motor respectively, and is configured to acquire the motor speed adjustment signal and pump flow adjustment signal output by the cooperative signal generation subunit 272.

[0297] This unit has a built-in dedicated variable frequency drive for pumps; in this embodiment, the ABB ACS580 series variable frequency drive is selected. This variable frequency drive has the following technical characteristics: supports vector control mode, frequency adjustment range of 0 to 400Hz, response time ≤20ms, built-in PID control function, can directly receive 4 to 20mA or 0 to 10V analog control signals, and is compatible with three-phase asynchronous motors or permanent magnet synchronous motors.

[0298] The working mechanism of the motor drive execution unit 31 is as follows:

[0299] First, in response to the motor speed regulation signal, the variable frequency drive (VFD) precisely adjusts the real-time speed of the driven motor to nopt by changing the output frequency and voltage, based on the optimal motor speed command value nopt carried in the signal. This adjustment process uses a combination of open-loop feedforward and closed-loop feedback: the feedforward part directly calculates the corresponding output frequency based on nopt, while the feedback part collects the actual motor speed in real time through a built-in encoder or Hall sensor, compares it with the target value, and then fine-tunes the output to ensure that the steady-state speed accuracy is better than ±0.5%.

[0300] Secondly, in response to the pump flow rate adjustment signal, the variable frequency drive and the PID coefficients tuned by the edge computing module 2 work together to perform variable frequency closed-loop regulation of the drive motor. Specifically, the pump flow rate adjustment signal is not a direct speed command, but rather an adjustment amount that includes the result of PID calculation. After receiving this adjustment amount, the variable frequency drive adds it to the current output frequency, causing the motor speed to be dynamically adjusted according to the PID control law, thereby changing the pump outlet flow rate. At the same time, the current loop and speed loop inside the variable frequency drive remain active, ensuring smooth motor operation during dynamic adjustment.

[0301] Through the aforementioned dual adjustment mechanism, the pump outlet flow rate is consistent with the optimal pump outlet flow rate determined by the parameter optimization subunit 262. The real-time deviation between the two flows is controlled within a preset flow deviation threshold. In this embodiment, the preset flow deviation threshold is ±2%, which is calibrated through multiple step response experiments in industrial settings. This ensures control accuracy while avoiding frequent actuator movements.

[0302] The technical effect of the motor drive execution unit 31 is that it organically combines the feedforward control of the optimal motor speed with the PID-based flow closed-loop control, enabling the motor to respond quickly to changes in operating conditions (through feedforward) and accurately eliminate steady-state deviations (through closed-loop), thus solving the problem of response lag or overshoot caused by traditional frequency converter control relying solely on a single feedback loop.

[0303] III. Pipeline regulation execution unit 32;

[0304] The pipeline regulation execution unit 32 is connected to the electric regulating valve of the outlet pipeline and the coordination signal generation subunit 272 respectively, and is configured to acquire the valve opening regulation signal output by the coordination signal generation subunit 272.

[0305] The unit is equipped with a VQ series electric ball valve for electric control. Its technical parameters include: nominal diameter DN50 to DN200, working pressure PN16 to PN40, control signal 4 to 20mA, full stroke time 30 to 60s, and positioning accuracy ±1%. The valve is located in the main pipeline at the pump outlet to adjust the cross-sectional area of ​​the pipeline flow, thereby controlling the pipeline resistance characteristics and flow distribution.

[0306] The working mechanism of the pipeline regulation execution unit 32 is as follows:

[0307] This unit receives valve opening adjustment signals, which carry the optimal pipeline valve opening determined by the parameter optimization subunit 262. The servo controller built into the electric control valve analyzes this signal and drives the valve actuator (electric actuator) to adjust the valve core position to match... The corresponding opening degree. During the adjustment process, the valve's built-in position sensor provides real-time feedback on the actual opening degree, which is compared with the target value to form a closed loop, ensuring that the steady-state opening degree error is ≤±1%.

