A method and system for controlling pneumatic transport based on pressure wave correlation

By synchronously collecting and separating pressure fluctuation components in the pneumatic delivery pipeline, and using pressure wave correlation technology to predict arrival time and control timing, the problem of mismatch between pressure disturbance propagation state and end-point control timing is solved, achieving more efficient pneumatic delivery control.

CN123105286APending Publication Date: 2026-11-17SHANDONG YINCHI ENVIRONMENTAL PROTECTION EQUIP CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611629724.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-21
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In existing pneumatic delivery pipelines, the propagation state of pressure disturbances does not correspond sufficiently with the timing of end-point control. The identification of the actual arrival time of disturbances after the superposition of pressure changes is easily interfered with, resulting in a mismatch between the control timing and the actual propagation state of disturbances.

Method used

By synchronously collecting pressure time-series data from pressure measuring points in the pneumatic delivery pipeline, separating pressure fluctuation components, and performing pressure wave correlation on adjacent pressure measuring points within the propagation time window, the propagation delay of the same pressure disturbance is confirmed, the arrival time is estimated, and the control sequence is arranged in advance. The control is performed using the response time of the air supply valve and blower to ensure that the control action is executed before the pressure disturbance arrives, and the end pressure deviation is adjusted through feedback and correction.

Benefits of technology

It achieves a full correspondence between the propagation state of pressure disturbance and the timing of end-point control, reduces the interference of changes in execution pressure on the identification of the actual arrival time of disturbance, and improves the control efficiency and accuracy of pneumatic delivery pipelines.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123105286A_ABST
    Figure CN123105286A_ABST
Patent Text Reader

Abstract

This application relates to the field of pneumatic conveying control technology, and discloses a pneumatic conveying control method and system based on pressure wave correlation. The method includes: synchronously collecting pressure time-series data from pressure measuring points in the pneumatic conveying pipeline and separating the pressure baseline to obtain pressure fluctuation components; performing pressure wave correlation to obtain continuous propagation delay; determining the expected arrival time based on the peak time of the end-point measuring point, the continuous propagation delay, and the remaining pipeline distance; arranging the control sequence according to the response time of the air supply valve and the blower; allocating the air supply valve opening and blower frequency; confirming the actual response time using pressure wave correlation after stripping the execution pressure response; and correcting the expected arrival time of the next pressure disturbance by limiting the time deviation and adjusting the air supply valve opening and blower frequency. This technical solution is used for pneumatic conveying control, achieving closed-loop control through propagation identification and execution verification, which can reduce lag and arrival judgment interference, and meet the need for advance control.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of pneumatic conveying control technology, specifically to a pneumatic conveying control method and system based on pressure wave correlation. Background Technology

[0002] In existing technologies, the operation and regulation of pneumatic conveying pipelines typically rely on pressure measurements along the pipeline or at the end. This involves comparing real-time pressure, pressure difference, or flow rate with set ranges and adjusting the gas delivery volume, replenishment status, or power equipment operating parameters to meet the requirements of continuous conveying and prevent blockages. However, existing pressure feedback-based pneumatic conveying regulation has some significant shortcomings.

[0003] In actual conveying processes, fluctuations in material feed, gas source pulsations, pipeline turning, and changes in the two-phase flow state of the material and gas all generate pressure fluctuations that propagate along the pipeline. These pressure fluctuations change in amplitude and shape after passing through different pipe sections. Conventional control primarily relies on the current pressure at a specific measuring point or the deviation from the terminal pressure to trigger adjustments. Typically, the operating status of the air supply valve or blower is changed only after a deviation occurs. For scenarios with long pipelines, dispersed measuring points, and actuators with response times, the measurement, judgment, and execution processes consume adjustment time during the propagation of the same pressure disturbance towards the terminal. After the control action takes effect, new pressure changes will occur at the terminal, superimposed on the original pressure disturbance. Judging the arrival time of the disturbance solely based on the terminal pressure peak is susceptible to interference from the execution process, causing a deviation between the control time and the actual disturbance propagation state, which is then transmitted to subsequent adjustments.

[0004] Therefore, it is evident that existing technologies often suffer from problems such as insufficient correspondence between the propagation state of pressure disturbances in pneumatic delivery pipelines and the timing of end-point control, and susceptibility to interference in identifying the actual arrival time of disturbances after the superposition of execution pressure changes. These are the shortcomings of existing technologies.

[0005] In view of this, it is necessary to provide a pneumatic transport control method and system based on pressure wave correlation to solve the above-mentioned defects in the prior art. Summary of the Invention

[0006] The purpose of this application is to address the shortcomings of the prior art, namely, the insufficient correspondence between the pressure disturbance propagation state and the timing of end-point control in pneumatic conveying pipelines, and the susceptibility of the identification of the actual disturbance arrival time after the superposition of execution pressure changes, by providing a pneumatic conveying control method and system based on pressure wave correlation to solve the above-mentioned technical problems.

[0007] To achieve the above objectives, this application provides the following technical solution: In a first aspect, this application provides a pneumatic transport control method based on pressure wave correlation, comprising: Step S1: Synchronously collect pressure time-series data from pressure measuring points in the pneumatic conveying pipeline and separate the pressure baseline to obtain the pressure fluctuation component; Step S2: Perform pressure wave correlation on the pressure fluctuation components of adjacent pressure measuring points within the propagation time window, confirm the propagation delay of the same pressure disturbance, and shorten the response time of the next pressure measuring point accordingly to obtain the continuous propagation delay; Step S3: Take the nearest pressure measuring point upstream of the terminal as the terminal front measuring point, determine the estimated arrival time based on the peak time, continuous propagation delay and remaining pipeline distance of the terminal front measuring point, and arrange the control sequence in advance according to the response time of the air replenishment valve and the response time of the blower. Step S4: According to the control sequence, allocate the opening degree of the air supply valve and the frequency of the blower based on the pressure fluctuation component of the terminal measuring point and the terminal pressure deviation, and perform control before the same pressure disturbance arrives; Step S5: Based on the execution feedback and calibration pressure response, separate the execution pressure response from the pressure fluctuation component of the end pressure measurement point, associate the remaining pressure fluctuation with the execution pressure wave of the pressure fluctuation component of the end measurement point, and determine the actual response time only when the same correlation peak is always unique within the execution pressure response error range. Step S6: After unique confirmation, the expected arrival time of the next pressure disturbance is corrected according to the deviation between the actual response time and the expected arrival time. If not confirmed, the time correction is frozen. The opening of the air supply valve and the frequency of the blower are adjusted according to the end pressure deviation and the control is executed.

[0008] By adopting the above technical solution, a processing relationship is established around the same pressure disturbance, from pressure fluctuation component acquisition, continuous propagation identification, estimated arrival time determination, pre-arrival control, actual response time confirmation to the next estimated arrival time correction. This achieves time connection and execution result verification between the pressure disturbance propagation process and the control process of the air supply valve and blower. It enables the end-point control basis to be shifted from the already formed local pressure deviation to the pressure disturbance propagation process. After the control action changes the end pressure, a disturbance time reference that can be used for subsequent correction is obtained again. This meets the requirement that the pressure disturbance propagation state and the end-point control timing in the pneumatic conveying pipeline correspond sufficiently and that the identification of the actual disturbance arrival time after the superposition of execution pressure changes is less affected by interference.

[0009] Specifically, synchronously acquiring the pressure fluctuation components at each pressure measuring point provides a consistent time reference for subsequent processing, which helps reduce the interference of slow pressure changes on propagation judgment and provides continuous input for tracing the same pressure disturbance along the pipeline. Establishing segment-by-segment associations around adjacent pressure measuring points and tightening the subsequent response range constrains candidate responses to the preceding propagation state, reducing the entry of irrelevant candidate responses into the time chain and making the continuous propagation delay a traceable basis for end-point prediction. Connecting the pressure fluctuation components of the end-point measuring points with the continuous propagation delay and the remaining pipeline distance, and incorporating the response time of the air supply valve and the blower into the same control sequence, allows the air supply valve and blower to act in advance to reach the unified target, reducing the possibility of execution lagging behind the same pressure disturbance. Coordinating based on the control sequence The opening degree of the air supply valve and the frequency of the blower constrain the adjustment amount by combining the terminal pressure deviation with the same pressure disturbance that is propagating. This retains the current terminal state and ensures that the control action corresponds to the expected arrival process in time. After the control is executed, the execution pressure response is stripped and the remaining pressure fluctuations are re-correlated. The actual response time is then constrained by the correlation peak. This distinguishes the execution pressure response from the arrival response of the same pressure disturbance, allowing the verification result to continue to be included in the preceding propagation time chain. The time deviation between the actual response time and the expected arrival time is limited and fed back to the next expected arrival time. At the same time, the opening degree of the air supply valve and the frequency of the blower are corrected based on the terminal pressure deviation. This can suppress the amplification of the single time deviation, so that the propagation identification, early control, response verification and subsequent correction form a continuous closed control process.

[0010] Preferably, step S2 specifically includes: Within the propagation window, cross-correlation calculations are performed on the pressure fluctuation components of the first pair of adjacent pressure measurement points. The effective propagation delay of the first segment is determined based on the relationship between the pressure change direction, main peak order, and duration of the correlation peaks. Starting from the next pair of adjacent pressure measuring points, the propagation time is calculated based on the previous effective propagation delay and the ratio of the length of the adjacent pipe segment. The propagation time is then correlated with the upstream peak time, and the response time of the next pressure measuring point is shortened within the propagation time window. Calculate the relevant peak during the response period of the next pressure measuring point, perform cross-correlation calculation on the pressure fluctuation components separated by one pressure measuring point to obtain the cross-segment propagation delay, and accumulate the propagation delay of adjacent pipe segments between the same start and end measuring points to obtain the cumulative propagation delay; When multiple related peaks appear in the same response period, candidate related peaks are selected based on the deviation between the cumulative propagation delay and the cross-segment propagation delay. Valid related peaks are determined based on the direction of pressure change, the order of the main peaks, and the duration. Continuous propagation delays are formed according to the pipeline sequence.

[0011] By adopting the above technical solution, the effective propagation delay of the first segment is used as the time constraint for identifying subsequent adjacent pipe segments. The cumulative propagation delay and the cross-segment propagation delay are used to cross-check the candidate correlation peaks within the same response period. This ensures that the direction of pressure change, the order of the main peaks, and the duration are all within the range defined by the preceding propagation time relationship to jointly confirm the effective correlation peaks. This can compress erroneous candidates introduced by local noise or similar pressure fluctuations when propagating segment by segment along the pipeline, ensure that the continuous propagation delay is consistent between adjacent pipe segments, and provide a more stable propagation time basis for the expected arrival time.

[0012] Preferably, when multiple related peaks appear in the same response period, candidate related peaks are screened based on the deviation between the cumulative propagation delay and the cross-segment propagation delay. Valid related peaks are determined based on the direction of pressure change, the order of the main peaks, and their duration. The steps for forming a continuous propagation delay according to the pipeline sequence specifically include: Based on the lengths of adjacent pipe segments on both sides of the pipeline turning position, the effective propagation delay corresponding to a single historical pressure disturbance is converted into the propagation velocity before and after the turning, and a correspondence between the propagation velocity before and after the turning is established. Based on the interpolation of the current propagation speed before turning in the corresponding relationship, the propagation speed after turning and the corresponding propagation time are determined. The current upstream main peak time is shifted backward along the corresponding propagation time to determine the predicted position of the main peak after turning. Connect adjacent pipe segments on both sides of the pipeline turning position with the propagation delay corresponding to each candidate correlation peak, calculate the deviation between the cumulative propagation delay and the cross-segment propagation delay between the same start and end measurement points, and calculate the time difference between the candidate correlation peak and the predicted position. The candidate correlation peaks are selected by the deviation between the cumulative propagation delay and the cross-segment propagation delay from smallest to largest. For candidate correlation peaks with the same deviation, they are compared by the time difference from smallest to largest. Only when the direction of pressure change, the order of the main peaks and the duration match and the candidate correlation peak is unique are the effective correlation peaks confirmed and a continuous propagation delay is formed.

[0013] Based on the above scheme, the historical propagation velocity correspondence on both sides of the pipeline turning position is introduced into the candidate correlation peak screening, so that the current propagation velocity before turning can form the predicted position of the main peak after turning, and together with the deviation of cumulative propagation delay and cross-segment propagation delay, the candidate range is constrained, thereby reducing the interference of the propagation velocity change caused by the pipeline turning position on the candidate correlation peak pairing, and ensuring that the continuous propagation delay remains consistent with the pressure disturbance propagation sequence after crossing the pipeline turning position.

