Multi-dimensional collaborative self-adaptive pneumatic conveying control system and method

Abstract

The application discloses a kind of multi-dimension collaborative self-adaptive pneumatic conveying control system and method, belong to industrial process intelligent control technical field.Method includes: obtaining gas pressure data and material flow data;Calculate pipe resistance coefficient, and calculate material compensation coefficient;Calculate target gas pressure and target material flow;Based on adjusted actual gas pressure and actual material flow, calculate pressure deviation, flow deviation and its change rate;Calculate electrostatic signal short-time energy;Judge whether to trigger self-adaptive flow pattern adjustment;Call pre-constructed fuzzy rule base, output the correction direction of pressure and flow;Using pre-trained neural network, obtain the PID control parameter of pressure regulation and flow regulation.The application realizes the promotion from "fixed parameter" to "full working condition self-adaptive", from "single-dimensional control" to "multi-dimensional coordination", enhances the stability and energy efficiency of pneumatic conveying under complex industrial environment.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for industrial processes, specifically to an intelligent control system for pneumatic units in pneumatic conveying processes, and more particularly to a multi-dimensional collaborative adaptive pneumatic conveying control system and method. Background Technology

[0002] Pneumatic conveying is a unit operation that transports powdery or granular materials through compressed gas in a closed pipeline. Based on material concentration and gas velocity, it can be divided into dilute-phase conveying (high gas velocity, low concentration) and dense-phase conveying (low gas velocity, high concentration). To balance conveying efficiency, pipeline wear, and energy consumption, modern industry tends to adopt mixed dilute-dense phase conveying, that is, dynamically adjusting the conveying mode according to operating conditions. However, achieving stable and efficient mixed conveying faces the following core challenges: 1. Rigid control targets: In traditional control strategies, the target pressure and flow rate of the dilute or dense phase are set as fixed values ​​offline. When the properties of the material batch (such as density, particle size, and moisture content) change or the pipeline resistance changes due to material accumulation or wear, the fixed target values ​​cannot guarantee the optimal conveying state, and may even cause material blockage or excessive energy consumption.

[0003] 2. Limitations of Single-Variable Control: Most existing PID controllers only adjust for pressure deviations, ignoring the system dynamics reflected by flow deviations. Pneumatic conveying is a system with strong coupling between pressure and flow; single pressure control cannot cope with flow pattern instability caused by flow fluctuations.

[0004] 3. Frequent Actuator Actions and Fixed Thresholds: To reduce valve wear, event triggering mechanisms have been introduced, but existing trigger thresholds are mostly fixed values. During dense phase transport, the inherent pressure fluctuations are large, and fixed thresholds can easily lead to excessively frequent triggering; during dilute phase transport, the fluctuations are smaller, and fixed thresholds may cause slow responses. Furthermore, triggering indicators often only consider the pressure dimension, ignoring the precursory flow pattern changes implied by flow fluctuations and electrostatic signals. Summary of the Invention

[0005] The technical problem to be solved by the present invention is: in order to solve the problems of poor adaptability and single control dimension in the existing pneumatic conveying control, the present invention provides a multi-dimensional collaborative adaptive pneumatic conveying control system and method.

[0006] The technical solution adopted by this invention to solve its technical problem is: On one hand, the present invention provides a multi-dimensional collaborative adaptive pneumatic conveying control method, comprising: S1. Acquire gas pressure data and material flow data in pneumatic conveying; S2. Calculate the pipeline resistance coefficient based on gas pressure data and material flow data, and calculate the material compensation coefficient based on material characteristics. S3. Based on the pipeline resistance coefficient and material compensation coefficient, calculate the target gas pressure and target material flow rate, and control the actuator to make preliminary adjustments to the pressure and flow rate; S4. Based on the adjusted actual gas pressure and actual material flow rate, calculate the pressure deviation, pressure deviation rate of change, flow rate deviation, and flow rate of change. S5. Obtain the real-time electrostatic sampling value in pneumatic conveying, and calculate the short-time energy of the electrostatic signal based on the real-time electrostatic sampling value; S6. Compare the pressure deviation change rate, flow deviation change rate, and short-time energy of electrostatic signal with the preset pressure deviation change rate threshold, preset flow deviation change rate threshold, and electrostatic fluctuation energy threshold, respectively, to determine whether adaptive flow pattern adjustment is triggered. If yes, proceed to step S7; otherwise, maintain the control command of step S3. S7. Based on pressure deviation, pressure deviation change rate, flow deviation, and flow deviation change rate, call the pre-built fuzzy rule base and output the correction direction of pressure and flow. S8. Input the pressure deviation, pressure deviation change rate, flow deviation, and flow deviation change rate into the pre-trained neural network to obtain the PID control parameters for pressure regulation and flow regulation. S9. Based on the correction direction and PID control parameters, the PID controller calculates and outputs the adjustment command of the actuator, and returns to step S1.