[0308] The goal of adjusting the pipe network opening is to align the pipe network resistance characteristics with the optimal motor speed. Optimal pump outlet flow rate Matching. According to fluid mechanics principles, under a given pipe network structure, the pipe network resistance is proportional to the square of the flow rate, and this proportionality coefficient is determined by the valve opening. By setting the valve opening to... This allows the pipeline resistance curve to pass exactly through the system operating point. , This ensures that the pump's output flow and pressure can overcome pipeline resistance and be delivered to the terminal, while avoiding additional throttling losses due to insufficient valve opening or flow loss due to excessive valve opening.

[0309] The technical effect of the pipeline regulation execution unit 32 is that it realizes the coordinated matching between the pipeline side and the pump source, so that the entire fluid transportation system operates at the optimal energy consumption point, and solves the problem of additional energy loss caused by the mismatch between pump and pipeline parameters in traditional systems.

[0310] IV. Active lubrication actuator 33;

[0311] The active lubrication execution unit 33 is connected to both the cooperative signal generation subunit 272 and the active lubrication system, and is configured to acquire the lubrication parameter adjustment signal output by the cooperative signal generation subunit 272. This signal includes at least the optimal lubrication flow rate adjustment command, the optimal lubrication pressure adjustment command, and the optimal lubrication temperature adjustment command.

[0312] The active lubrication actuator 33 consists of three sub-components, which correspond to the precise adjustment of lubrication flow, pressure and temperature, respectively.

[0313] (a) Lubrication pump frequency converter;

[0314] The frequency converter drive for the lubrication pump uses a Delta VFD007 series miniature frequency converter with a power range of 0.75 to 2.2 kW, a control accuracy of ±0.1 Hz, and supports Modbus communication. This drive is connected to the lubrication pump (usually a gear pump or screw pump) to adjust the pump speed, thereby controlling the oil flow rate in the lubrication circuit.

[0315] Its working mechanism is as follows: it receives the optimal lubrication flow rate adjustment command, which contains the optimal lubrication flow rate determined by the parameter optimization subunit 262. Variable frequency drive according to The target speed was calculated based on the flow-speed characteristic curve of the lubrication pump (obtained through experimental calibration), and the pump speed was adjusted to this target value by changing the output frequency. During the adjustment process, the flow sensor at the lubrication circuit outlet collected the actual flow rate in real time. The data is fed back to edge computing module 2 to form an outer closed-loop control; the speed loop inside the frequency converter driver forms an inner loop control, ensuring fast speed response and high steady-state accuracy.

[0316] (ii) Miniature electromagnetic proportional valve;

[0317] The miniature electromagnetic proportional valve uses the EVP series industrial-grade proportional valve, with a response time ≤10ms, linearity ≤±3%, control signal 0 to 10V, and pressure adjustment range 0 to 2MPa. This valve is installed on the main pipeline of the lubrication circuit for precise adjustment of the oil supply pressure of the lubrication system.

[0318] Its working mechanism is as follows: it receives the optimal lubrication pressure adjustment command, which includes the optimal lubrication pressure. The proportional valve's built-in electronic controller adjusts the valve core opening according to the command value, changing the throttling area of ​​the valve orifice, thereby regulating the lubrication circuit pressure to... Because proportional valves have extremely high response speed and can track pressure command changes in real time, they are particularly suitable for scenarios requiring rapid response to changes in tooth surface fatigue. A downstream pressure sensor collects the actual pressure in real time. It is then fed back to the edge computing module for deviation judgment and iterative adjustment.

[0319] (iii) Electric three-way regulating valve;

[0320] The electric three-way regulating valve is a T941H series electric three-way valve, with a nominal diameter of DN15 to DN50, an operating temperature range of -20 to 120℃, a control signal of 4 to 20mA, and a full stroke time of 15 to 30 seconds. This valve is located in the lubrication and cooling branch and is used to regulate the proportion of lubricating oil flowing through the cooler, thereby controlling the temperature of the lubricating medium.