[0014] Preferably, step S3 specifically includes: Pair the continuous propagation delay with the corresponding pipe segment length, calculate the propagation velocity of each pipe segment upstream of the end measuring point, and match the calibration velocity ratio of the remaining pipe segments according to the pipe turning position. The propagation time of the remaining pipeline distance is calculated based on the ratio of the propagation speed of each pipe segment to the calibration speed. The propagation time is then superimposed on the peak time of the measuring point before the end to form the estimated arrival time. Match historical single pressure disturbances to the current operating range, determine the upper limit of arrival time error by the maximum absolute value of the deviation between the historical estimated arrival time and the historical actual response time, and advance the estimated arrival time by the control lead amount greater than the upper limit of arrival time error to determine the common effective time; Based on the common effective time, the response time of the air supply valve and the response time of the blower are traced back to obtain the command time of the air supply valve and the command time of the blower. The control timing is determined by the order of the air supply valve command and the blower command. When both commands are later than the current time, the corresponding execution trigger is set.

[0015] In the above scheme, the propagation speed of each pipe segment upstream of the end measuring point is proportionally connected with the calibration speed of the remaining pipe segment. Based on this, the propagation time of the remaining pipeline distance can be calculated. The lead amount is controlled by the upper limit of the arrival time error. Then, the timing of the air supply valve command and the timing of the blower command are traced back according to the response time of the air supply valve and the response time of the blower, so that the expected arrival time, the timing of the air supply valve command and the timing of the blower command are kept at the same time reference, reducing the effect misalignment.

[0016] As a preferred method, the steps of obtaining the command times of the air supply valve and the blower are derived by tracing back along the response times of the air supply valve and the blower, respectively, based on the common effective time, specifically include: During the historical operation period when the air supply valve changes its opening independently and the blower frequency is stable, the air supply valve operation delay is determined based on the air supply valve command time and the air supply valve feedback start point. The air supply valve pressure delay is determined based on the air supply valve feedback start point and the moment when the end pressure change first reaches the calibrated effective amplitude. During the historical operation period in which the blower frequency is changed independently and the air supply valve opening is stable, the blower operation delay is determined based on the blower command time and the blower feedback start point. The blower pressure delay is determined based on the time when the blower feedback start point and the end pressure change first reach the calibrated effective amplitude. Match historical action segments according to the current operating range, take the upper limit of the cumulative duration of the air supply valve action delay and the air supply valve pressure delay as the air supply valve response time, and take the upper limit of the cumulative duration of the blower action delay and the blower pressure delay as the blower response time. Using the common effective time as the endpoint, trace back along the response time of the air supply valve and the response time of the blower, and round forward according to the sampling period to determine the command time of the air supply valve and the command time of the blower.

[0017] Based on the above processing, the action delay of the air supply valve, the pressure delay of the air supply valve, the action delay of the blower, and the pressure delay of the blower are extracted from the historical action segments of the individual actions of the air supply valve and the blower. The response time of the air supply valve and the response time of the blower are formed according to the upper limit of the cumulative duration of the current operating interval. This allows the action process and the end pressure display process to be included in the instruction backtracking basis, avoiding insufficient advance caused by estimating the response based solely on the control instruction time, and ensuring that the instruction time of the air supply valve and the instruction time of the blower maintain clear execution boundaries.

[0018] Preferably, step S4 specifically includes: Based on the historical amplitude ratio, the pressure fluctuation component of the terminal measuring point is converted into the terminal pressure change at the expected arrival time, and the opening degree of the air supply valve and the frequency of the blower are allocated in combination with the terminal pressure deviation. When the air supply valve command time is earlier than the blower command time, the terminal pressure deviation is obtained at the blower command time. The execution pressure change that has not yet been reflected is predicted according to the actual opening change of the air supply valve. The execution pressure change is included in the terminal pressure change, and the blower frequency is updated in combination with the terminal pressure deviation. When the blower command time is earlier than the air supply valve command time, the terminal pressure deviation is obtained at the air supply valve command time. The execution pressure change that has not yet been reflected is predicted according to the actual frequency change of the blower. The execution pressure change is included in the terminal pressure change, and the air supply valve opening is updated in combination with the terminal pressure deviation. The amplitude and rate of change of the air supply valve opening and the blower frequency are limited. Pneumatic conveying control is performed at the time of the air supply valve command and the time of the blower command, and is performed synchronously when the two are the same.

[0019] Based on the above processing logic, when there is a sequence between the time of the air replenishment valve command and the time of the blower command, the execution pressure change that has not yet been reflected is predicted according to the actual change in the opening degree of the air replenishment valve or the actual change in the frequency of the blower. Then, the blower frequency or the opening degree of the air replenishment valve is updated in combination with the end pressure deviation at the time of the blower command or the time of the air replenishment valve command, so that the opening degree of the air replenishment valve and the frequency of the blower correspond to the end state at the corresponding command time. Then, the adjustment range is constrained by the amplitude and the rate of change to reduce redundant compensation or excessive adjustment.

[0020] Preferably, step S5 specifically includes: Read the calibration pressure response of the individual operation of the air supply valve and the blower, match the action direction and operating range according to the execution feedback, identify the unit response by execution change, process it according to the baseline separation method of pressure time series data, and obtain the response sequence corresponding to the unit execution change; Based on the feedback time of the actual change in the air supply valve opening and the actual change in the blower frequency, the response sequence is scaled according to the change amount of each execution and superimposed along the time axis to obtain the execution pressure response at each sampling time. The pressure response is extracted point by point from the pressure fluctuation components of the end pressure measurement point. The relevant search intervals on both sides of the expected arrival time are determined according to the historical arrival time deviation. Pressure wave correlation is performed on the remaining pressure fluctuation and the pressure fluctuation components of the end measurement point. Within the error range of the pressure response, relevant peaks are compared. Based on the direction of pressure change, the order of main peaks, and the duration, a unique relevant peak that conforms to the continuous propagation delay order is identified, and the corresponding terminal main peak time is determined as the actual response time.

[0021] By adopting the above technical solution, the calibration pressure response of the individual operation of the air supply valve and the blower is scaled and superimposed according to the change amount and feedback time of each execution. This can construct the execution pressure response corresponding to the current control action. The pressure fluctuation component of the terminal pressure measurement point is sampled and separated time by time, so that the remaining pressure fluctuation retains the propagation component of the original pressure disturbance. Then, the unique correlation peak is confirmed in the relevant search interval limited by the historical arrival time deviation, combined with the propagation sequence, which enhances the time consistency between the actual response time and the previous continuous propagation delay.

[0022] Preferably, the steps of comparing relevant peaks within the error range of the pressure response, identifying a unique relevant peak that conforms to the continuous propagation delay order based on the direction of pressure change, the order of the main peaks, and their duration, and determining the corresponding terminal main peak time as the actual response time, specifically include: Each sampling time within the relevant search interval is used as the candidate relevant peak position. The remaining pressure fluctuation and the pressure fluctuation components of the terminal measurement point are aligned according to the corresponding propagation time to determine the rising segment and the falling segment. Calculate the amplitude ratio of the rising segment and the falling segment relative to the corresponding waveband of the end measuring point, reconstruct the end pressure waveform according to the amplitude ratio, and compare the reconstructed end pressure waveform with the remaining pressure fluctuation to obtain the waveform difference; The error range of the executed pressure response at each sampling time is determined by the repeated measurement difference of the calibrated pressure response. Unnormalized cross-correlation is used for calculation. The endpoints of the error range are selected according to the sign of the pressure fluctuation component at the end measurement point. The upper and lower bounds of the correlation values ​​of each candidate correlation peak are calculated. When the lower bound of the correlation value of a candidate correlation peak is higher than the upper bound of the correlation values ​​of other candidate correlation peaks and the waveform difference meets the calibration waveform deviation range, the same pressure disturbance is confirmed by combining the pressure change direction, the order of the main peaks and the duration, and the corresponding terminal main peak time is determined as the actual response time.

[0023] In the above scheme, the sampling time within the relevant search interval is used as the candidate correlation peak position. The terminal pressure waveform is reconstructed by the amplitude ratio of the rising and falling segments, so that the candidate position is simultaneously subject to the constraints of waveform difference and pressure wave correlation. The error range of the executed pressure response is formed based on the repeated measurement difference of the calibrated pressure response. The upper and lower bounds of the correlation values ​​of the candidate correlation peaks are obtained by unnormalized cross-correlation calculation. Only the results with the lower bound of the correlation value higher than the upper bound of the correlation value of the remaining candidate correlation peaks and satisfying the calibration waveform deviation range are retained. This can incorporate the uncertainty of the executed pressure response into the peak competition boundary, reduce the misjudgment of correlation peaks, and improve the verifiability of unique confirmation of the actual response time.

[0024] Preferably, step S6 specifically includes: After the actual response time is uniquely confirmed, the propagation time correction is updated according to the time deviation between the actual response time and the expected arrival time, and the single update amplitude is limited according to the preset ratio of the propagation duration corresponding to the continuous propagation delay. Based on the peak time of the last measuring point before the next pressure disturbance, the continuous propagation delay, and the remaining pipeline distance, the estimated arrival time is calculated, and the estimated arrival time is corrected by the propagation time correction after amplitude limiting during the matching of the operating range. After the actual response time, the terminal pressure deviation is obtained, and the remaining adjustment range of the air supply valve opening and blower frequency is determined according to the execution feedback. The remaining pressure fluctuation is aligned with the peak time of the terminal front measurement point according to the actual response time to confirm the disturbance persistence status. During the ongoing disturbance, the adjustment amount corresponding to the end pressure deviation is allocated according to the remaining adjustment range. When the disturbance ends and the end pressure deviation is within the allowable range, the air supply valve opening and blower frequency are gradually adjusted back to the previous values ​​and pneumatic delivery control is executed.

[0025] Based on the above scheme, after the actual response time is uniquely confirmed, the propagation time correction is updated by the time deviation between the actual response time and the expected arrival time. The single update amplitude is limited according to the propagation duration corresponding to the continuous propagation delay, so that the expected arrival time of the next pressure disturbance absorbs the time deviation and avoids excessive jumps. Then, the remaining adjustment range of the terminal pressure deviation, the air supply valve opening and the blower frequency are combined with the disturbance duration state, so that the adjustment amount in the disturbance duration state is constrained by the remaining adjustment range. After the disturbance ends, it gradually returns to the previous value, forming a subsequent closed loop that connects time correction and terminal pressure regulation, reducing the accumulation of time deviation and excessive regulation in continuous delivery.

[0026] Secondly, this application also provides a pneumatic conveying control system based on pressure wave correlation, comprising: The pressure acquisition module is used to synchronously acquire pressure time-series data from pressure measuring points in the pneumatic conveying pipeline and separate the pressure baseline to obtain the pressure fluctuation component. The time delay correlation module is used to perform pressure wave correlation on the pressure fluctuation components of adjacent pressure measurement points within the propagation time window, confirm the propagation time delay of the same pressure disturbance, and thereby shorten the response time of the next pressure measurement point to obtain the continuous propagation time delay. The arrival prediction module is used to determine the estimated arrival time based on the peak time of the nearest upstream pressure measurement point of the terminal, the continuous propagation delay, and the remaining pipeline distance of the terminal measurement point, and to arrange the control sequence in advance according to the response time of the air supply valve and the response time of the blower. The control and distribution module is used to allocate the opening degree of the air supply valve and the frequency of the blower according to the control sequence, based on the pressure fluctuation component of the end measuring point and the end pressure deviation, and to perform control before the same pressure disturbance arrives. The response verification module is used to separate the execution pressure response from the pressure fluctuation component of the end pressure measurement point according to the execution feedback and calibration pressure response, associate the remaining pressure fluctuation with the execution pressure wave of the pressure fluctuation component of the end measurement point, and determine the actual response time only when the same correlation peak is always unique within the execution pressure response error range. The closed-loop correction module is used to correct the expected arrival time of the next pressure disturbance by limiting the deviation between the actual response time and the expected arrival time after unique confirmation. If not confirmed, the correction is frozen. The module also corrects the opening of the air supply valve and the frequency of the blower based on the end pressure deviation and performs regulation. Attached Figure Description

[0027] To facilitate the explanation of the technical solutions of the embodiments of this application, the accompanying drawings used in this application will be briefly described below.

[0028] Figure 1 This is a flowchart of a pneumatic transport control method based on pressure wave correlation provided in this application; Figure 2 This is a schematic diagram of a pneumatic conveying control system based on pressure wave correlation provided in this application.

[0029] The module comprises: 1. Pressure acquisition module; 2. Time delay correlation module; 3. Arrival prediction module; 4. Control and allocation module; 5. Response verification module; and 6. Closed-loop correction module. Detailed Implementation

[0030] The embodiments of this application will now be described with reference to the accompanying drawings. It should be understood that the following embodiments are only used to illustrate the technical solutions of this application and do not limit the scope of protection of this application. Those skilled in the art can make equivalent substitutions or improvements based on the disclosure of this application.

[0031] It should be noted that in the description of this application, the terms "comprising," "including," "having," and their synonyms are used to indicate the presence of the described features, structures, steps, operations, elements, components, or combinations thereof, but do not exclude the presence of other features, structures, steps, operations, elements, components, or combinations thereof.