[0007] It avoids system oscillations caused by frequent adjustments and can quickly respond to changes in flow patterns, significantly improving the stability, energy efficiency, and material adaptability of the pneumatic conveying process, and preventing pipe blockage and energy waste.

[0008] Furthermore, the equation for solving the pipeline resistance coefficient is as follows: ; ; In the formula, For the pressure drop of the gas-solid two-phase flow in the main pipeline; This represents the current pipeline resistance coefficient. The total equivalent horizontal length; The inner diameter of the pipe; The density of the gas; Apparent gas velocity; Add a resistance coefficient to the material; This refers to the material-to-gas mixing ratio; This is the actual length of the horizontal pipe; This refers to the vertical height of the pipe. This represents the number of pipe bends.

[0009] Taking into account the pipeline geometry (horizontal, vertical, elbows), gas-solid two-phase flow characteristics, and additional material resistance, it can accurately and in real-time calculate the current pipeline resistance coefficient. This provides a reliable physical basis for subsequent calculations and avoids control bias caused by empirical assumptions.

[0010] Furthermore, in step S2, the material compensation coefficient is calculated based on the material characteristics, including: The bulk density, median particle size, and moisture content of the material are obtained through an online material parameter detection unit. The bulk density, median particle size, and moisture content are then input into a pre-trained feedforward neural network, which outputs pressure compensation coefficient and flow compensation coefficient.

[0011] It can dynamically compensate for the influence of different material properties on pressure and flow control targets, so that the system can maintain the optimal target setpoint when conveying multiple materials or when the material properties fluctuate, thus enhancing the robustness and versatility of the control method.

[0012] Furthermore, the formula for calculating the target gas pressure is as follows: ; The formula for calculating the target material flow rate is: ; In the formula, The target gas pressure; Reference gas pressure; This is the pressure compensation coefficient; This represents the current pipeline resistance coefficient. The initial resistance coefficient of the clean pipeline; , This is a resistance index; For the target material flow rate; Reference material flow rate; This is the flow compensation coefficient.

[0013] Based on the resistance of clean pipelines, and combined with the real-time pipeline resistance coefficient and material compensation coefficient, the system dynamically generates target pressure and flow rate that match the current operating conditions. This ensures both the conveying capacity and avoids energy waste and pipeline wear caused by excessive pressure or flow, achieving a balance between energy saving and reliability.

[0014] Furthermore, the calculation formula for the preset pressure deviation change rate threshold is as follows: ; ; In the formula, The preset pressure deviation change rate threshold is used; The threshold for the rate of change of the reference pressure deviation; This is the sensitivity coefficient; The pressure variation coefficient; The standard deviation of pressure; This represents the average pressure.

[0015] By introducing a pressure variation coefficient to adaptively adjust the threshold, the threshold automatically widens when pressure fluctuations are large (such as when conveying unstable materials), avoiding false triggering of adjustments; and tightens when the pressure is stable, improving control response sensitivity. This mechanism effectively reduces the redundancy of adjustment actions and enhances the system's anti-interference capability.

[0016] Furthermore, the formula for calculating the preset flow deviation change rate threshold is as follows: ; CV Q = σ Q / ; In the formula, The preset threshold for the rate of change of flow deviation; The threshold for the rate of change of the baseline flow rate; This is the sensitivity coefficient; Here is the coefficient of variation of flow rate; The standard deviation of the flow rate; This represents the average pressure.

[0017] Furthermore, the formula for calculating the electrostatic fluctuation energy threshold is as follows: ; In the formula, The threshold value for electrostatic fluctuation energy; The baseline electrostatic fluctuation energy threshold; This represents the average pressure inside the pipe. For reference pressure; This is for the safety factor.

[0018] The threshold is increased at high pressure to avoid adjustments triggered by weak electrostatic fluctuations; the threshold is decreased at low pressure to improve the detection sensitivity of flow pattern changes during dilute phase transport.

[0019] Furthermore, step S7 includes: S71. Map the four variables—pressure deviation, pressure deviation rate of change, flow deviation, and flow deviation rate of change—to a unified map. The fuzzy domain of discourse; S72. Define multiple fuzzy subsets for the four variables: pressure deviation, pressure deviation rate of change, flow deviation, and flow deviation rate of change. S73. Using Gaussian membership functions, calculate the membership degrees of pressure deviation, pressure deviation rate of change, flow deviation, and flow deviation rate of change to each fuzzy subset for use in the fuzzy rule base. S74. Based on the membership degree, call the pre-built fuzzy rule base and output the correction direction of pressure and flow.