[0321] Its working mechanism is as follows: it receives the optimal lubrication temperature adjustment command, which includes the optimal lubrication temperature. The electric three-way valve adjusts the valve core position according to the command value, changing the flow distribution ratio between the hot oil bypass and the cooler branch: when cooling is required, the flow ratio to the cooler is increased; when heating is required, the bypass ratio is increased. A temperature sensor at the lubricating oil tank outlet collects the actual oil temperature in real time. This feedback is then sent to edge computing module 2 to form a closed-loop control, stabilizing the oil temperature at [temperature value missing]. Nearby. During the adjustment process, temperature changes have significant inertia, therefore an incremental PID algorithm is used to avoid overshoot.

[0322] (iv) Feedback and iteration mechanism;

[0323] After completing the adjustment, the active lubrication actuator 33 will adjust the real-time lubrication flow rate. Real-time lubrication pressure Real-time lubrication temperature The detection module feeds back the feedback to the edge computing module. Upon receiving these feedback values, the edge computing module compares them with the corresponding optimal lubrication parameters to determine whether the deviation of each parameter exceeds a preset lubrication adjustment deviation threshold. In this embodiment, the preset lubrication adjustment deviation threshold is ±5%, which is calibrated through lubrication system response characteristic experiments to achieve a balance between control accuracy and actuator lifespan.

[0324] If the deviation of any lubrication parameter exceeds ±5%, the edge computing module 2 will repeatedly execute the active lubrication coordinated adjustment algorithm, recalculate the optimal lubrication parameters, and output the adjustment command again until the deviation meets the requirements. If all parameter deviations are within the allowable range, the current lubrication parameters will be maintained until the operating conditions change, triggering a new round of adjustment.

[0325] The technical advantages of the active lubrication actuator 33 are: it achieves independent and precise control of three dimensions—lubrication flow rate, pressure, and temperature—allowing the lubrication system to dynamically optimize its output parameters based on tooth surface fatigue, external disturbance intensity, and meshing clearance. Under severe fatigue conditions, the lubrication pump's frequency converter increases flow rate, the miniature proportional valve increases pressure, and the electric three-way valve reduces temperature, forming a "high flow rate - high pressure - low temperature" enhanced lubrication strategy, effectively delaying crack propagation. Under mild fatigue conditions, it reverts to an energy-saving mode of "reference flow rate - reference pressure - reference temperature," avoiding energy waste caused by over-lubrication. Experiments show that this unit reduces lubrication system energy consumption by 40% while extending tooth surface life by 42%.

[0326] V. Execution Module 3 Collaborative Workflow;

[0327] During system operation, the collaborative signal generation subunit 272 of the edge computing module 2 synchronously outputs four signals: motor speed adjustment signal, pump flow rate adjustment signal, valve opening adjustment signal, and lubrication parameter adjustment signal. The three units of the execution module 3 receive and respond to these signals in parallel, realizing synchronous adjustment of multiple links.

[0328] Taking a complete adjustment cycle as an example:

[0329] The parameter optimization subunit 262 calculates the optimal operating parameters based on the current operating conditions: =1450rpm =20.5m 3 / h、 =65%, =1.8L / min =0.7MPa =38℃.

[0330] The cooperative signal generation subunit 272 encodes these parameters into analog or digital signals and simultaneously sends them to the execution module 3 via the Profinet bus.

[0331] The motor drive actuator 31 responds within 10ms: the frequency converter increases the motor speed from the current value of 1400rpm to 1450rpm; simultaneously, it adjusts the output frequency according to the pump flow regulation signal, directing the flow to 20.5m³. 3 / h convergence.

[0332] The pipeline regulation execution unit 32 completes the valve opening adjustment within 30 seconds (limited by the mechanical action time), stabilizing the opening at 65%.

[0333] The active lubrication actuator 33 completes the lubrication parameter adjustment within 100ms: the lubrication pump frequency converter adjusts the speed to the corresponding frequency of 1.8L / min; the proportional valve adjusts the pressure to 0.7MPa; the three-way valve starts to adjust the temperature, which takes 2 to 3 minutes to stabilize at 38℃ due to thermal inertia.