[0032] It should be noted in advance that, for ease of understanding of the technical solution of this application, the following explanations are provided regarding the terminology and related technologies involved in this application: 1. Cross-correlation calculation: By comparing the correspondence between two time series signals at different time shifts, the result of the change in the similarity between the two time series signals with time shift is obtained. The time shift corresponding to the correlation peak is used to characterize the propagation time relationship between signal components with common change characteristics.

[0033] 2. Unnormalized cross-correlation calculation: The corresponding sampled values ​​of two time-series signals at different time shifts are directly multiplied and accumulated without normalizing the signal energy. This preserves the influence of the signal amplitude on the correlation result and is suitable for comparing candidate correlation peaks by combining amplitude changes and error boundaries.

[0034] To address the issues of insufficient correspondence between the propagation state of pressure disturbances and the timing of end-point control in pneumatic conveying pipelines, and the susceptibility of the identification of the actual arrival time of disturbances after the superposition of execution pressure changes, traditional regulation relying on current pressure deviation triggers a lag when there is a time difference between long pipelines and execution responses, making it difficult to meet the actual need for continuous control before the arrival of pressure disturbances. This application discloses a pneumatic conveying control method and system based on pressure wave correlation. It organizes the expected arrival time and control sequence with continuous propagation delay, and confirms the actual response time after execution, so that the air supply valve and blower act in advance around the same pressure disturbance. At the same time, it distinguishes between the execution pressure response and the end-point arrival response of the same pressure disturbance. The actual response time is then used to correct the expected arrival time of the next pressure disturbance and to adjust the opening degree of the air supply valve and the frequency of the blower, forming a closed relationship of prediction, execution, verification and correction, reducing the accumulation of time deviations, and making the end-point control sequence close to the actual propagation process of the same pressure disturbance.

[0035] like Figure 1 As shown, this embodiment provides a pneumatic transport control method based on pressure wave correlation, including: Step S1: Synchronously collect pressure time-series data from pressure measuring points in the pneumatic conveying pipeline and separate the pressure baseline to obtain the pressure fluctuation component; Step S2: Perform pressure wave correlation on the pressure fluctuation components of adjacent pressure measuring points within the propagation time window, confirm the propagation delay of the same pressure disturbance, and shorten the response time of the next pressure measuring point accordingly to obtain the continuous propagation delay; Step S3: Take the nearest pressure measuring point upstream of the terminal as the terminal front measuring point, determine the estimated arrival time based on the peak time, continuous propagation delay and remaining pipeline distance of the terminal front measuring point, and arrange the control sequence in advance according to the response time of the air replenishment valve and the response time of the blower. Step S4: According to the control sequence, allocate the opening degree of the air supply valve and the frequency of the blower based on the pressure fluctuation component of the terminal measuring point and the terminal pressure deviation, and perform control before the same pressure disturbance arrives; Step S5: Based on the execution feedback and calibration pressure response, separate the execution pressure response from the pressure fluctuation component of the end pressure measurement point, associate the remaining pressure fluctuation with the execution pressure wave of the pressure fluctuation component of the end measurement point, and determine the actual response time only when the same correlation peak is always unique within the execution pressure response error range. Step S6: After unique confirmation, the expected arrival time of the next pressure disturbance is corrected according to the deviation between the actual response time and the expected arrival time. If not confirmed, the time correction is frozen. The opening of the air supply valve and the frequency of the blower are adjusted according to the end pressure deviation and the control is executed.

[0036] This embodiment uses the pressure fluctuation components acquired and separated synchronously as the propagation identification starting point, enabling each pressure measuring point to enter the processing based on a unified time basis. This reduces the impact of slow pressure changes on the judgment of disturbance patterns and provides continuous input for tracking the same pressure disturbance along the pipeline. By associating the pressure waves of adjacent pressure measuring points and progressively narrowing the subsequent response range, the confirmed time relationship becomes the identification boundary for the next segment, compressing the space of irrelevant candidate responses, and ensuring that the continuous propagation delay provides a verifiable basis for end-point prediction. The estimated arrival time is formed by combining the peak time of the measuring point before the end point, the continuous propagation delay, and the remaining pipeline distance. The response time of the air supply valve and the blower are incorporated into the control sequence, allowing the air supply valve and blower to obtain advance allowance, so that the control action enters the effective stage before the same pressure disturbance arrives at the end point. The opening degree of the air supply valve and the frequency of the blower are coordinated around the control sequence, so that the same pressure disturbance and the end pressure deviation jointly constrain the adjustment direction and amplitude, reducing the time lag of restarting the adjustment after the deviation is formed, and converting the estimated arrival time into the execution boundary. After the action changes the terminal pressure, the execution pressure response is separated based on the execution feedback and the calibration pressure response. Then, the remaining pressure fluctuations are correlated with pressure waves, and the actual response time is constrained by the correlation peak. This reduces the mixing of the execution pressure response with the judgment of the arrival of the same pressure disturbance, allowing the result after execution to continue to be connected to the preceding propagation time chain. The time deviation between the actual response time and the expected arrival time is limited and fed back to the next expected arrival time. The opening of the air supply valve and the frequency of the blower are corrected based on the terminal pressure deviation, so that the time deviation and the terminal state are respectively included in the subsequent time sequence and the update of the air supply valve opening and the blower frequency, suppressing the diffusion of single deviations into the continuous control process. Overall, each major technical link forms a closed effect chain along the same pressure disturbance, so that the upstream propagation state can constrain the terminal action in advance, and the terminal execution result verifies the next prediction in reverse. Under the condition that there is a time difference between the propagation state change and the execution response, the control time reference is maintained continuously, the action connection is clear, and the matching degree between the terminal control and the actual propagation process of the same pressure disturbance is improved.

[0037] The above steps will be specifically described below based on the embodiments of this application.

[0038] In step S1, the core task is to map the pressure electrical signals obtained along the pipeline to the same sampling time, separate the slowly changing pressure baseline, and retain the pressure fluctuation components that can be compared between different pressure measurement points; the pressure measurement point at the end also retains the pressure value before separation for subsequent calculation of the end pressure deviation.

[0039] In this embodiment, pressure timing data from pressure measuring points in the pneumatic conveying pipeline can be collected synchronously, and the pressure baseline can be separated to obtain the pressure fluctuation component. Pressure sensors are set at the beginning, both sides of the turning position, and the end along the conveying direction. The analog-to-digital conversion of each channel is triggered by the same hardware sampling clock, and the sampling time corresponding to the trigger count is written into the data of each channel. The pipe section length is measured along the pipeline centerline, and the lengths of the straight pipe sections before and after the turning position are recorded respectively. The nearest pressure measuring point upstream of the end is fixed as the end measuring point. When inspecting or replacing sensors or changing pipeline connections, the order of pressure measuring points and pipe section lengths are rechecked, and the pressure wave correlation is enabled after the corresponding channel calibration is completed. Specifically, the pressure measuring ports of each pressure sensor are avoided from the direct impact position of the air injection jet, and flush diaphragms or pressure-feeding channels of consistent length and dynamic calibration are used. The sensor range is selected according to the allowable working pressure of the conveying equipment, and at least three traceable pressure points are applied within the range. The offset at zero pressure is deducted for each channel, and the pressure is calculated according to the change in electrical output between adjacent calibration pressure points. If any channel exhibits a hysteresis exceeding the sensor's allowable error during repeated pressure application, the channel is isolated and retested. The pressure values ​​of qualified channels are then converted to kPa. Subsequently, a pressure step is applied simultaneously to all channels, and the feedback starting point is compared. If the sum of the hardware synchronization error and the channel response difference exceeds 0.1 sampling cycles, the sampling connection is adjusted or the mismatched sensor is replaced until the synchronization requirements are met.

[0040] Based on this, the sampling period is selected according to the highest effective frequency of the pressure disturbance to be identified, and an analog low-pass filter is set before analog-to-digital conversion, with the cutoff frequency lower than half of the sampling frequency. For example, for pressure disturbances with an effective frequency not exceeding 20Hz, 200Hz synchronous sampling and 40Hz analog low-pass filtering are used, and the passband amplitude and channel phase difference are confirmed by frequency sweep input. During continuous sampling, the trigger count is checked frame by frame. If there are missing frames, repeated counts, range saturation, or timestamp reversal, all associated windows at the abnormal moment are isolated and covered. Only after restoring complete sampling and reacquiring baseline initialization data is the pressure measurement point restored to participate in propagation confirmation.

[0041] Furthermore, the pressure baseline is formed using a fixed-coefficient moving average that reads only current and past samples, with an averaging window of 400 sampling points. At the start of acquisition, 400 valid pressure samples are first obtained. Then, the earliest sample is removed point by point and incorporated into the current sample. The current pressure baseline is obtained by dividing the sum of the window values ​​by 400. The pressure fluctuation component is obtained by subtracting the current pressure baseline from the current pressure value. This processing is integrated throughout online acquisition and actuator calibration, maintaining a constant averaging window length during the event. Although the current sample generates processing results instantly, the waveform is still affected by fixed filtering characteristics. Therefore, subsequent propagation history, amplitude ratios, and executed pressure responses are all calibrated using the same processed data. The peak times before and after separation cannot be mixed into the same time-delay record.

[0042] For identifying the start and end of pressure disturbances, pressure wave correlation tracking is used to capture the pressure fluctuation envelope formed by the flow of material and gas within the calibration frequency band and propagation time window. The propagation speed is determined by the actual time difference between measurement points. Bands caused by rapid gas phase pressure transmission and slower material concentration changes are recorded separately, and only bands whose morphology matches the propagation history are pre-adjusted. First, within a 10-second sampling period when the gas supply valve opening and blower frequency are stable, the maximum absolute value of the pressure fluctuation component is obtained, and the pressure sensor calibration error is superimposed to form a static steady-state noise boundary. When three consecutive samples exceed twice the static steady-state noise boundary, the first of these three samples is taken as the starting point of the pressure disturbance; when the pressure fluctuation component falls back to within the static steady-state noise boundary and remains there for five consecutive samples, the last sample before the holding period is taken as the ending point of the pressure disturbance. The absolute amplitudes within the start and end intervals are compared, and the time of the unique maximum amplitude is taken as the main peak time. When the maximum amplitude remains constant, the sampling time corresponding to the midpoint of the platform is taken; for even-length platforms, the earlier sampling time is taken. The peak signs of positive and negative pressure disturbances are retained and used when comparing the direction of pressure changes later.

[0043] Baseline initialization is complete, but pressure disturbances have not yet ended. Pressure time-series data is continuously acquired, but propagation confirmation requiring a complete descent phase is not output yet. Each output includes the actual acquisition time of the last required sample. The propagation delay is calculated using the peak time, and control triggering is scheduled using the current time at which the calculation is completed. If two pressure disturbance start and end intervals overlap and cannot be separated, the entire overlapping interval is left unconfirmed, awaiting subsequent independent pressure disturbances. When the influence period of any actuator's action covers the upstream template of a new pressure disturbance, that template is paused for early control. After the action is completed and all upstream pressure measurement points have returned to the static steady-state noise boundary for 400 consecutive samples, independent bands are received. Continuous readings of the original terminal pressure are retained so that limited control can still be performed based on the terminal pressure deviation during the propagation confirmation pause.

[0044] At this point, step S1 generates time-aligned pressure fluctuation components, complete pressure disturbance bands, and corresponding peak times, and sends the effective bands to step S2; the terminal raw pressure and execution feedback continue to be updated according to the same sampling time base.

[0045] In step S2, the core task is to confirm the same pressure disturbance segment by segment along the pipeline, convert the propagation delay already confirmed in the previous segment into the search boundary of the subsequent segment, and eliminate inconsistent connections by combining cross-segment measurements and the propagation relationship in the case of multiple peaks.

[0046] In this embodiment, pressure wave correlation can be performed on the pressure fluctuation components of adjacent pressure measuring points within the propagation time window to confirm the propagation delay of the same pressure disturbance. Based on this, the response time period of the next pressure measuring point can be shortened to obtain a continuous propagation delay. Specifically, cross-correlation can be calculated on the pressure fluctuation components of the first pair of adjacent pressure measuring points within the propagation time window. Based on the pressure change direction, main peak order, and duration relationship corresponding to the correlation peak, the effective propagation delay of the first segment can be determined. First, the complete pressure disturbance waveband of the first pressure measuring point is used as a template, and the equal-length window is moved to each sampling position of the downstream pressure measuring point. All template samples are used at each position to obtain the correlation value within the positive propagation delay range.