[0020] By using a unified fuzzy domain mapping and Gaussian membership function, the pressure and flow deviations and their rates of change are smoothly fuzzified. Combined with a fuzzy rule base constructed from expert knowledge, it can output correction directions that conform to human operational experience (such as "increase pressure, decrease flow"), thereby improving the rationality of control decisions.

[0021] On the other hand, the present invention provides a pneumatic conveying control system, comprising: The data acquisition layer is used to collect gas pressure data, material flow data, and real-time electrostatic sampling values ​​in pneumatic conveying. The intelligent control layer includes a data preprocessing module, an adaptive target generation module, a multi-input fuzzy neural network PID controller, and an adaptive event triggering module, wherein: The data preprocessing module is used to preprocess the data collected by the data acquisition layer. The adaptive target generation module is used to calculate the target gas pressure and target material flow rate based on gas pressure data, material flow rate data and material characteristics. The multi-input fuzzy neural network PID controller is used to output pressure-flow coordinated control commands based on pressure deviation, pressure deviation change rate, flow deviation and flow deviation change rate. The adaptive event triggering module is used to compare the pressure deviation change rate, flow deviation change rate and short-time energy of electrostatic signal with the corresponding thresholds to determine whether to trigger the multi-input fuzzy neural network PID controller. The actuator layer includes electric regulating valves, frequency converters, and quick-release valves.

[0022] It adopts a three-layer architecture consisting of a data acquisition layer, an intelligent control layer, and an execution drive layer. It has an adaptive target generation and event triggering mechanism, and the multi-input fuzzy neural network PID controller is only started when the triggering conditions are met, thus avoiding the resource consumption caused by continuous PID controller calculations.

[0023] Furthermore, the intelligent control layer also includes a blockage warning and self-healing module, which is used to identify the blockage condition based on gas pressure data and control the system to perform blockage clearing actions, including opening a quick pressure relief valve.

[0024] With the addition of a blockage warning and self-healing module, the system can not only proactively adjust to prevent pipe blockage, but also automatically perform rapid pressure relief and unblocking when initial blockage occurs, eliminating the accident in its infancy. This significantly reduces the time cost and material loss associated with manual pipe clearing, and improves the reliability and safety of continuous operation of the pneumatic conveying system. Attached Figure Description

[0025] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0026] Figure 1 This is an overall flowchart of the multi-dimensional collaborative adaptive pneumatic conveying control method of the present invention; Figure 2 This is a flowchart of step S7 of the pneumatic conveying control method of the present invention; Figure 3 This is a schematic diagram of the pneumatic conveying control system of the present invention. Detailed Implementation

[0027] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0028] Example 1: like Figure 1 As shown, this embodiment provides a multi-dimensional collaborative adaptive pneumatic conveying control method, including: S1. Acquire gas pressure and material flow data in pneumatic conveying. Specifically, gas pressure data at multiple points along the axial direction of the conveying pipeline is collected through a pressure transmitter. Multi-point pressure is monitored collaboratively to form a pressure distribution curve along the conveying process. The material flow rate is measured in real time through a material flow meter.

[0029] Preferably, before proceeding to subsequent steps, the collected gas pressure data and material flow rate data are filtered for noise reduction, normalized, and divided into time windows. A sliding time window method is used to divide the continuous time-series signal into input sample frames of fixed length.

[0030] S2. Based on gas pressure data and material flow rate data, calculate the pipeline resistance coefficient, and based on material characteristics, calculate the material compensation coefficient, which includes the pressure compensation coefficient. and flow compensation coefficient This design can automatically adapt to batch differences in materials and gradual changes in resistance caused by long-term pipeline operation, achieving dynamic optimization of the target value. In actual cement plant fly ash conveying systems, conveying efficiency is increased by 12%~18%, and unit conveying energy consumption is reduced by 8%~15%.

[0031] Specifically, the equation for solving the pipe resistance coefficient is: ; ; In the formula, For the pressure drop of the gas-solid two-phase flow in the main pipeline; This represents the current pipeline resistance coefficient. The total equivalent horizontal length refers to the total length of the actual pipeline (including horizontal, vertical, and elbows) converted into the equivalent horizontal straight pipe with the same resistance loss. The inner diameter of the pipe; The density of the gas; Apparent gas velocity; The additional drag coefficient for materials represents the additional drag caused by the presence of solid particles; This refers to the material-to-gas mixing ratio; This is the actual length of the horizontal pipe; This refers to the vertical height of the pipe. This represents the number of pipe bends.

[0032] Of the parameters mentioned above, besides , , Other parameters can be calculated using production line instruments, among which, , For solid mass flow rate, Let be the gas mass flow rate. Therefore, in the above formula for the pressure drop of the gas-solid two-phase flow in the main pipeline, , Treating them as variables, pressure drop in two gas-solid two-phase flows was collected on-site to construct a model for... , The system of two equations can be solved. , The value of .