[0334] During the adjustment process, each sensor continuously reports the actual values. Edge computing module 2 detected a flow deviation of +1.2% (within ±2%), a lubrication flow deviation of +3% (within ±5%), and a lubrication pressure deviation of -2% (within ±5%), all of which meet the requirements. Therefore, the current parameters are maintained without iteration.

[0335] If the flow deviation exceeds ±2% due to some disturbance, the edge computing module 2 will immediately start the next iteration, recalculate the optimal parameters and output the adjustment signal to make the system quickly return to steady state.

[0336] VI. Technical effects of Example 3;

[0337] This embodiment achieves the following technical effects through the coordinated operation of the motor drive execution unit 31, the pipeline regulation execution unit 32, and the active lubrication execution unit 33:

[0338] In this embodiment, the execution module 3, through the precise coordination of the three units of motor drive, pipeline regulation and active lubrication, accurately transforms the optimization decision of the edge computing module 2 into the action response in the physical world, enabling the internal gear pump system to maintain high precision, low energy consumption and long service life under complex working conditions, thus forming the final execution link of the fully closed-loop intelligent control circuit.

[0339] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principle of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. An intelligent control system for an internal gear pump, characterized in that, include: The detection module is used to collect the operating parameters, external interference parameters, and tooth surface fatigue-lubrication parameters of the internal gear pump in real time. The operating parameters include the meshing clearance, the external interference parameters include power grid voltage fluctuations, environmental vibrations, sudden changes in medium viscosity, and pipeline pressure shocks, and the tooth surface fatigue-lubrication parameters include tooth surface meshing acoustic emission signals, metal abrasive content, and lubrication status parameters. The edge computing module, connected to the detection module, is configured as follows: Real-time volumetric efficiency is calculated based on meshing clearance and lubrication status parameters; External interference parameters are dimensionless and weighted fused to generate comprehensive interference coefficients; The degree of tooth surface fatigue damage is calculated based on acoustic emission signals, metal abrasive content, meshing clearance, and duration. A full-link energy consumption model of pump body, motor, pipeline network and lubrication system is constructed. The meshing clearance, interference comprehensive coefficient and tooth surface fatigue damage degree are used as correction factors. Real-time full-link energy consumption is calculated by combining real-time volumetric efficiency and operating parameters. With the goal of minimizing energy consumption across the entire chain, the optimization weights are adjusted based on the interference comprehensive coefficient and the degree of fatigue damage to solve for the optimal operating parameters, which include the optimal pump outlet flow rate, optimal motor speed, optimal pipeline valve opening, and optimal lubrication parameters. Using the optimal pump outlet flow rate as a given value, the PID coefficients are tuned according to the meshing clearance, interference comprehensive coefficient, and tooth surface fatigue damage degree; A coordinated control signal is generated based on the tuned PID coefficients and the optimal operating parameters. An execution module, connected to the edge computing module, is used to respond to the collaborative control signal and adjust the motor speed, the pipeline valve opening, and the lubrication parameters of the active lubrication system according to the optimal motor speed, the optimal pipeline valve opening, and the optimal lubrication parameters. The edge computing module also iteratively executes the above operations based on the feedback operating parameters.

2. The intelligent control system for the internal gear pump according to claim 1, characterized in that, The edge computing module includes: The volumetric efficiency calculation unit, connected to the detection module, is configured as follows: The detection module acquires real-time meshing clearance, real-time pump body temperature rise, real-time medium viscosity, and real-time lubrication flow rate, as well as a preset initial volumetric efficiency and a preset reference lubrication flow rate. The meshing clearance correction amount is generated based on the real-time meshing clearance and the preset meshing clearance influence coefficient; the thermal effect correction amount is generated based on the real-time pump body temperature rise and the preset temperature correction coefficient; the medium viscosity correction amount is generated based on the absolute value of the deviation between the real-time medium viscosity and the rated medium viscosity and the preset viscosity correction coefficient; and the lubrication state correction amount is generated based on the ratio of the real-time lubrication flow rate to the preset reference lubrication flow rate and the preset lubrication effect correction coefficient. The real-time volumetric efficiency is generated by subtracting the meshing clearance correction, the thermal effect correction, the medium viscosity correction, and the lubrication state correction from the initial volumetric efficiency. Among them, the preset meshing clearance influence coefficient, the preset temperature correction coefficient, the preset viscosity correction coefficient, and the preset lubrication effect correction coefficient are all configurable thresholds calibrated through experiments.