[0047] Specifically, the propagation time window is obtained by calibrating independent pressure disturbances within the same pipeline and operating range. During calibration, the actuator remains stationary, and the peak time difference of the upstream and downstream single-peak bands in the same direction is recorded. The minimum time difference is extended forward by two sampling periods, and the maximum time difference is extended backward by two sampling periods, with the starting point still later than zero delay. The operating range is divided according to material type, feed rate, initial air supply valve opening, and initial blower frequency. Continuous quantities use a left-closed, right-open interval, with the highest interval including the right endpoint; the corresponding history is read only if all four items match simultaneously. An event is bound to the operating range based on the initial operating condition before the first instruction is sent. The calibration library simultaneously registers the actual opening, frequency, and action trajectory range that the event is allowed to traverse. Changes in the execution within this plan do not alter the initial operating range; changes in feed or execution trajectory exceeding the coverage area result in the cessation of this advance control. At least 10 non-overlapping effective pressure disturbances are obtained for each operating range, and an additional 5 are taken for verification. If the verification time difference exceeds the propagation time window, the calibration range is expanded and re-verification is performed. When a qualified calibration is lacking, data collection continues without initiating early intervention. For example, for each allowable delay, the upstream sample is multiplied by the downstream sample corresponding to the shifted point, and the results are summed. This summation is then divided by the square root of the product of the energy values ​​of the two sample segments. The normalized cross-correlation value can be written as: , in, Indicates the first [unit] along the conveying direction The and the first The pressure measurement point was moved downstream. Normalized cross-correlation values ​​at each sampling point; for the same template comparison of these two pressure measurement points. and These represent the template starting point shifting backward for the first pressure measurement point. The pressure fluctuation component at the first sample, and the pressure fluctuation component at the second pressure measurement point at the corresponding number of sampling points after the sample, both in kPa. and These represent the unified sampling sequence number of the complete template starting point of the first pressure measurement point and the total number of samples in that template, respectively. The offset within the template ranges from 0 to the total number of samples minus 1, with the downstream shift point number being a positive integer. The second pressure measurement point is located downstream of the first pressure measurement point. The normalized cross-correlation value is dimensionless, ranging from -1 to 1. The calculation terminates when the energy of any band does not exceed the product of the square of the static steady-state noise boundary and the number of samples. The correlation peak must be higher than the adjacent correlation values ​​on both sides; for peaks with the same sign, the downstream main peak must be later than the upstream main peak, and the start-end duration ratio must be within the independent calibration range before multiplying the number of shifted sampling points of the effective correlation peak by the sampling period to confirm the effective propagation delay of the first segment; if multiple peaks cannot be uniquely identified after morphological comparison, the first segment remains unconfirmed.

[0048] After confirming the effective propagation delay of the first segment, starting from the next pair of adjacent pressure measuring points, the propagation duration is calculated based on the ratio of the previous effective propagation delay to the length of the adjacent pipe segment. The current adjacent pipe segment length is divided by the length of the previous adjacent pipe segment, and then multiplied by the previous effective propagation delay. Before reading the two lengths, the common pressure measuring point is checked. If the previous adjacent pipe segment has no effective propagation delay or its length is not positive, the extension to the next segment stops. Subsequently, the propagation duration is correlated with the upstream peak time. Within the propagation time window, the response time of the next pressure measuring point is contracted. Taking the upstream peak time plus the converted propagation duration as the center, the maximum absolute value of the calibration duration deviation is used to expand both sides, and the intersection with the absolute time interval corresponding to the original propagation time window is calculated. The contracted response time can be written as: , in, and They represent the first The response time period after the pressure measurement point contracts and the absolute time interval converted from the original propagation time window; Indicates the current upstream number The peak time of each pressure measuring point. Taking this measuring point as a common endpoint, and These represent the effective propagation time delay and pipe segment length from the previous pressure measuring point to the current pressure measuring point, respectively. and These represent the pipe segment length from the current pressure measuring point to the next pressure measuring point and the calibration deviation boundary of the converted propagation time, respectively. Time is expressed in seconds (s), and length in meters (m). Adjacent measuring point numbers differ by 1, and the intersection operation only retains the portion covered by both time intervals. The left end of the intersection is taken from the most recent sampling time backward, and the right end is taken from the most recent sampling time backward. If the intersection is empty, the current propagation chain ends, and the boundary sampling times are retained as allowed candidates.

[0049] Furthermore, after the response time segment is determined, the correlation peak is calculated position by position within the response time segment of the next pressure measurement point. Each move reads a complete downstream band of the same length as the current upstream complete template, and the direction of pressure change, the order of the main peaks, and the duration are rechecked. If a position where no end-of-window sample has been obtained is encountered, the current event is retained and the process waits for all samples to be available. New pressure disturbances generated during the waiting period are queued separately, and their correlation peaks are not included in the current propagation chain. If a complete band is still not available even after reaching the calibrated maximum duration, the current confirmation is terminated, and the next iteration starts again from the first pair of adjacent pressure measurement points. Based on this, the cross-correlation of pressure fluctuation components separated by one pressure measurement point is calculated to obtain the cross-segment propagation delay. Using the same template from the upstream pressure measurement point, the process crosses the intermediate pressure measurement point, and performs an independent correlation search on the downstream pressure measurement point. The allowable delay range is obtained by adding the two original propagation time windows. If there are multiple peaks in the cross-segment correlation that cannot be eliminated by morphology, the current cross-segment result remains unconfirmed. After obtaining the unique cross-segment propagation delay, the propagation delays of adjacent pipe segments with the same start and end measurement points are accumulated to obtain the cumulative propagation delay. The endpoint of each adjacent connection is checked to ensure consistency with the starting point of the next connection. If duplicates, breaks, or reverse sequences occur during the accumulation process, the current accumulation is cancelled. The absolute value of the difference between the cumulative propagation delay and the cross-segment propagation delay is used for subsequent candidate selection.

[0050] In some embodiments of this application, when multiple related peaks occur within the same response period, candidate related peaks can be screened based on the deviation between the cumulative propagation delay and the cross-segment propagation delay. Valid related peaks are determined according to the direction of pressure change, the order of the main peaks, and their duration, forming continuous propagation delays in pipeline sequence. Specifically, based on the lengths of adjacent pipe segments on both sides of the pipeline turning position, the effective propagation delay of a single historical pressure disturbance can be converted into a propagation velocity before and after the turning, establishing a correspondence between the propagation velocities before and after the turning. The lengths and delays on both sides of the turning are extracted from the same independent pressure disturbance record, and each length is divided by the positive delay, then arranged into paired nodes in ascending order of the propagation velocity before the turning. When the same propagation velocity before the turning corresponds to multiple propagation velocities after the turning, the median of the propagation velocities after the turning is taken to form one node, while the maximum deviation is retained as the calibration error of that node.

[0051] Based on this, interpolation is performed in the corresponding relationship according to the current propagation speed before turning to determine the propagation speed and corresponding propagation duration after turning. First, the left and right adjacent nodes are found, and then the propagation speed after turning is obtained by linear interpolation according to the distance ratio between the current propagation speed before turning and the two nodes. The length of the adjacent pipe segment after turning is divided by this result to obtain the corresponding propagation duration. When the current value is equal to the node, the node value is read directly. If it exceeds the node coverage area or the interpolation result is not positive, the branch is stopped and waiting for supplementary calibration. Next, the current upstream main peak time is shifted backward along the corresponding propagation duration to determine the predicted position of the main peak after turning. The unrounded absolute time is retained as the comparison benchmark. The candidate related peaks still correspond to the sampling time, and the predicted positions are not made equal by pre-rounding.

[0052] Furthermore, after the predicted location is determined, the propagation delay corresponding to each candidate related peak is used to connect adjacent pipe segments on both sides of the pipeline turning position. The deviation between the cumulative propagation delay and the cross-segment propagation delay between the same start and end measurement points is calculated. Each time, only the connection of the current candidate is replaced, while the other confirmed connections remain unchanged. If a candidate causes a reversal or causes the cumulative deviation to exceed the tolerance obtained by independent verification, the candidate is removed from the confirmable range. For candidates that are still comparable, the time difference between the candidate related peak and the predicted location is calculated. The sampling time of the corresponding downstream main peak is subtracted from the predicted location, and the absolute value is taken, keeping the same unit as the cumulative deviation in seconds. At the same time, two independent comparison results are retained. Based on this, candidates are first screened according to the deviation between the cumulative propagation delay and the cross-segment propagation delay from smallest to largest. Candidate related peaks with the same deviation are then compared according to the time difference from smallest to largest. The cumulative deviation is calculated as an integer sampling point difference, and only those with the same integer value are compared in terms of time difference. For example, if the cumulative deviation of both candidates is 2 sampling points, and the time difference from the predicted position is 0.010s and 0.025s respectively, the former is ranked first; if the cumulative deviation is 1 and 2 sampling points respectively, the one with the smaller cumulative deviation is retained first. After sorting and morphology verification, only when the pressure change direction, main peak order and duration match and the candidate related peak is unique, is the effective related peak confirmed and a continuous propagation delay is formed. If the peaks are tied for best or all morphologies do not match, the effective propagation delay is not output to the next pipe segment.

[0053] At this point, step S2 applies the adjacent correlation, cross-segment verification, and multi-peak screening to the same pressure disturbance, and outputs the continuous propagation delay connected in the order of physical pipelines; only when the propagation chain actually reaches the end measurement point does it proceed to the expected arrival calculation in step S3.

[0054] In step S3, the core task is to convert the confirmed pressure propagation relationship into the expected arrival time of the terminal, and to incorporate the action lag and pressure lag of the air supply valve and the blower into the instruction arrangement, so that the two-way regulation has an executable advance effective sequence.

[0055] In this embodiment, the nearest upstream pressure measuring point can be used as the pre-terminal measuring point. The estimated arrival time is determined based on the peak time, continuous propagation delay, and remaining pipeline distance of the pre-terminal measuring point. The control sequence is then arranged in advance according to the response time of the air supply valve and the blower. Specifically, the continuous propagation delay can be paired with the corresponding pipeline segment length to calculate the propagation velocity of each pipeline segment upstream of the pre-terminal measuring point. The calibration velocity ratio of the remaining pipeline segment is matched according to the pipeline turning position. Subsequently, the positive time delay and centerline length are read segment by segment according to the start and end positions of adjacent pressure measuring points, and the length is divided by the propagation delay. For pipeline segments with mismatched endpoints, length changes that have not been recalibrated, or time delays of zero, the estimated arrival calculation is cancelled. Here, the propagation velocity of the pressure disturbance in the corresponding pipeline segment is obtained, and the subsequent remaining distance is also calculated based on the same pipeline centerline diameter.

[0056] Furthermore, the remaining pipe segments are matched with calibration speed ratios based on their turning positions. The remaining pipe segments between the pre-end measuring point and the end pressure measuring point are divided according to their turning positions. During calibration, pressure sensors are temporarily and synchronously placed at the dividing points to measure the propagation velocity of each remaining pipe segment with a single pressure disturbance. This velocity is then matched with the propagation velocity of an upstream reference pipe segment with the same turning relationship, and the median of the velocity ratios is taken as the calibration speed ratio. Straight pipe segments are preferentially matched with the nearest upstream straight pipe segment of the same diameter and the same delivery direction. Pipe segments after turning are matched with historical records of the same bend angle and bend radius. If there is no matching reference pipe segment, the remaining segment is calibrated separately without using another pipe ratio.

[0057] After the remaining pipe segments are matched, the propagation time of the remaining pipeline distance is calculated based on the ratio of the propagation speed of each segment to the calibrated speed, and this propagation time is then added to the peak time of the end-point measuring point. During this process, the reference propagation speed is multiplied by the calibrated speed ratio for each segment, and then the remaining pipe segment length is divided by the converted speed to accumulate the propagation time of each remaining pipe segment. The propagation time correction amount already saved in the current operating section is read and added according to the time correction relationship in step S6 to form the estimated arrival time for the current actual use, which can be written as: , in, Indicates the estimated arrival time of the current pressure disturbance; Indicates the first measurement point before the end point. Peak moment of each pressure measurement point; This represents the total number of remaining pipe segments. For the [number]th... One remaining pipe section, and These represent the pipe section length and the ratio of the matched calibration speed, respectively. This indicates the sequence number of the starting pressure measurement point of the upstream reference pipe segment matched for the remaining pipe segment. This indicates the propagation velocity of the reference pipe segment from the pressure measuring point with this serial number to the next pressure measuring point; the remaining pipe segments are accumulated segment by segment from 1 to the total number. Indicates the first Each operating range is completed The propagation time correction is saved after each effective update. Initially, both the update count and the correction are zero. Length is expressed in meters (m), velocity in meters per second (m / s), the calibration velocity ratio is dimensionless, and all other time quantities are expressed in seconds (s). If any calibration velocity ratio or propagation velocity is not positive, no predicted arrival time is generated. The actual predicted value used in this instance remains unchanged, and the time correction after terminal confirmation is used for the next pressure disturbance.