[0033] Specifically, based on material characteristics, the material compensation coefficient is calculated, including the following steps: An online material parameter detection unit is configured at the inlet of the feeding unit to obtain the material bulk density. Median particle size and moisture content The bulk density of the material Median particle size and moisture content The input to a pre-trained feedforward neural network (or multivariate regression model) outputs a pressure compensation coefficient. and flow compensation coefficient The network training data comes from historical data from a laboratory pneumatic conveying test platform, or simulation data generated based on classic pneumatic conveying pressure drop models (such as the Barth additional pressure drop model). For example, when the material's moisture content increases, leading to enhanced adhesion, Increase (appropriately increase the conveying pressure to overcome adhesion). Reduce (lower the feed rate to prevent blockage); as particle size increases, the critical fluidization velocity increases. It needs to be adjusted upwards to maintain suspension.

[0034] S3. Based on the pipeline resistance coefficient and material compensation coefficient, calculate the target gas pressure and target material flow rate, and control the actuator to make preliminary adjustments to the pressure and flow rate according to the above target values.

[0035] Specifically, the formula for calculating the target gas pressure is: ; The formula for calculating the target material flow rate is: ; In the formula, The target gas pressure; The reference gas pressure is (preferably 0.3~0.6MPa). This is the pressure compensation coefficient; This represents the current pipeline resistance coefficient. The initial resistance coefficient of the clean pipeline (preferably 0.02~0.03); , The resistance index is preferably 0.3 to 0.6, which reflects the engineering strategy of increasing pressure compensation and appropriately reducing flow rate when resistance increases; For the target material flow rate; The reference material flow rate is (preferably 10~30t / h). This is the flow compensation coefficient.

[0036] This step implements a "one-item-one-policy, real-time matching" strategy, which is particularly suitable for multi-variety, small-batch production scenarios (such as specialty chemicals and pharmaceuticals). In the future, it can be further integrated with a Manufacturing Execution System (MES) to directly receive material batch information and achieve fully automated intelligent generation of control targets.

[0037] S4. Calculate the pressure deviation based on the adjusted actual gas pressure and actual material flow rate. Pressure deviation change rate Flow deviation and rate of change of flow deviation .

[0038] Wherein, pressure deviation is the difference between the adjusted actual gas pressure and the target gas pressure, and flow deviation is the difference between the adjusted actual material flow rate and the target material flow rate; the rate of change of pressure deviation is the rate of change of pressure deviation over time, and the rate of change of flow deviation is the rate of change of flow deviation over time.

[0039] S5. Collect real-time electrostatic sampling values ​​during pneumatic conveying using an electrostatic sensor, and calculate the short-time energy of the electrostatic signal based on the real-time electrostatic sampling values. .

[0040] The electrostatic sensor outputs a broadband AC signal. Static electricity is generated when particles rub against the pipe wall as they flow through the pipe. In a dilute phase, the particle concentration is low, and the electrostatic signal energy is weak; in a dense phase with stagnant flow, the particle clusters rub intensely against the pipe wall, resulting in a strong electrostatic signal energy. The short-time energy of the electrostatic signal is defined as follows: In the formula, For the first The sampled values ​​of the electrostatic sensor This represents the number of sampling points (e.g., 1000 points correspond to 1 second of data, with a sampling rate of 1kHz).

[0041] S6, Pressure Deviation Change Rate Flow deviation change rate Short-time energy of electrostatic signals The values ​​are compared with the preset pressure deviation change rate threshold, the preset flow deviation change rate threshold, and the electrostatic fluctuation energy threshold to determine whether adaptive flow pattern adjustment is triggered. If yes, the process jumps to step S7; otherwise, the control command of step S3 is retained.

[0042] The core of this design lies in balancing control performance and actuator lifespan. Furthermore, a fixed threshold cannot adapt to all operating conditions of pneumatic conveying: in dense phase conveying, pressure fluctuations are large, and setting the threshold too low will cause frequent valve vibration; in dilute phase conveying, fluctuations are weak, and setting the threshold too high will cause the control system to become "sluggish" and unable to suppress disturbances in a timely manner. In addition, a single pressure deviation change rate index ignores the flow pattern transformation information implied by material flow fluctuations and electrostatic characteristics. Therefore, a three-dimensional adaptive threshold system and collaborative judgment logic are proposed. The threshold self-tunes online with the pressure variation coefficient and flow variation coefficient, achieving intelligent adjustment characteristics that are "sensitive when operating conditions are stable and tolerant when operating conditions fluctuate."