3. The intelligent control system for the internal gear pump according to claim 2, characterized in that, The edge computing module includes: The interference quantization unit, connected to the detection module, is configured as follows: The detection module acquires real-time grid voltage, real-time environmental vibration acceleration, real-time medium viscosity, and real-time pipeline pressure impact peak value. A voltage interference quantification value is generated based on the ratio of the absolute value of the deviation between the real-time grid voltage and the rated grid voltage to the rated grid voltage. Based on the ratio of the real-time environmental vibration acceleration to the preset maximum allowable vibration acceleration threshold, a vibration interference quantification value is generated; Based on the ratio of the absolute value of the deviation between the real-time medium viscosity and the rated medium viscosity to the rated medium viscosity, a viscosity interference quantification value is generated. Based on the ratio of the real-time pipeline pressure shock peak value to the pump outlet rated pressure, a pressure shock interference quantification value is generated. Obtain preset voltage interference weighting coefficients, vibration interference weighting coefficients, viscosity interference weighting coefficients, and pressure shock interference weighting coefficients; then perform a weighted summation of the quantized values ​​of voltage interference, vibration interference, viscosity interference, and pressure shock interference to generate a comprehensive interference coefficient. Among them, the preset maximum allowable vibration acceleration threshold, the preset voltage interference weight coefficient, the preset vibration interference weight coefficient, the preset viscosity interference weight coefficient, and the preset pressure impact interference weight coefficient are all configurable thresholds calibrated through experiments, and the sum of each of the weight coefficients is a fixed threshold.

4. The intelligent control system for the internal gear pump according to claim 3, characterized in that, The edge computing module includes: The fatigue prediction unit, connected to the detection module, is configured as follows: The detection module acquires real-time meshing acoustic emission signals, real-time metal abrasive content, real-time meshing clearance, and cumulative running time. The real-time meshing acoustic emission signal is subjected to low-pass filtering below a preset filtering frequency threshold, and the peak characteristics of the filtered signal are extracted. Obtain the preset acoustic emission peak threshold, the preset maximum allowable metal abrasive content threshold, the preset maximum allowable meshing clearance threshold, and the preset tooth surface design fatigue life threshold; An acoustic emission contribution factor is generated based on the ratio of the peak characteristic to the preset acoustic emission peak threshold; an abrasive contribution factor is generated based on the ratio of the real-time metal abrasive content to the preset maximum allowable metal abrasive content threshold; a clearance contribution factor is generated based on the ratio of the real-time meshing clearance to the preset maximum allowable meshing clearance threshold; and a duration contribution factor is generated based on the ratio of the cumulative running time to the preset tooth surface design fatigue life threshold. Obtain preset acoustic emission weighting coefficients, abrasive particle weighting coefficients, gap weighting coefficients, and duration weighting coefficients. Then, perform a weighted summation on the acoustic emission contribution factor, the abrasive particle contribution factor, the gap contribution factor, and the duration contribution factor to generate the degree of tooth surface fatigue damage. Among them, the preset filter frequency threshold, the preset acoustic emission peak threshold, the preset maximum allowable metal abrasive content threshold, the preset maximum allowable meshing clearance threshold, the preset tooth surface design fatigue life threshold, and each of the weight coefficients are all configurable thresholds calibrated through experiments, and the sum of each of the weight coefficients is a fixed threshold.