[0058] After obtaining the estimated arrival time, historical single pressure disturbances are matched to the current operating range. The upper bound of the arrival time error is determined by the maximum absolute value of the deviation between the historical estimated arrival time and the historical actual response time. The estimated arrival time is advanced by the control lead amount greater than the upper bound of the arrival time error to determine the common effective time. This upper bound constraint has verified the control arrangements within the operating condition. When online observations exceed the boundary, the boundary is updated and untriggered plans are suspended. Historical samples simultaneously satisfy the requirements of a complete propagation chain, a single pressure disturbance, and reliable terminal peak confirmation. Upon initial activation, the actuator is stably acquired for 10 independent pressure disturbances, and the actual response time is directly determined from the complete terminal band. Each time, the historical estimated arrival time is recalculated using the available upstream data at that time. Terminal observations cannot be used to reverse-write historical predictions. Subsequently, only events uniquely confirmed in step S5 are included; operating condition out-of-bounds and unconfirmed events are not included in the error statistics. Based on this, the estimated arrival time is advanced by a control lead greater than the upper bound of the arrival time error to determine the common effective time. The control lead is taken as the upper bound of the arrival time error plus two sampling periods and the measured upper bound of the command transmission jitter. The common effective time is obtained by subtracting the control lead from the estimated arrival time. The command transmission jitter is measured through 100 consecutive no-load trigger tests. If it exceeds this range during operation, any untriggered plans are canceled and communication is rechecked. When the operating parameters exceed the coverage range of historical samples, the upper bound of the arrival time error is deactivated and recalibrated. When the historical maximum deviation increases with the addition of new valid samples, the updated control lead is used for subsequent control. The timing of existing events is recalculated before either command is triggered. If either command has already been triggered, feedback correction is initiated, and expired commands are not retroactively reissued.

[0059] In some embodiments of this application, the timing of the air supply valve command and the timing of the blower command can be obtained by tracing back along the response time of the air supply valve and the blower command time, respectively, using the common effective time as a reference. Specifically, within the historical action segment where the air supply valve changes its opening independently and the blower frequency is stable, the air supply valve action delay can be determined based on the air supply valve command time and the air supply valve feedback start point. Then, the air supply valve pressure delay can be determined based on the air supply valve feedback start point and the moment when the end pressure change first reaches the calibrated effective amplitude. In the statically stable segment without upstream pressure disturbance, positive and negative opening changes are applied respectively. The first moment when the actual opening leaves the feedback statically stable noise band for three consecutive samples is taken as the air supply valve feedback start point; the first moment when the end pressure first exceeds three times the pressure noise boundary along the calibrated direction and continues for three samples is taken as the pressure effective start point. The two start points are traced back to the first sample confirming continuity.

[0060] Correspondingly, within the historical operating segment where the blower frequency changes independently and the air supply valve opening remains stable, the blower operating delay is determined based on the blower command time and the blower feedback start point. Then, the blower pressure delay is determined based on the blower feedback start point and the moment the terminal pressure change first reaches the calibrated effective amplitude. The actual output frequency of the frequency converter is read, and the blower feedback start point is determined according to the same rule of three consecutive samples. For both calibrations, three variation amplitudes are selected within the allowable operating range, and each amplitude and direction is repeated at least 10 times. Only operating segments where the feedback from the other actuator remains within the static stable noise band and there is no pressure disturbance upstream are included. If feedback is not initiated, the pressure change direction is reversed, or the calibrated effective amplitude is not reached, the calibration is stopped, the execution status is checked, and the operating segment is re-acquired.

[0061] In this embodiment, historical action segments are matched according to the current operating interval. The upper limit of the cumulative duration of the replenishing valve action delay and the replenishing valve pressure delay is taken as the replenishing valve response time, and the upper limit of the cumulative duration of the blower action delay and the blower pressure delay is taken as the blower response time. For each valid action segment, the two delays of this segment are first accumulated, and then the maximum value of the cumulative duration of all matched action segments is taken, and a time quantization margin of one sampling period is added. Each accumulated value comes from the same action, and the feedback starting points of different actions cannot be spliced ​​together. The response time is matched according to the planned action direction, total change amplitude, and gradual change rate. Each increment of the planned trajectory must be within the coverage range of the calibration group. When the actual action amplitude exceeds the calibration coverage range, the amplitude within the allowable range is reselected. If the pressure demand still cannot be covered, the advance arrangement is stopped. Subsequently, with the common effective time as the endpoint, the response time of the replenishing valve and the blower are traced back, and the sampling period is rounded forward to determine the replenishing valve command time and the blower command time. In this embodiment of the application, taking the zero time of the unified sampling time base as the rounding starting point, the instruction time can be written as: , in, and They represent the first The command timing of the actuator and the response time of the matched action trajectory; the first and second paths correspond to the air supply valve and the blower, respectively. This indicates the need to control the lead time. This indicates a positive sampling period. The lower bracket rounds the inner quotient towards negative infinity. All times are measured in seconds and relative to the same sampling zero point. Backtracking and rounding are processed using integer counting based on a unified time base, and the boundary is retained if it is exactly at the sampling grid boundary. The obtained command time is no later than the original backtracking time, and the common effective time is the latest time when the two pressure actions reach the target time of matching calibration effective amplitude. The two actual response starting points can be different.

[0062] Based on this, a control sequence is formed according to the order of the air supply valve command time and the blower command time. The actuator corresponding to the earlier time is listed as the first to act, and the actuator corresponding to the later time is listed as the last to act. If the two rounding results are the same, the same sampling trigger is arranged. When the arrangement is completed, the current time is reread. If both the air supply valve command time and the blower command time are later than the current time, the corresponding execution trigger is set, and it is checked that the remaining time from the earlier command time is sufficient to complete the transmission. For example, when the expected arrival time is 13.70s, the control advance is 0.15s, the two response times are 0.24s and 0.39s respectively, and the sampling period is 0.005s, the two command times are 13.31s and 13.16s. If the current time is 13.20s, the entire plan cannot be executed, the advance command that has not yet been triggered is canceled, and the process proceeds to step S4, which involves end pressure limiting feedback control.

[0063] At this point, step S3 generates the estimated arrival time, the co-effective time, and the two instruction times after the futureness check; step S4 allocates the execution quantity according to the control timing; and step S5 retains the estimated arrival time as the time reference for related searches and subsequent corrections.

[0064] In step S4, the core task is to convert the expected end pressure change and the current end pressure deviation into the actual executable air supply valve opening and blower frequency, and to deduct the pressure effects that have been manifested and those that have not yet been manifested when there is a sequence between the two actions.

[0065] In this embodiment, the control can be performed according to the control sequence, based on the pressure fluctuation component of the pre-terminal measuring point and the terminal pressure deviation, by allocating the air supply valve opening and blower frequency before the arrival of the same pressure disturbance. Specifically, the pressure fluctuation component of the pre-terminal measuring point can be converted into the terminal pressure change at the expected arrival time according to the historical amplitude ratio. Then, the air supply valve opening and blower frequency are allocated in combination with the terminal pressure deviation. From the historical single pressure disturbance of the actuator's stability, the ratio of the main peak amplitude of the terminal and the pre-terminal measuring point in the same direction is obtained, and the median is taken as the historical amplitude ratio within the matched operating range. The current terminal pre-terminal measuring point band is shifted along the expected propagation time and multiplied by the historical amplitude ratio to obtain the expected terminal band; when the main peak amplitude falls outside the historical amplitude coverage range, the amplitude prediction is stopped. In this process, the terminal pressure deviation is the current terminal original pressure minus the terminal target pressure given by the process; a positive value indicates that the pressure is too high. The expected arrival time of the terminal band is subtracted from the current value to obtain the remaining disturbance pressure increment. If the current time is earlier than the expected start time of the band, the current value is treated as zero. This increment is added to the current terminal pressure deviation to obtain the pressure change that the two actions need to offset. The directions of the two pressures are determined by individual action calibration: the original terminal pressure change caused by a 1 percentage point increase in the actual opening of the air supply valve, and the original terminal pressure change caused by a 1 Hz increase in the actual frequency of the blower, both retain the measured positive and negative signs; when the gain approaches the static steady noise boundary or changes sign with the action amplitude, the action range is reduced and recalibrated.

[0066] Based on this, the allowable opening degree variation and frequency variation are discretized according to the actual resolution of the actuator. The remaining end pressure deviation at the expected arrival time is calculated for each pair, and the combination with the smallest absolute value is selected. The selected combination can be written as: , in, This indicates the combination of the selected air supply valve opening change and the blower frequency change. and These represent the total change in the air supply valve opening and the total change in the blower frequency, respectively, which are being compared. This represents a finite set of candidate combinations that simultaneously satisfy the constraints of amplitude, rate of change per sample, calibration, and arrival time. This indicates the current end pressure deviation. This indicates the incremental disturbance pressure that is yet to occur. and These represent the signed effective pressure gain at the expected arrival time after the summation of the total changes of the two candidate types through their respective planned trajectories. The change in the supplementary air valve is expressed as a percentage point, and the change in the blower is expressed as a frequency (Hz). The two types of gains are expressed in kPa per percentage point and per Hz, respectively. The pressure product and the terminal pressure deviation are both in kPa. A finite search selects the combination with the smallest absolute residual deviation. When two combinations are tied, the combination with the smaller absolute value of frequency change is selected first, followed by the combination with the smaller absolute value of opening change. If they are still tied, a unique combination is determined by the order of increasing opening change and frequency change. Zero change is represented by zero pressure contribution.

[0067] In this process, for each candidate total change, first, allocate unidirectional increments that meet the rate of change limit according to the sampling period from the expected feedback starting point. The last increment is taken as the remaining change, forming the planned execution trajectory. The expected feedback starting point is a conservative upper bound of the planned instruction time plus the matching action delay. According to the original unit pressure response identified in step S5, multiply each increment of the planned trajectory by the response value corresponding to the expected arrival time and sum them up. Divide the resulting pressure effect by the non-zero total change to obtain the signed pressure gain of this candidate; zero total change directly takes zero pressure contribution. When the response reading position falls between two calibrated sampling positions, linear interpolation is used, and negative time differences are taken as zero. If the planned trajectory cannot be completed within the calibration timing coverage, or if any increment is in the unverified direction, amplitude, and rate range, the corresponding combination is removed from the candidate combination set. For example, the current terminal pressure deviation is 0.8 kPa, the expected disturbance pressure increment is 0.4 kPa, the two effective pressure gains are 0.2 kPa per percentage point and 0.5 kPa per Hz, respectively. A reduction of 4 percentage points in opening and a reduction of 0.8 Hz in frequency can offset 1.2 kPa at the expected time. If the actual available change is insufficient, the unoffset part is retained, and subsequent limited adjustment is carried out based on terminal pressure feedback.

[0068] Furthermore, for two scheduled different command times, when the air supply valve command time is earlier than the blower command time, the terminal pressure deviation is obtained at the blower command time. The execution pressure change that has not yet been reflected is predicted based on the actual opening change of the air supply valve, and this execution pressure change is included in the terminal pressure change. The blower frequency is then updated based on the terminal pressure deviation. The actual opening of the air supply valve is read sample by sample, and the difference between adjacent actual feedbacks is taken as one execution change. The original unit pressure response is read starting from the time of this feedback. The pressure effect expected to arrive at the expected time minus the pressure effect that has already occurred is the part that needs to be added to the subsequent action calculation. For example, this difference is summed according to each actual execution change, and can be written as: , in, Indicates the first The first actuator in the path is in the The changes in execution pressure that are not yet reflected at the unified sampling time but will be reflected at the expected arrival time; This represents the total number of valid feedback changes received by this actuator up to the current calculation. For this actuator, sorted by feedback time... The second change and These represent the actual amount of change and the corresponding unified sampling sequence number, respectively. and These represent the values ​​of the original unit pressure response matched for this change at the expected arrival time and the current sampling time, respectively. Both complete independent variables are the time difference relative to the feedback start point. The original unit pressure response refers to the original terminal pressure change caused by the actual execution step of the unit, selected according to the matching direction, operating range, and verification range; the actual change is expressed as a percentage point in the air supply valve branch and in Hz in the blower branch, and the response is expressed as kPa caused by the corresponding unit execution change, with the cumulative difference being kPa. The feedback change is accumulated from 1 to the total number of valid times, with the response at negative time positions being zero, and the current sampling sequence number being the value corresponding to the time of the subsequent action command. Execution pressure changes that have not yet been reflected are included in the disturbance pressure increment, the executed air supply valve action is fixed, and the action is only searched for in the remaining allowable trajectory of the blower.

[0069] Correspondingly, when the blower command time is earlier than the air supply valve command time, the terminal pressure deviation is obtained at the air supply valve command time. The unreflected execution pressure change is predicted based on the actual frequency change of the blower, and the execution pressure change is included in the terminal pressure change. The air supply valve opening is then updated based on the terminal pressure deviation. A feedback change sequence is formed using the actual output frequency of the frequency converter. The expected action and the reflected action are read sequentially using the original unit pressure response of the blower. At this time, the blower action that has already been executed is maintained, and only the remaining allowable opening of the air supply valve is searched. If the action feedback arrives first but the actual change is zero, the unreflected execution pressure change is zero. If the feedback has not been updated for more than two sampling cycles, the action command is paused and the early activation confirmation is canceled. The calculation is recalculated based on the current terminal pressure deviation after the feedback is restored.