[0043] Specifically, the preset pressure deviation change rate threshold is adaptively adjusted according to the intensity of pressure fluctuation, and the calculation formula is as follows: ; ; In the formula, The preset pressure deviation change rate threshold is used; The threshold value for the rate of change of the reference pressure deviation is obtained through experimental calibration and is preferably 0.05~0.2MPa / s; This is the sensitivity coefficient, set according to the material conveying type (0.5~1.0 for viscous materials, and 0.2~0.5 for materials with good flowability). The standard deviation of pressure; Average pressure; The pressure variation coefficient is used to characterize the intensity of local pressure fluctuations. It can be calculated by selecting pressure values ​​from approximately N cycles (e.g., 20 cycles). In practical applications, during dense phase transport... It can reach 0.1~0.2, at this point Automatic relaxation to avoid frequent triggering due to high-frequency, low-amplitude fluctuations; during dilute phase transport, As low as 0.02~0.05, Keep the value low to ensure a timely response to external disturbances.

[0044] Specifically, the formula for calculating the preset flow deviation change rate threshold is as follows: ; CV Q = σ Q / ; In the formula, The preset threshold for the rate of change of flow deviation; The threshold for the rate of change of the baseline flow deviation is obtained through experimental calibration, and is preferably 0.5~2 (t / h) / s; As a sensitivity coefficient, in actual operation, flow rate fluctuations are often more sensitive than pressure fluctuations, therefore Typically set slightly larger For example, the value can be 1.1; Here is the coefficient of variation of flow rate; The standard deviation of the flow rate; This represents the average pressure.

[0045] Specifically, the formula for calculating the electrostatic fluctuation energy threshold is: ; In the formula, The threshold value for electrostatic fluctuation energy; The baseline electrostatic fluctuation energy threshold; This represents the average pressure inside the pipe. For reference pressure; For a safety factor, such as 1.3.

[0046] During the debugging phase, the system was set to "learning mode" and ran for 5 minutes each in a stable dilute phase state and a stable dense phase state, as shown in Table 1. The short-time energy of the electrostatic signal in the two stable states was recorded respectively. and electrostatic standard deviation and according to The reference electrostatic fluctuation energy thresholds for stable dilute phase and stable dense phase states were obtained, respectively. .

[0047] Table 1

[0048] In some feasible implementations, , ,and The values ​​of are shown in Table 2.

[0049] Table 2

[0050] The adaptive flow pattern adjustment trigger logic is shown in Table 3: Table 3

[0051] This logic can be expressed as follows: when both pressure and flow rate deviations exceed their limits, a direct trigger is initiated (indicating a real external disturbance to the system); when only one of the pressure or flow rate deviations exceeds its limit, triggering is only initiated if electrostatic energy also exceeds its limit (indicating a change in the flow pattern); if only electrostatic energy exceeds its limit, triggering is not initiated (this may be due to particle electrostatic accumulation rather than a change in the flow state). This effectively filters out false triggering caused by single-channel random noise. Furthermore, the program includes a safety exception: if any rate of change exceeds twice its threshold, unconditional triggering is initiated to handle sudden strong interference. In practical applications, the operating frequency of actuators (electric regulating valves, frequency converters) is reduced by 35% to 50%, control performance is not compromised, and the expected valve lifespan is extended by more than three times.

[0052] S7. Based on pressure deviation, pressure deviation change rate, flow deviation, and flow deviation change rate, call the pre-built fuzzy rule base and output the correction direction for pressure and flow.

[0053] Specifically, such as Figure 2 As shown, step S7 includes: S71, Adjust pressure deviation Pressure deviation change rate Flow deviation and rate of change of flow deviation Four variables are uniformly mapped to The fuzzy domain of discourse.

[0054] S72, Pressure Deviation Pressure deviation change rate Flow deviation and rate of change of flow deviation Each variable defines multiple fuzzy subsets, such as NB (negative large), NM (negative medium), NS (negative small), ZE (zero), PS (positive small), PM (positive medium), and PB (positive large).

[0055] S73. Using Gaussian membership functions, the membership degrees of pressure deviation, pressure deviation rate of change, flow deviation, and flow deviation rate of change to each fuzzy subset are calculated for use in the fuzzy rule base. The Gaussian membership function centers are not uniformly distributed, but are denser near zero to improve steady-state accuracy.

[0056] S74. Based on membership, call the pre-built fuzzy rule base and output the correction direction of pressure and traffic.