5. The intelligent control system for the internal gear pump according to claim 4, characterized in that, The edge computing module includes: The energy consumption modeling unit, connected to the volumetric efficiency calculation unit, the interference quantification unit, the fatigue prediction unit, and the detection module, is configured as follows: The detection module acquires pump outlet pressure, pump outlet flow rate, motor speed, frequency converter frequency, pipeline length, pipeline diameter, medium density, real-time lubrication pressure, and real-time lubrication flow rate. Based on the pump outlet pressure, the pump outlet flow rate, and the real-time volumetric efficiency, combined with a preset pump body mechanical efficiency threshold, the energy consumption of the pump body sub-stage is calculated. Based on the energy consumption of the pump body sub-component, the motor speed, the frequency conversion frequency and the preset motor efficiency mapping relationship, combined with the voltage interference quantification value and the preset voltage interference correction coefficient, the energy consumption of the motor sub-component is calculated. Based on the pipeline length, the pipeline diameter, the medium density, the pump outlet flow rate, and the preset pipeline friction coefficient threshold, combined with the pressure shock interference quantification value and the preset pressure shock correction coefficient, the energy consumption of the pipeline sub-link is calculated. Based on the real-time lubrication pressure, the real-time lubrication flow rate, and the preset lubrication system efficiency threshold, calculate the energy consumption of the lubrication system sub-processes. The energy consumption of the pump body sub-component, the energy consumption of the motor sub-component, the energy consumption of the pipeline sub-component, and the energy consumption of the lubrication system sub-component are summed to generate the basic energy consumption value of the entire link; Among them, the preset pump body mechanical efficiency threshold, the preset voltage interference correction coefficient, the preset pipeline friction coefficient threshold, the preset pressure shock correction coefficient, the preset lubrication system efficiency threshold, and the preset motor efficiency mapping relationship are all configurable thresholds calibrated through experiments.

6. The intelligent control system for the internal gear pump according to claim 5, characterized in that, The edge computing module includes: The energy consumption correction unit, connected to the energy consumption modeling unit, the interference quantization unit, and the fatigue prediction unit respectively, is configured as follows: The energy consumption modeling unit generates the full-link basic energy consumption value, the interference quantification unit generates the interference comprehensive coefficient, and the fatigue prediction unit generates the tooth surface fatigue damage degree. Obtain the preset interference correction coefficient threshold and the preset fatigue correction coefficient threshold; A first correction factor is generated based on the interference comprehensive coefficient and the preset interference correction coefficient threshold, and a second correction factor is generated based on the degree of tooth surface fatigue damage and the preset fatigue correction coefficient threshold. The full-link basic energy consumption value is multiplied by the first correction factor and the second correction factor respectively to generate the real-time full-link energy consumption; The preset interference correction coefficient threshold and the preset fatigue correction coefficient threshold are both configurable thresholds calibrated through experiments.

7. The intelligent control system for the internal gear pump according to claim 5, characterized in that, The edge computing module includes: The optimization solution unit is connected to the energy consumption modeling unit, the interference quantification unit, and the fatigue prediction unit, respectively; the optimization solution unit includes: The weight adjustment subunit is configured as follows: Obtain the interference comprehensive coefficient generated by the interference quantization unit and the degree of tooth surface fatigue damage generated by the fatigue prediction unit; The system has a first interference level threshold, a second interference level threshold, a first fatigue level threshold, a second fatigue level threshold, and at least four sets of optimization weight allocation strategies. Each set of strategies includes at least energy consumption optimization weight, control accuracy weight, and lubrication optimization weight. The interference comprehensive coefficient is compared with the first interference level threshold and the second interference level threshold, and the tooth surface fatigue damage degree is compared with the first fatigue level threshold and the second fatigue level threshold. When the degree of fatigue damage to the tooth surface is greater than the second fatigue level threshold, a first optimization weight allocation strategy is selected, wherein the value of the lubrication optimization weight in the first optimization weight allocation strategy is greater than the values ​​of the energy consumption optimization weight and the control accuracy weight. When the interference comprehensive coefficient is greater than the second interference level threshold, a second optimization weight allocation strategy is selected, wherein the value of the control accuracy weight in the second optimization weight allocation strategy is greater than the values ​​of the energy consumption optimization weight and the lubrication optimization weight. When the degree of tooth surface fatigue damage is less than or equal to the first fatigue level threshold and the interference comprehensive coefficient is less than or equal to the first interference level threshold, a third optimization weight allocation strategy is selected. In the third optimization weight allocation strategy, the value of the energy consumption optimization weight is greater than the value of the control accuracy weight and the lubrication optimization weight. When the degree of tooth surface fatigue damage is between the first fatigue level threshold and the second fatigue level threshold, and the interference comprehensive coefficient is between the first interference level threshold and the second interference level threshold, a fourth optimization weight allocation strategy is selected, wherein the energy consumption optimization weight, the control accuracy weight, and the lubrication optimization weight are all equal in the fourth optimization weight allocation strategy. Wherein, the first interference level threshold, the second interference level threshold, the first fatigue level threshold, the second fatigue level threshold, and the weight values ​​in each of the optimized weight allocation strategies are all configurable thresholds calibrated through experiments.