[0070] The amplitude and rate of change of the air supply valve opening and blower frequency are limited. Pneumatic conveying control is performed separately at the time of the air supply valve command and the time of the blower command, and synchronously when they are the same. Specifically, before the two commands are submitted, the target execution quantity is constrained by the upper and lower limits of the equipment. Then, the planned trajectory is checked by sampling from the current actual execution quantity. The change in the feedback of adjacent plans does not exceed the product of the allowable rate of change and the sampling period. The entire trajectory intersects with the calibrated direction, amplitude, rate and cumulative change range. The candidate combination set consists of target grid points that can be reached by the qualified trajectory. The target change quantity can be completed across multiple sampling periods. The actual increment in each period is limited. Existing incomplete commands occupy the corresponding trajectory. It is forbidden to issue duplicate targets. After the subsequent action is recalculated, the response time and command time of the corresponding trajectory are checked again. If the requirement for joint effectiveness cannot be met, the advance arrangement is cancelled and the restricted feedback control is retained. When the equipment pressure protection limit is triggered, the prediction allocation is terminated and the existing material stop or shutdown interlock of the equipment is executed. Advance control is allowed only after the protection is released, the execution feedback is restored and an effective propagation chain is re-established.

[0071] After allocation according to the above restrictions, pneumatic conveying control is executed at the time of the air supply valve command and the time of the blower command, respectively. When the time of the air supply valve command and the time of the blower command are the same, they are executed synchronously. The two commands at the same time read the same pressure and execution feedback, calculate only one combination and submit them together, and verify the receipt confirmation of the two channels separately. If the feedback of the first action is reversed, the action times out, or the driver rejects the command, the subsequent action plan is canceled, and the commands that have already been issued are not resent. Only when the feedback is available and the pressure is still within the allowable operating range of the equipment, the pressure evaluation time is determined by the larger value of the current time plus the response time of the two channels, and the corresponding original unit pressure response is reread. The restricted combination search is performed with the current end pressure deviation, the zero disturbance prediction increment, and the execution pressure change that has not yet been reflected. After the pressure enters the process allowable deviation zone, the actual execution volume is maintained, waiting for a new complete pressure disturbance.

[0072] At this point, step S4 implements the control timing into two restricted instructions, and corrects the subsequent actions based on the actual feedback of the first action; each actual execution change, along with the feedback time, enters step S5 for point-by-point reconstruction of the execution pressure response.

[0073] In step S5, the core task is to use two-way execution feedback to reconstruct the end pressure effect introduced by the regulation, separate this effect from the pressure fluctuation component of the end pressure measuring point, and transfer the calibration error to the comparison range of all candidate correlation peaks. Only when the same candidate remains unique within the allowable error will the actual response time be output.

[0074] In this embodiment, the execution pressure response can be separated from the pressure fluctuation component of the end pressure measuring point according to the execution feedback and calibration pressure response. The remaining pressure fluctuation is then correlated with the execution pressure wave of the pressure fluctuation component of the end measuring point. The actual response time is determined only when the same correlation peak is always unique within the execution pressure response error range. Specifically, the calibration pressure response of the air supply valve and blower acting alone can be read and processed according to the pressure baseline separation method of the pressure time series data. The original end pressure and actual feedback of the individual action test in step S3 are saved. First, the original unit pressure response is identified according to the actual feedback increment. Then, the identification result, together with 400 zero response samples before the starting point, is sent to the same moving average processing as in step S1 to obtain the pressure fluctuation corresponding to the unit execution change point by point. The zero response interval before the feedback starting point is retained each time, so that there is a unified starting point when shifting according to the actual feedback.

[0075] Furthermore, the calibration pressure responses of the individual actions of the air supply valve and blower are read, and the action direction and operating range are matched according to the execution feedback. The unit response is identified by execution change, and processed according to the baseline separation method of the pressure time series data to obtain the response sequence corresponding to the unit execution change. Specifically, for each effective individual action, the difference between adjacent actual feedbacks is calculated point by point, and the static steady-state average value before the action is subtracted from the original pressure at the end, forming the input increment sequence and the pressure change sequence respectively. The original unit pressure response at each sampling delay from zero delay is taken as the unknown quantity; each pressure sample corresponds to one row, and the coefficient of a certain delay column is taken as the actual feedback increment at the corresponding delay before that sample, and the negative time coefficient is set to zero. The rows of multiple actions in the same direction and operating range are merged in chronological order. The calibration record must cover the stable response segment, and the number of observation rows is not less than the number of unknown quantities. The QR decomposition with column pivot is used to solve for the unit response that minimizes the square error of pressure reconstruction. If the absolute value of the diagonal term does not exceed one millionth of the largest diagonal term, it is considered that the rank is insufficient, the identification of that group is stopped, and re-acquisition is carried out. Forward and reverse directions are identified separately. Three amplitudes in the same direction are fitted together and verified point-by-point using five actions not involved in the fitting. If the verification error exceeds the allowable error for pressure measurement and control, the applicable range is narrowed and re-identification is performed. The unit response is averaged and separated using the same 400-point method to form an online reconstruction sequence. For gradual actions, the unit response is obtained from the incremental relationship of the entire actual feedback segment; the actual feedback increment is only included once during online superposition. The superposition error is then checked using independent dual-actuator action segments, and it is confirmed that the cumulative absolute value of the actual change within the window is within the coverage range of the dual-action verification; if it exceeds the range, the current confirmation is stopped.

[0076] When receiving actual feedback online, the response sequence is scaled and superimposed along the time axis according to the feedback time of the actual change in the air supply valve opening and the actual change in the blower frequency, to obtain the execution pressure response at each sampling time. First, adjacent feedback values ​​are compared to obtain the signed actual execution change, and then the unit response in the corresponding direction is shifted to the sampling time of that feedback. If the same feedback sequence number arrives repeatedly, it is only included once. Late feedback is added before the confirmed output, and the affected sampling points are recalculated. Continuous changes are processed using the measured increment within each sampling interval. The execution pressure response can be written as: , in, Indicates the first The execution pressure response estimated at a uniform sampling time, in kPa; Indicates the first The first actuator The baseline separation unit response matched by the actual executed change is the value taken at the difference between the current sampling number and the feedback sampling number. This unit response is obtained by applying the same fixed coefficient baseline separation online to the original unit pressure response, and the unit is kPa caused by the corresponding unit executed change. Negative sampling offset is taken as zero. After 400 consecutive samples are collected from the tail of the response and return to the static steady noise boundary, the subsequent tail is taken as zero. The maximum residual amplitude before truncation is included in the error range. Gaps exceeding the allowable feedback interval are not filled with fictitious changes, and the confirmation of the current actual response is suspended.

[0077] After the execution pressure response is generated, it is extracted point by point from the pressure fluctuation components at the terminal pressure measurement points. The correlation search intervals on both sides of the expected arrival time are determined based on the historical arrival time deviation. Pressure wave correlation is then performed on the remaining pressure fluctuations and the pressure fluctuation components at the terminal measurement points. Specifically, the execution pressure response at the same moment is subtracted from the pressure fluctuation components at the terminal pressure measurement points according to a unified sampling sequence number. Before subtraction, the operating range used for the response is checked to be consistent with the current execution feedback. If any path is outside the uncalibrated direction or action range, the original terminal sample is retained, but no identifiable remaining pressure fluctuation is output. When the execution feedback is restored and all affected samples can be reconstructed, the entire correlation window is recalculated. Subsequent correlation values, waveform differences, and boundary comparisons are all read from the recalculated results.

[0078] Furthermore, the relevant search intervals on both sides of the expected arrival time are determined based on the historical arrival time deviation. The historical actual response time within the matching interval is subtracted from the historical expected arrival time, and the minimum and maximum values ​​are taken. These values ​​are then extended outwards by two sampling periods, and finally added to the expected arrival time to obtain the two boundaries. When the boundary falls between the sampling times, the most recent sampling time is taken outwards. When the historical maximum error increases, the entire search interval is first extended, and then all candidates are recalculated. In step S3, when correcting the prediction using the propagation time correction, historical residuals with the same correction caliber are read simultaneously to avoid directly applying the deviation of the uncorrected prediction to the corrected prediction.

[0079] After the relevant search interval is determined, pressure wave correlation is performed on the remaining pressure fluctuations and the pressure fluctuation components of the terminal measuring point within the relevant search interval. The complete waveband of the terminal measuring point is fixed as the template for this operation, and the complete waveband of the terminal is aligned position by position. The start and end of the window remain unchanged relative to the main peak of the template. If the edge candidate lacks subsequent samples, the system waits for the arrival of the last sample required by the latest candidate. If the output is still incomplete after the calibrated maximum duration, the entire output is not confirmed. Before all comparisons are completed, the expected arrival time and the current execution record are kept unchanged, and subsequent pressure control continues to read the original terminal pressure.

[0080] In some embodiments of this application, relevant peaks can be compared within the error range of the pressure response. A unique relevant peak that conforms to the continuous propagation delay order is identified based on the direction of pressure change, the order of the main peaks, and their duration. The corresponding terminal main peak time is then determined as the actual response time. Specifically, each sampling time within the relevant search interval can be used as a candidate relevant peak position. The remaining pressure fluctuation and the pressure fluctuation components of the terminal pre-measurement point are aligned according to the corresponding propagation duration to determine the rising and falling segments. The candidate positions are directly assumed to be the sampling positions where the terminal main peak is located. The template is then shifted according to the time difference between the candidate and the terminal pre-measurement point's main peak. During alignment, the differences between adjacent template samples are compared point by point. Positive differences correspond to the rising segment, and negative differences correspond to the falling segment. Platform samples are assigned to the band corresponding to the previous non-zero difference, and starting platform samples are assigned to the band corresponding to the next non-zero difference. If any band is missing or contains multiple indivisible main peaks, the candidate does not meet the waveform confirmation conditions.

[0081] For candidates with two wavebands, the amplitude ratios of the rising and falling segments relative to the corresponding wavebands at the end-of-terminal measurement points are calculated separately. The end-of-terminal pressure waveform is reconstructed according to these amplitude ratios. The reconstructed end-of-terminal pressure waveform is then compared with the remaining pressure fluctuations to obtain the waveform differences. Specifically, within each waveband, the template sample is multiplied and summed item by item with the aligned remaining pressure fluctuations, and then divided by the sum of the squares of the template samples for that waveband to obtain the amplitude ratio for that waveband. The two denominators are compared with the square of the static steady-state noise boundary multiplied by their respective sample numbers. If either denominator is not large enough, the reconstruction of the candidate waveform is stopped. If the ratio is not positive or exceeds the amplitude ratio range of independent pressure disturbances under the same operating condition, the relevant value boundary of the candidate is retained, but it is not allowed to become the final confirmed result. Positive pressure peaks and negative pressure troughs are segmented according to the actual rising and falling directions. The entire process from the starting point of a negative pressure disturbance to the main peak cannot be referred to as the rising segment. Based on this, the terminal pressure waveform is reconstructed according to the amplitude ratio of each of the two bands. The reconstructed terminal pressure waveform is compared with the residual pressure fluctuation to obtain the waveform difference. The amplitude ratio of the corresponding band is read for each template sample, scaled, and subtracted from the corresponding residual pressure fluctuation. The sum is then expressed as the root mean square waveform difference, which can be written as: , in, This indicates that the candidate peak at the end is located at the unified sampling sequence number. Differences in the root mean square waveform at different times; This indicates the sample offset of the main peak of the template at the end of the test point relative to the template starting point. This indicates that after aligning the candidate peak, the template starting point is moved backwards by the [number]th peak. The residual pressure fluctuations at the end of each sample. This indicates the amplitude proportion of the current sample in the template at the candidate peak position. This indicates that the offset sample of the template is assigned to the band mapping of the rising or falling segment, and the amplitude ratio is dimensionless. For the first... A complete template for each pressure measurement point , and The parameters are, in order: total number of template samples, unified sampling sequence number of the starting point, and pressure fluctuation component at the corresponding sample number after the starting point; all candidates use the same template. Waveform differences and pressure fluctuations are both expressed in kPa. The upper limit is allowed to be the maximum difference obtained by reconstructing independent verification events in the same segment plus the pressure calibration error. If it equals the upper limit, it is retained; otherwise, it is not recognized.

[0082] To address the uncertainty in the stripping process, the error range of the executed pressure response at each sampling time is determined based on the repeated measurement differences of the calibrated pressure response. Unnormalized cross-correlation is used for calculation, and the endpoints of the error range are selected based on the sign of the pressure fluctuation component at the end-point measurement. The upper and lower bounds of the correlation values ​​for each candidate correlation peak are calculated. Specifically, the unit responses of repeated actions are identified using the same method, and subtracted point-by-point from the unit responses of the corresponding directions and operating intervals obtained through merging identification, with the maximum absolute value taken. Then, five independent actions not involved in the identification are used for verification; if any exceedance occurs, the error envelope is expanded and re-verified. The unit response error is scaled according to the absolute value of the actual execution change, superimposed along the feedback sampling time, and includes the uncovered superposition bias of the independent dual-action verification, the response truncation margin, and the online end-point pressure measurement margin. The error half-width can be written as: , in, Indicates the first The non-negative error half-width after pressure response stripping is performed at a uniform sampling time; Indicates the first The first actuator The unit response error envelope of the actual execution change matching is the value taken at the difference between the current sampling number and the feedback sampling number; This represents the sum of the maximum deviation not covered by the sum of the errors of the two paths in the independent dual-action verification, the response truncation margin, and the online end-point pressure measurement margin. The unit response error envelope is calculated in kPa based on the corresponding unit execution change, and the accumulated error half-width is kPa. The execution pressure response error is the difference between the actual action and the estimated action, and the end-point measurement margin is incorporated into the symmetrical interval, with the upper and lower endpoints being the error half-width and its opposite, respectively. When the initial operating conditions such as feeding change, the actual trajectory exceeds the verification range, or the response tail continuously exceeds the limit, the actual response confirmation is stopped, the propagation time correction is frozen, and a qualified error envelope is re-established.