[0057] Specifically, the fuzzy rule base is based on the fusion of two aspects of knowledge: 1. Physical mechanism analysis: Starting from the gas-solid two-phase flow phase diagram (Zenz phase diagram) and transport characteristics, the coupling law between pressure and flow rate is summarized. For example, dense phase thrombosis requires high pressure and low velocity, while dilute phase suspension requires low pressure and high velocity; the precursor to pipe blockage is a rise in pressure and a sudden drop in flow rate. 2. Operational expert experience: The adjustment strategies of skilled operators are summarized, such as "when both pressure and flow rate are low, prioritize increasing pressure and gas supply," and "when pressure is normal but flow rate is high, slightly reduce pressure and reduce feed." These two aspects are integrated to form a preliminary fuzzy rule base and to establish the synergistic influence law between pressure and flow rate, thereby further guiding the changes in adaptive operating parameters in practical engineering applications. The total number of rules is 7. 4 =2401 rules, after simplification using symmetry and screening by expert experience, about 200 core rules were retained, covering all typical working conditions.

[0058] S8. Input the pressure deviation, pressure deviation change rate, flow deviation, and flow deviation change rate into the pre-trained neural network to obtain the PID control parameters for pressure regulation and flow regulation.

[0059] Specifically, a single-hidden-layer BP neural network is constructed, with 4 neurons in the input layer ( , , , The network consists of 15 hidden layers (using the ReLU activation function) and 6 output layers, corresponding to KpP (proportional coefficient of the pressure control loop), KiP (integral coefficient of the pressure control loop), KdP (differential coefficient of the pressure control loop), KpQ (proportional coefficient of the flow control loop), KiQ (integral coefficient of the flow control loop), and KdQ (differential coefficient of the flow control loop), respectively. The output layer activation function is a constrained sigmoid. Online training of the network weights uses gradient descent with a momentum term. The objective function is defined as: ; In the formula, The weights represent the material flow error, reflecting the degree of importance attached to material flow fluctuations. Training samples are derived from critical operating condition data stored in the current control cycle and historical data, with 5-10 iterations per cycle. This network complements the fuzzy rule base: fuzzy rules provide experience-based coarse-grained adjustments, while the neural network provides data-based fine-tuning, balancing robustness and accuracy.

[0060] S9. Based on the correction direction and PID control parameters provided by the fuzzy rule base, the PID controller calculates and outputs the adjustment command of the actuator, and returns to step S1.

[0061] Specifically, the pressure channel control output is sent to the electric regulating valve to change the air supply pressure; the flow channel control output is sent to the frequency converter to change the speed of the compressed air source, thereby achieving flow regulation.

[0062] By introducing flow deviation, this design expands the control domain of the PID controller from a one-dimensional pressure line to a two-dimensional pressure-flow plane. When the system experiences temporary signs of blockage (pressure has not yet increased significantly, but flow has decreased significantly), a traditional pressure PID controller will not activate, while this system will detect the deviation. When the flow rate is NB (indicating severely low flow), the corresponding rule in the fuzzy rule base is immediately triggered to increase the gas flow rate in advance, eliminating the risk of blockage before the pressure rises sharply. Simulations and experiments show that, under conditions of sudden changes in material moisture content, increasing the flow input in the control loop can reduce pressure overshoot by 40%, shorten response time by 35%, and improve delivery stability by 28%.

[0063] Example 2: like Figure 3 As shown, this embodiment provides a multi-dimensional collaborative adaptive pneumatic conveying control system, including: The pneumatic unit assembly layer includes a compressed air source unit, a feeding unit, a gas-solid mixing unit, a conveying pipeline, and a separation unit. Specifically, the compressed air source unit uses a variable frequency drive air compressor or a Roots blower to provide stable compressed air. The feeding unit includes a hopper and a rotary feeder, used to feed powdery or granular materials into the gas-solid mixing unit at a controllable rate. The gas-solid mixing unit (such as an ejector or mixing chamber) thoroughly mixes the compressed air with the material before sending it into the conveying pipeline. The conveying pipeline connects each unit, and a separation unit (such as a cyclone separator or bag filter) is installed at the end to achieve gas-solid separation.

[0064] The data acquisition layer includes an array of pressure transmitters, material flow meters, and electrostatic sensors deployed at multiple points along the axial direction of the conveying pipeline. These sensors are used to collect gas pressure data, material flow data, and real-time electrostatic sampling values ​​during pneumatic conveying. All data is uploaded to the intelligent control layer in real time via industrial Ethernet or fieldbus.

[0065] Specifically, the pressure transmitter array includes at least the following monitoring points: a pressure monitoring point on the main gas supply line; a pressure monitoring point at the outlet of the gas-solid mixing unit; a pressure monitoring point in the middle of the delivery pipeline; and a pressure monitoring point at the inlet of the terminal separation unit. The main gas supply line is the main gas supply pipeline connecting the outlet of the compressed gas source unit and the inlet of the gas-solid mixing unit. Through multi-point pressure collaborative monitoring, the pressure distribution curve along the delivery process can be obtained, providing key data support for subsequent control decisions.