8. The intelligent control system for the internal gear pump according to claim 7, characterized in that, The optimization solution unit further includes: The parameter optimization subunit, connected to the weight adjustment subunit, the energy consumption modeling unit, and the detection module, is configured as follows: The system acquires the real-time end-to-end energy consumption model generated by the energy consumption modeling unit, the current operating condition parameters collected by the detection module, and the current optimized weight allocation strategy output by the weight adjustment subunit. With the goal of minimizing energy consumption across the entire chain, and with pump outlet flow, motor speed, pipeline valve opening, and lubrication parameters of the active lubrication system as variables to be optimized, a weighted multi-objective optimization function is constructed based on the current optimization weight allocation strategy. The gradient descent method is used to iteratively optimize the variable to be optimized. Each variable is updated along the negative gradient direction with a preset learning rate threshold until the difference in the total energy consumption of two adjacent iterations is less than a preset convergence threshold or the number of iterations reaches a preset maximum number of iterations threshold. The pump outlet flow rate, motor speed, pipeline valve opening, and lubrication parameters at the end of the iteration are determined as the optimal pump outlet flow rate, optimal motor speed, optimal pipeline valve opening, and optimal lubrication parameters, respectively. The preset learning rate threshold, the preset convergence threshold, and the preset maximum number of iterations threshold are all configurable thresholds calibrated through experiments.

9. The intelligent control system for the internal gear pump according to claim 8, characterized in that, The edge computing module includes: A PID collaborative control unit is connected to the optimization solution unit, the detection module, the interference quantization unit, and the fatigue prediction unit, respectively; the PID collaborative control unit includes: The PID tuning subunit is configured as follows: The optimal pump outlet flow rate determined by the parameter optimization subunit is obtained as the control setpoint, and the real-time meshing clearance collected by the detection module, the interference comprehensive coefficient generated by the interference quantification unit, and the degree of tooth surface fatigue damage generated by the fatigue prediction unit are obtained. The system includes preset thresholds for PID initial proportional coefficient, PID initial integral coefficient, PID initial derivative coefficient, design meshing clearance, clearance correction coefficient, first disturbance correction coefficient, second disturbance correction coefficient, third disturbance correction coefficient, first fatigue correction coefficient, second fatigue correction coefficient, and third fatigue correction coefficient. Based on the deviation between the real-time meshing clearance and the designed meshing clearance threshold, the clearance correction coefficient threshold, and the PID initial proportional coefficient threshold, combined with the product of the interference comprehensive coefficient and the first interference correction coefficient threshold, and the product of the tooth surface fatigue damage degree and the first fatigue correction coefficient threshold, the tuned proportional coefficient is generated. Based on the deviation between the real-time meshing clearance and the designed meshing clearance threshold, the clearance correction coefficient threshold, and the PID initial integral coefficient threshold, combined with the product of the interference comprehensive coefficient and the second interference correction coefficient threshold, and the product of the tooth surface fatigue damage degree and the second fatigue correction coefficient threshold, the tuned integral coefficient is generated. Based on the deviation between the real-time meshing clearance and the designed meshing clearance threshold, the clearance correction coefficient threshold, and the PID initial differential coefficient threshold, combined with the product of the interference comprehensive coefficient and the third interference correction coefficient threshold, and the product of the tooth surface fatigue damage degree and the third fatigue correction coefficient threshold, the tuned differential coefficient is generated. The collaborative signal generation subunit, connected to both the PID tuning subunit and the optimization solution unit, is configured as follows: The adjusted proportional coefficient, the adjusted integral coefficient, and the adjusted differential coefficient are obtained, as well as the optimal motor speed, optimal pipeline valve opening, and optimal lubrication parameters determined by the parameter optimization subunit. Based on the deviation between the optimal pump outlet flow rate and the real-time pump outlet flow rate, PID calculation is performed using the adjusted proportional coefficient, the adjusted integral coefficient, and the adjusted derivative coefficient to generate a pump flow rate adjustment signal. Based on the optimal motor speed, optimal pipeline valve opening, and optimal lubrication parameters, motor speed adjustment signals, valve opening adjustment signals, and lubrication parameter adjustment signals are generated respectively. The pump flow rate adjustment signal, motor speed adjustment signal, valve opening adjustment signal, and lubrication parameter adjustment signal are synchronously output to the execution module. Among them, the initial proportional coefficient threshold of the PID, the initial integral coefficient threshold of the PID, the initial derivative coefficient threshold of the PID, the design meshing clearance threshold, the clearance correction coefficient threshold, the first interference correction coefficient threshold, the second interference correction coefficient threshold, the third interference correction coefficient threshold, the first fatigue correction coefficient threshold, the second fatigue correction coefficient threshold, and the third fatigue correction coefficient threshold are all configurable thresholds calibrated through experiments.