[0083] Based on this, unnormalized cross-correlation calculation is used. The endpoints of the error range are selected according to the sign of the pressure fluctuation component at the end-point measuring point, and the upper and lower bounds of the correlation values ​​for each candidate correlation peak are calculated. The actual disturbance fluctuation equals the residual pressure fluctuation minus the executed pressure response error. Therefore, when the template sample is positive, the lower bound is selected from the upper error endpoint, and the upper bound from the lower error endpoint; when the template sample is negative, the endpoints are swapped; when the sample is zero, the contribution of this term is zero. Endpoints are selected and accumulated item by item. The lower and upper bounds of the correlation values ​​can be written as follows: , , in, and These respectively indicate that the candidate main peak is located in the unified sampling sequence number. Lower and upper bounds of the correlation value at time; This indicates the number of peaks after alignment, from the template start point backwards. The half-width of the error at the last sampling time corresponding to each sample is paired with the residual pressure fluctuation at that time. The absolute value of the template sample multiplied by the half-width of the error is the maximum increase or decrease of the product of the corresponding samples. The correlation value and upper and lower bounds are all in kPa². Each multiplication and addition retains double precision and is rounded outwards towards negative infinity and positive infinity respectively to obtain the lower and upper bounds. The display digits are not rounded before comparison. All candidates use the same template, sample number and error caliber; the template is fixed according to the measurement sequence obtained this time. Missing, saturated or noise exceeding the limit is blocked by step S1. The validity of candidate comparison is limited to the given template and the verified error set.

[0084] After all relevant value boundaries are formed, when the lower boundary of the correlation value of a candidate correlation peak is higher than the upper boundary of the correlation values ​​of the other candidate correlation peaks and the waveform difference meets the calibration waveform deviation range, the same pressure disturbance is confirmed by combining the pressure change direction, main peak sequence, and duration, and the corresponding terminal main peak time is determined as the actual response time. First, compare the candidate lower boundary with the maximum value of all other candidate upper boundaries. If it is strictly greater than the maximum value, the verification continues. If the boundaries are equal, overlap, or there are less than two comparable candidates, the confirmation is not made. Then, check the peak sign, whether the terminal main peak is later than the main peak of the previous measurement point, whether the start and end durations are within the calibration range, and whether the terminal main peak continues the propagation sequence of step S2. Then, check whether the main peak of the remaining pressure fluctuation coincides with the candidate position according to the main peak rule of step S1. If they do not coincide, the confirmation is not made. After all conditions are met, the actual response time is calculated according to the candidate sampling sequence number, and the event is locked to be output only once.

[0085] For example, within a reduced calculation window, the template pressure samples are -1 kPa, 2 kPa, and -1 kPa, respectively, and the remaining pressure samples are 0 kPa, -1 kPa, 2 kPa, -1 kPa, and 0 kPa, respectively. The nominal correlation values ​​of the three complete candidate windows are -4, 6, and -4, respectively. When the half-width of the pressure response error for each sample is 0.1 kPa, the maximum offset of each correlation value is 0.4, and the lower bound of the intermediate candidate is 5.6, which is higher than the upper bound of the other candidates, -3.6. When the half-width of the error expands to 3 kPa, the lower bound of the intermediate candidate becomes -6, and the upper bound of the other candidates becomes 8, making it impossible to confirm the actual response time. The actual operation still uses the complete pressure disturbance band for calculation, and the overlap of correlation value boundaries only produces unconfirmed results. Step S6 does not write the time correction amount based on this.

[0086] At this point, step S5 generates the actual response time after verification by the execution pressure response error, all candidate competition, and waveform relationships; unconfirmed events continue to retain the original pressure feedback control, and only the uniquely confirmed event enters the propagation time correction in step S6.

[0087] In step S6, the core task is to use the uniquely confirmed actual response time for the time correction of the next pressure disturbance, and to continue to adjust according to the current end pressure and the actual execution margin; the time parameter update and pressure adjustment check their respective conditions.

[0088] In this embodiment, after unique confirmation, the expected arrival time of the next pressure disturbance can be corrected by limiting the deviation between the actual response time and the expected arrival time. If not confirmed, the time correction is frozen. The opening of the air supply valve and the frequency of the blower are adjusted and controlled according to the terminal pressure deviation. Specifically, after unique confirmation of the actual response time, the propagation time correction amount is updated according to the time deviation between the actual response time and the expected arrival time. The single update amplitude is limited according to a preset ratio of the propagation duration corresponding to the continuous propagation delay. First, it is checked that the event corresponding to the actual response time is consistent with the predicted event retained in step S3. The actual response time is taken by subtracting the expected arrival time actually used in this instance. A positive deviation is obtained if the actual arrival is late, and a negative deviation is obtained if the actual arrival is early. If the same event has already been written into the propagation time correction amount, subsequent repeated feedback will only check the original record and will not be updated again. If the candidate is not uniquely confirmed or the operating range changes, the propagation time correction amount before the update is maintained.

[0089] To limit the changes caused by a single effective observation, the single update range of the propagation time correction is limited according to a preset ratio of the propagation time corresponding to the continuous propagation delay. The effective propagation delay from the first pressure measuring point to the last measuring point is accumulated, and 10% of the accumulated propagation time is used as the allowable increase or decrease range for a single observation. The updated propagation time correction can be written as: , in, This indicates the exact moment of the confirmed response to the current event. This indicates the percentage of a single update; in this embodiment, it is set to 0.1. Indicates the first The propagation time correction amount after this effective update is completed in the [number] operating intervals. This update is the cumulative [number]th [number]. This is the first effective update. The summation covers all adjacent effective propagation delays from the first pressure measurement point to the last measurement point, measured in seconds. Two extreme value operations limit the current time deviation to a positive or negative boundary equal to the cumulative propagation time multiplied by the update ratio. When the deviation exceeds the boundary, the boundary in the same direction is used. If the cumulative propagation time is not positive or the propagation chain is incomplete, the value before the update is retained. For example, with a cumulative propagation time of 0.8s, a correction amount of 0.02s before the update, and a current time deviation of 0.11s, an increment of 0.08s is allowed, and the correction amount after the update is 0.10s.

[0090] Subsequently, based on the peak time, continuous propagation delay, and remaining pipeline distance of the next pressure disturbance's upstream measuring point, the estimated arrival time is recalculated. During operating interval matching, the estimated arrival time is corrected using the propagation time correction after amplitude limiting. Specifically, sampling and propagation confirmation are fully executed, and then the remaining propagation time is obtained by matching the remaining pipeline speed with the sum of the peak time of the upstream measuring point. The peak time, continuous propagation delay, and actual response time of the previous pressure disturbance cannot be substituted into the propagation calculation of the next pressure disturbance; the propagation time correction is only read when the next pressure disturbance itself has a complete propagation chain. When the pipeline length, sensor response, or sampling period changes, the original calibration and propagation time correction are discontinued, and the matching history is re-established. Based on this, during the matching of operating intervals, the estimated arrival time is corrected using the propagation time correction after amplitude limiting. According to the estimated arrival time formula in step S3, the propagation time correction for the operating interval corresponding to the next pressure disturbance is added to the sum of the newly obtained peak time and the remaining propagation time. A positive value shifts the prediction backward, while a negative value shifts the prediction forward. If no valid operating interval is matched, the time correction is discontinued, and independent calibration for that operating interval continues to be obtained. The corrected estimated arrival time is then re-checked through the common effective time retrospective and future command check in step S3. If the propagation time correction causes any command time to expire, the advance arrangement for that event is revoked, without forcibly changing the command time to fabricate an advance effective time. When the historical deviation exceeds the current empirical upper limit, the error range of the corresponding interval is expanded, and the control lead is rechecked.

[0091] For the current pressure disturbance, the terminal pressure deviation is obtained after the actual response time. Based on the execution feedback, the remaining adjustment range of the air supply valve opening and blower frequency is determined. The remaining pressure fluctuation is aligned with the peak time of the terminal measurement point before the actual response time to confirm the disturbance's persistence. Specifically, the target pressure is subtracted from the original terminal pressure sample arriving at and after the unique confirmation completion time. The sampling time of the sample must be later than the actual response time; the old pressure before the confirmation calculation cannot be used. Subsequently, the remaining adjustment range of the air supply valve opening and blower frequency is determined based on the execution feedback. The margins between the allowable upper and lower limits are calculated using the latest actual opening and actual frequency, respectively. Then, an achievable trajectory is formed according to the sampling rate of change limit, and its intersection with the calibration range is determined. Actions that have been issued but whose feedback has not yet been completed are retained in the execution response reconstruction; unused command margins cannot be treated as released actual margins.

[0092] Based on this, the remaining pressure fluctuations are aligned with the peak time of the pre-terminal measurement point according to the actual response time to confirm the duration of the disturbance. The complete band of the pre-terminal measurement point is then shifted to the confirmed peak position, and the end position of the aligned end band and the remaining pressure fluctuations thereafter are read. If the end position has not yet crossed the current sampling time, or if there are still samples exceeding the static steady-state noise boundary after the end position, the disturbance continues. If the end position has crossed the boundary and five consecutive complete samples return to the static steady-state noise boundary, the disturbance is confirmed to have ended. Continuous counting is interrupted for missing samples, and a new set of five complete samples is obtained before further judgment is made, maintaining restricted feedback adjustment throughout the process.

[0093] Furthermore, while the disturbance persists, the adjustment amount corresponding to the terminal pressure deviation is allocated according to the remaining adjustment range. When the disturbance ends and the terminal pressure deviation is within the allowable range, the opening of the air supply valve and the blower frequency are gradually reverted to the previous values, and pneumatic delivery control is executed. Specifically, the pressure evaluation time is determined by the larger of the current time and the response time of the two paths. The original unit pressure response is read according to the difference between this time and each actual feedback time. Following the finite combination search in step S4, the pressure increment of the disturbance that is yet to appear is set to zero, and the pressure changes that have been executed but not yet reflected are included in the expected effect. The next instruction is calculated only within the remaining adjustment range of the two paths. When the terminal pressure deviation is within the allowable range, the option to maintain it is selected. When it exceeds the allowable range, the option to reduce the deviation without exceeding the execution boundary is selected. When there is no combination that can reduce the deviation, the increase in adjustment amount is stopped, and the pressure protection interlock is retained. When a new upstream pressure disturbance forms an effective propagation chain, the current recovery intention is terminated, and the new event is rearranged according to the current actual execution amount.

[0094] Furthermore, when the disturbance ends and the terminal pressure deviation is within the allowable range, the opening of the air supply valve and the blower frequency are gradually reverted to the previous control values, and pneumatic delivery control is executed. The previous control value is the last set of valid actual feedback before the first command is sent. Each sampling cycle moves only an increment corresponding to the allowable rate of change towards the previous control value, and the actual feedback after the reversion is read to determine whether to continue. If a new pressure disturbance occurs, the terminal pressure deviation exceeds the allowable range, or the feedback times out, the subsequent reversion is immediately stopped, and the current pressure deviation is returned to limited adjustment. The recovery ends after both actual feedbacks enter the feedback resolution range of the previous control value and are maintained for 5 consecutive valid samples.

[0095] At this point, step S6 limits the effective time observation to the prediction of the next pressure disturbance in the same operating range, and completes the continued adjustment or gradual recovery based on the current remaining pressure fluctuation, terminal pressure deviation and execution feedback. The next pressure disturbance starts again from synchronous sampling and propagation confirmation.

[0096] In summary, this method synchronously collects pressure time-series data along the pipeline and separates the pressure baseline. It confirms the same pressure disturbance through adjacent correlation, cross-segment verification, and turning propagation relationships. It predicts the terminal arrival time based on the remaining pipeline distance and schedules advance adjustments according to the response times of the air supply valve and blower. During the adjustment period, the executed pressure response is reconstructed and stripped based on actual execution feedback. All candidate correlation peaks are compared within the calibration error range. The predicted arrival time of the next pressure disturbance is corrected using the uniquely confirmed actual response time as a limiting factor. This ensures seamless integration of propagation identification, command execution, and feedback adjustment, reducing arrival misidentification caused by superimposed executed pressures and improving the correspondence between the timing of pneumatic conveying control and the propagation state of pressure disturbances.