[0066] Specifically, the material flow meter is installed at the outlet of the gas-solid mixing unit or at the beginning of the conveying pipeline to measure the material flow rate in real time. The electrostatic sensor is installed in the middle section of the conveying pipeline to reflect the flow pattern characteristics by using the electrostatic fluctuation signal generated by the friction between particles and the pipe wall in the gas-solid two-phase flow.

[0067] The intelligent control layer includes a data preprocessing module, an adaptive target generation module, a multi-input fuzzy neural network PID controller, and an adaptive event triggering module, wherein: The data preprocessing module is used to preprocess the data acquired by the data acquisition layer; the adaptive target generation module is used to calculate the target gas pressure and target material flow rate based on gas pressure data, material flow rate data, and material characteristics; the multi-input fuzzy neural network PID controller is used to output pressure-flow coordinated control commands based on pressure deviation, pressure deviation change rate, flow deviation, and flow deviation change rate; the adaptive event triggering module is used to compare the pressure deviation change rate, flow deviation change rate, and short-time energy of the electrostatic signal with the corresponding thresholds to determine whether to trigger the multi-input fuzzy neural network PID controller.

[0068] The execution drive layer includes an electric regulating valve and a variable frequency drive. The electric regulating valve receives gas pressure regulation commands from the intelligent control layer and adjusts the valve opening in real time to control the pressure of the main gas supply pipe. The variable frequency drive adjusts the motor speed of the compressed air source unit to achieve continuous adjustment of the compressed air flow over a wide range.

[0069] Example 3 Unlike Embodiment 2, the intelligent control layer of this embodiment also includes a blockage warning and self-healing module, and the execution drive layer also includes a rapid pressure relief valve. The blockage warning and self-healing module is used to identify the blockage condition based on gas pressure data and control the system to perform blockage clearing actions.

[0070] Specifically, when the multi-point pressure transmitter array detects that the pressure at the end of the pipeline continues to rise above the warning threshold (e.g., 0.4 MPa) while the pressure at the beginning of the pipeline decreases synchronously, it is determined to be a blockage condition. The blockage warning and self-healing module immediately triggers the following sequence of actions: First, it opens the quick-release valve to release the back pressure in the pipeline and prevent further pressure buildup; second, it increases the air supply flow rate of the compressed air source unit to a preset multiple (e.g., 2 times) of the maximum rated value, using high-speed airflow to impact the blocked section for self-healing unblocking; after unblocking, the system returns to the original conveying mode. If the self-healing action fails to resolve the blockage within a specified time (e.g., 10 seconds), an alarm signal is issued to prompt manual intervention.

[0071] This design can identify risks four cycles (approximately 8 seconds) before a complete blockage occurs, reducing the blockage rate by more than 70%, eliminating unplanned downtime caused by complete blockage, ensuring production continuity, and avoiding severe pressure shocks to the pipeline.

[0072] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A multi-dimensional collaborative adaptive pneumatic conveying control method, characterized in that, include: S1. Acquire gas pressure data and material flow data in pneumatic conveying; S2. Calculate the pipeline resistance coefficient based on gas pressure data and material flow data, and calculate the material compensation coefficient based on material characteristics. S3. Based on the pipeline resistance coefficient and material compensation coefficient, calculate the target gas pressure and target material flow rate, and control the actuator to make preliminary adjustments to the pressure and flow rate; S4. Based on the adjusted actual gas pressure and actual material flow rate, calculate the pressure deviation, pressure deviation rate of change, flow rate deviation, and flow rate of change. S5. Obtain the real-time electrostatic sampling value in pneumatic conveying, and calculate the short-time energy of the electrostatic signal based on the real-time electrostatic sampling value; S6. Compare the pressure deviation change rate, flow deviation change rate, and short-time energy of electrostatic signal with the preset pressure deviation change rate threshold, preset flow deviation change rate threshold, and electrostatic fluctuation energy threshold, respectively, to determine whether adaptive flow pattern adjustment is triggered. If yes, proceed to step S7; otherwise, maintain the control command of step S3. S7. Based on pressure deviation, pressure deviation change rate, flow deviation, and flow deviation change rate, call the pre-built fuzzy rule base and output the correction direction of pressure and flow. S8. Input the pressure deviation, pressure deviation change rate, flow deviation, and flow deviation change rate into the pre-trained neural network to obtain the PID control parameters for pressure regulation and flow regulation. S9. Based on the correction direction and PID control parameters, the PID controller calculates and outputs the adjustment command of the actuator, and returns to step S1.