10. The intelligent control system for the internal gear pump according to claim 9, characterized in that, The execution module includes: The motor drive execution unit, connected to both the cooperative signal generation subunit and the drive motor, is configured as follows: Acquire the motor speed regulation signal and the pump flow regulation signal; The built-in pump-specific variable frequency drive responds to the motor speed adjustment signal to adjust the real-time speed of the drive motor to the optimal motor speed, and responds to the pump flow adjustment signal and the calibrated PID coefficient to perform variable frequency closed-loop regulation of the drive motor, so that the deviation between the pump outlet flow and the optimal pump outlet flow is maintained within the preset flow deviation threshold. The pipeline regulation execution unit, connected to both the coordinated signal generation subunit and the electric regulating valve of the outlet pipeline, is configured as follows: Obtain the valve opening adjustment signal; In response to the valve opening adjustment signal, the real-time opening of the electric regulating valve is adjusted to the optimal pipeline valve opening, so that the pipeline resistance characteristics are matched with the optimal motor speed and the optimal pump outlet flow rate; The active lubrication execution unit, connected to both the cooperative signal generation subunit and the active lubrication system, is configured as follows: The lubrication parameter adjustment signal is acquired, and the lubrication parameter adjustment signal includes at least the optimal lubrication flow rate adjustment command, the optimal lubrication pressure adjustment command, and the optimal lubrication temperature adjustment command. The active lubrication actuator includes: The variable frequency drive for the lubrication pump adjusts the speed of the lubrication pump to a speed corresponding to the optimal lubrication flow rate in response to the optimal lubrication flow rate adjustment command; A miniature electromagnetic proportional valve, in response to the optimal lubrication pressure adjustment command, adjusts the lubrication circuit pressure to the optimal lubrication pressure; An electric three-way regulating valve, in response to the optimal lubrication temperature regulation command, adjusts the opening of the lubrication and cooling branch to the opening corresponding to the optimal lubrication temperature; The active lubrication execution unit also feeds back the adjusted real-time lubrication flow rate, real-time lubrication pressure and real-time lubrication temperature to the edge computing module through the detection module. The edge computing module is used to determine whether the deviation between each real-time lubrication parameter and the corresponding optimal lubrication parameter exceeds the preset lubrication adjustment deviation threshold. The preset flow deviation threshold and the preset lubrication adjustment deviation threshold are both configurable thresholds calibrated through experiments.

Images