2. The multi-dimensional collaborative adaptive pneumatic conveying control method according to claim 1, characterized in that, The equation for solving the pipeline resistance coefficient is as follows: ； ； In the formula, For the pressure drop of the gas-solid two-phase flow in the main pipeline; This represents the current pipeline resistance coefficient. The total equivalent horizontal length; The inner diameter of the pipe; The density of the gas; Apparent gas velocity; Add a resistance coefficient to the material; This refers to the material-to-gas mixing ratio; This is the actual length of the horizontal pipe; This refers to the vertical height of the pipe. This represents the number of pipe bends.

3. The multi-dimensional collaborative adaptive pneumatic conveying control method according to claim 1, characterized in that, In step S2, the material compensation coefficient is calculated based on the material characteristics, including: The bulk density, median particle size, and moisture content of the material are obtained through an online material parameter detection unit. The bulk density, median particle size, and moisture content are then input into a pre-trained feedforward neural network, which outputs pressure compensation coefficient and flow compensation coefficient.

4. The multi-dimensional collaborative adaptive pneumatic conveying control method according to claim 1, characterized in that, The formula for calculating the target gas pressure is: ； The formula for calculating the target material flow rate is: ； In the formula, The target gas pressure; Reference gas pressure; This is the pressure compensation coefficient; This represents the current pipeline resistance coefficient. The initial resistance coefficient of the clean pipeline; , This is a resistance index; For the target material flow rate; Reference material flow rate; This is the flow compensation coefficient.

5. The multi-dimensional collaborative adaptive pneumatic conveying control method according to claim 1, characterized in that, The formula for calculating the preset pressure deviation change rate threshold is as follows: ； ； In the formula, The preset pressure deviation change rate threshold is used; The threshold for the rate of change of the reference pressure deviation; This is the sensitivity coefficient; The pressure variation coefficient; The standard deviation of pressure; This represents the average pressure.

6. The multi-dimensional collaborative adaptive pneumatic conveying control method according to claim 1, characterized in that, The formula for calculating the preset flow deviation change rate threshold is as follows: ； CV Q = σ Q / ； In the formula, The preset threshold for the rate of change of flow deviation; The threshold for the rate of change of the baseline flow rate; This is the sensitivity coefficient; Here is the coefficient of variation of flow rate; The standard deviation of the flow rate; This represents the average pressure.

7. The multi-dimensional collaborative adaptive pneumatic conveying control method according to claim 1, characterized in that, The formula for calculating the electrostatic fluctuation energy threshold is as follows: ； In the formula, The threshold value for electrostatic fluctuation energy; The baseline electrostatic fluctuation energy threshold; This represents the average pressure inside the pipe. For reference pressure; This is for the safety factor.

8. The multi-dimensional collaborative adaptive pneumatic conveying control method according to claim 1, characterized in that, Step S7 includes: S71. Map the four variables—pressure deviation, pressure deviation rate of change, flow deviation, and flow deviation rate of change—to a unified map. The fuzzy domain of discourse; S72. Define multiple fuzzy subsets for the four variables: pressure deviation, pressure deviation rate of change, flow deviation, and flow deviation rate of change. S73. Using Gaussian membership functions, calculate the membership degrees of pressure deviation, pressure deviation rate of change, flow deviation, and flow deviation rate of change to each fuzzy subset for use in the fuzzy rule base. S74. Based on the membership degree, call the pre-built fuzzy rule base and output the correction direction of pressure and flow.

9. A pneumatic conveying control system applying the multi-dimensional collaborative adaptive pneumatic conveying control method according to any one of claims 1 to 8, characterized in that, include: The data acquisition layer is used to collect gas pressure data, material flow data, and real-time electrostatic sampling values ​​in pneumatic conveying. The intelligent control layer includes a data preprocessing module, an adaptive target generation module, a multi-input fuzzy neural network PID controller, and an adaptive event triggering module, wherein: The data preprocessing module is used to preprocess the data collected by the data acquisition layer. The adaptive target generation module is used to calculate the target gas pressure and target material flow rate based on gas pressure data, material flow rate data and material characteristics. The multi-input fuzzy neural network PID controller is used to output pressure-flow coordinated control commands based on pressure deviation, pressure deviation change rate, flow deviation and flow deviation change rate. The adaptive event triggering module is used to compare the pressure deviation change rate, flow deviation change rate and short-time energy of electrostatic signal with the corresponding thresholds to determine whether to trigger the multi-input fuzzy neural network PID controller. The actuator layer includes electric regulating valves, frequency converters, and quick-release valves.

10. The pneumatic conveying control system of the multi-dimensional collaborative adaptive pneumatic conveying control method according to claim 9, characterized in that, The intelligent control layer also includes a blockage warning and self-healing module, which is used to identify blockage conditions based on gas pressure data and control the system to perform blockage clearing actions, including opening a quick pressure relief valve.

Images