[0097] In some embodiments of this application, a pneumatic conveying control method based on pressure wave correlation is applied to a positive pressure pneumatic conveying pipeline for transporting polyethylene particles from a feed hopper to a receiving hopper. An industrial controller connects to pressure sensors along the pipeline, an air replenishment valve actuator, and a blower frequency converter, reading pressure electrical signals, the actual air replenishment valve opening, and the actual output frequency of the frequency converter. The controller simulates and executes a single feed fluctuation control operation using set pipeline and identification parameters: pipeline inner diameter 80mm, feed rate 600kg / h, target end pressure 220kPa, allowable pressure deviation ±10kPa, and lower and upper pressure protection limits of 200kPa and 240kPa, respectively. Before starting the control, the actual air replenishment valve opening is set to 40%, the blower frequency to 35Hz, and the initial end pressure to 223.2kPa.

[0098] A complete implementation process may include the following steps: Step 1: Complete the calibration of the pressure acquisition channel and verify the current operating range.

[0099] Five pressure measuring points are set along the conveying direction, numbered sequentially from 1 to 5. The fourth pressure measuring point is designated as the pre-terminal measuring point, and the fifth pressure measuring point is located at the end. The centerline lengths of the three segments from the first pressure measuring point to the pre-terminal measuring point are 20m, 20m, and 24m respectively, with the turning point located between the second and third pressure measuring points. The remaining pipeline from the pre-terminal measuring point to the end is divided into two segments of 40m and 80m according to the turning point, and the length of each segment is checked together with the measuring point location. The pressure measuring ports avoid the gas injection jet. The five channels use response-matched flush diaphragm pressure sensors, with a range of 0 to 300 kPa for each.

[0100] Calibration pressures of 0 kPa, 150 kPa, and 300 kPa were applied to each channel. After subtracting the zero bias, the pressure was converted based on adjacent calibration points. The calibration error limit for the simulated channel was set to 0.010 kPa. The simulated calibration points were checked against this limit. The sum of the channel synchronization error and response difference was set to 0.003 s, which is less than the allowable value of 0.005 s. A 4 Hz analog low-pass filter was set before analog-to-digital conversion, and 20 Hz synchronous sampling was used for the pressure fluctuation envelope with the highest effective frequency of 2 Hz. Take 0.05s; write the sampling time to each channel according to the same trigger count.

[0101] The controller uses material type, feeding rate, initial air supply valve opening, and initial blower frequency together for operating range matching. This time, calibration records were read for polyethylene granules, feeding rates of 550 kg / h to 650 kg / h, initial air supply valve openings of 38% to 42%, and initial blower frequencies of 34 Hz ​​to 36 Hz. All three continuous ranges include the lower boundary but not the upper boundary, and the operating range number is 3. When receiving pressure data, the trigger count is checked frame by frame. This data has no missing frames, duplicate counts, or range saturation, allowing entry into pressure wave correlation. If any subsequent window contains an abnormal frame, that window is isolated, and the original end pressure is read again.

[0102] Step 2: Read the propagation records and execution response calibration results for the same operating range.

[0103] In this embodiment, 10 independent pressure disturbances were used to establish the propagation history, and 5 additional independent pressure disturbances were used for verification. During the calibration period, the air supply valve and blower remained stable. The historical ranges of the first three main peak time differences were 2.4s to 2.6s, 2.35s to 2.65s, and 2.9s to 3.1s, respectively. Each of these ranges expanded by 0.1s to both sides to form a propagation time window, and the results of the 5 verifications all fell within the corresponding propagation time windows. The calibration range of the duration ratio of adjacent complete bands was 0.90 to 1.10, and the cumulative deviation tolerance across segments was taken at 2 sampling points. The length ratio conversion deviation boundary from the second pressure measurement point to the third pressure measurement point was taken as 0.2s, and the next adjacent pipe segment was taken as 0.15s.

[0104] Historical nodes with propagation velocities of 7 m / s and 9 m / s before the turn correspond to propagation velocities of 7 m / s and 9 m / s after the turn, respectively. During this remaining pipeline calibration, a temporary synchronous pressure sensor was added at the split point to obtain the velocity ratio between each remaining pipe segment and the matched upstream reference pipe segment; both medians were 1.00. A 40 m remaining pipe segment was matched with the reference pipe segment from the first pressure measuring point to the second pressure measuring point, and an 80 m remaining pipe segment was matched with the reference pipe segment from the third pressure measuring point to the end measuring point. Two sets of endpoint correspondences were retained for arrival prediction readings. The air supply valve was simulated to execute independent forward and reverse actions at 1, 2, and 4 percentage points, and the blower at 0.2 Hz, 0.4 Hz, and 0.8 Hz, respectively. Ten effective action segments were taken for each amplitude and direction. For each action, the feedback from the other actuator was checked for stability and no upstream pressure disturbance. The action feedback starting point was determined by the first sample after three consecutive samples left the feedback noise band, and the pressure effective starting point was determined by the end pressure exceeding three times the pressure noise boundary along the calibration direction and maintaining three samples. The maximum cumulative duration of the two matching action groups is set to 0.55s and 1.35s respectively. After adding a sampling period, the response time of the air supply valve is set to 0.6s and the response time of the blower is set to 1.4s.

[0105] The 10 historical arrival residuals were taken as -1.3s, -0.8s, -0.4s, -0.2s, -0.1s, 0, 0.1s, 0.2s, 0.3s, and 0.4s, respectively. The estimated arrival time for each arrival was calculated from the upstream complete band already acquired at that time; the terminal complete band was only used to determine the corresponding actual response time. The controller linked these independent records to the current operating range, both to generate control lead and to determine the terminal-related search boundaries. Historical records showing operating condition mismatches or where the terminal peak could not be uniquely identified were not included in these two statistics.

[0106] The controller arranges the actual feedback increments into a causal identification matrix. At least 60 pressure samples are read for each unit response duration. After merging the amplitude actions in the same direction, a column rank check is performed, and then the original unit pressure response is solved. The simulation data is generated using the piecewise response given in step eight. After solving, the unit response is obtained by averaging 400 points, and saved separately for both forward and reverse directions. Five additional actions in each direction are taken to verify the error envelope, and the superposition range is checked using independent dual-action segments. The cumulative absolute value of the actual opening change (4 percentage points) and the cumulative absolute value of the actual frequency change (0.8 Hz) are both within the coverage of the dual-action verification, allowing the use of matched responses for subsequent calculations.

[0107] Step 3: Obtain the complete band of this pressure disturbance and form the available sampling times.

[0108] Each pressure channel uses the same sampling zero time, generating a unified sampling time based on trigger count and a 0.05s sampling period. The data reception time is only used to verify whether the data arrives on time and does not participate in the calculation of propagation time difference between the main peaks. The controller retains the current sampled values ​​for the original pressure, pressure baseline, and pressure fluctuation component. Subsequent control reads the original pressure and propagates the correlation by reading the pressure fluctuation component. If a sampling interruption occurs, 400 valid samples are first reacquired before the pressure measurement point is restored to participate in propagation confirmation.

[0109] The controller first reads 400 complete pressure samples to initialize the pressure baseline. Then, it incorporates the current sample and removes the earliest sample each time, always using only the current and past samples. The maximum absolute value of the pressure fluctuation component (0.090 kPa) is obtained from the 10-second stable sampling segment of the actuator. After adding the pressure calibration error, the static steady-state noise boundary is set at 0.1 kPa. When three consecutive samples exceed 0.2 kPa, the starting point is determined by backtracking; when five consecutive samples return to within 0.1 kPa, the previous sampling point is determined as the ending point. The 400-point length of the current pressure baseline remains unchanged during the event. In this study, the four upstream pressure measurement points obtained positive peaks in the same direction, with the main peak times being 114s, 116.5s, 119s, and 122s, respectively. The complete band of the end-point measurement point starts at 121.5s and ends at 122.5s, with a total of 21 sampling points. The pressure fluctuation component increases from 1 kPa to 11 kPa per point, and then decreases back to 1 kPa at the same amplitude. The other three upstream complete bands have the same sample morphology as this band and are aligned according to their respective peak times. Only after the terminal measurement point acquires five consecutive statically stable samples between 122.55s and 122.75s can the complete band be ready for output. The controller retains both the peak time at 122s and the complete band acquisition time at 122.75s; the propagation delay is read from the former, and the control executability check is read from the latter.

[0110] Step 4: Confirm the adjacent propagation delay and verify the relationship between the segment and the turning point.

[0111] The controller uses each upstream complete band as a template, and moves the downstream equal-length window point by point within the corresponding propagation time window. For each location, all 21 samples are used to calculate the normalized cross-correlation. The template energy in this case is 891 kPa², exceeding the product of the square of the static steady-state noise boundary and the number of samples (0.21 kPa²), thus qualifying for correlation calculation. The first pair of adjacent pressure measurement points achieves a unique maximum correlation value of 1.0000 at a point 50 sampling points later, with correlation values ​​on both sides of it at 0.9882. The peak sign and duration are consistent, confirming the effective propagation delay of the first segment as 2.5 s. Based on this, and considering the adjacent length ratio, the controller shortens the response time of the third pressure measurement point to 118.8 s to 119.2 s, and obtains the next adjacent propagation delay of 2.5 s within the intersection boundary.

[0112] The response time of the fourth pressure measurement point was further narrowed to 121.85s to 122.15s. A unique correlation peak was obtained by shifting 60 sampling points backward, resulting in an effective propagation delay of 3s. The most recent sampling time was taken from the left end of each intersection backward and the right end forward. The current boundary itself corresponds to the sampling position, and boundary samples are retained. Before a complete downstream window is obtained, the controller continues to supplement sampling, without writing the correlation results of the incomplete descent segment into the propagation chain. The controller also independently searches for the third pressure measurement point using the template of the first pressure measurement point, obtaining a cross-segment propagation delay of 5s; and searches for the fourth pressure measurement point using the template of the second pressure measurement point, obtaining a cross-segment propagation delay of 5.5s. The difference between the corresponding cumulative propagation delay and the cross-segment propagation delay is zero, which meets the tolerance of two sampling points. The propagation velocity before the turn is 8m / s. Interpolation between two historical nodes yields a propagation velocity of 8m / s after the turn, consistent with the main peak position of the third pressure measurement point. This results in consecutive propagation delays of 2.5s, 2.5s, and 3s, with the propagation chain ending at the pre-terminal measurement point. If multiple comparable correlation peaks appear, the controller first compares the cumulative deviation of integer sampling points, then compares the time difference between the deviation from the predicted position after the turn, and stops the current propagation confirmation when they are tied for the best.

[0113] Step 5: Predict the arrival time of the terminal and schedule two future instructions.

[0114] The controller pairs the three consecutive propagation delays with the corresponding pipe segment lengths, resulting in a propagation velocity of 8 m / s for each reference pipe segment. Using centerline lengths of 40 m and 80 m respectively and a calibration velocity ratio of 1.00 for the two remaining pipe segments, the estimated arrival times are calculated according to the aforementioned arrival relationship and can be written as: , The remaining propagation times for the two segments are 5 seconds and 10 seconds, respectively. 137s was obtained. No valid time update has occurred within this operating interval, the propagation time correction is zero, and the controller reserves 137s for instruction scheduling, correlation search, and post-arrival comparison. The difference between the actual response time and the expected arrival time in 10 historical matches is a minimum of -1.3s and a maximum of 0.4s, with a maximum absolute value of 1.3s. The upper bound of the transmission jitter used in the 100-trigger simulation is 0.1s. After increasing two sampling periods... Take 1.5s, and the common effective time is 135.5s.

[0115] The controller reads the response times of the air supply valve and the blower respectively, rounding up to the sampling time base to obtain the air supply valve command time as 134.9s and the blower command time as 134.1s. Propagation confirmation and planning calculation are completed at 122.8s, still 11.3s from the earlier command time, exceeding the 0.1s transmission time budget for this operation. Both commands can be triggered in the future. The controller classifies the blower as the first actuator and the air supply valve as the second actuator, and saves the planned trigger times for each. If a calculation delay causes either command time to arrive later, the untriggered plan is canceled, and the system switches to restricted feedback control based on the end-pressure deviation.

[0116] Step 6: Allocate the allowable execution amount according to the expected end-point changes.

[0117] In historical data matching, the median ratio of the amplitude of the main peak in the same direction between the end and the preceding measurement points is 0.80. The current 11 kPa main peak falls within the historical amplitude coverage range, and the expected end main peak is 8.8 kPa. The planned formation time is earlier than the expected start point of the end band, therefore... Take 8.8 kPa. The controller obtains this from the current end-point raw pressure. The pressure is 3.2 kPa. The original unit pressure response is read according to the direction of action, amplitude and expected arrival time. The pressure gain of the air supply valve is 2 kPa per percentage point and the pressure gain of the blower is 5 kPa per Hz.

[0118] The allowable opening of the air supply valve is 30% to 50%, and the allowable frequency of the blower is 30Hz to 40Hz. The candidate opening variation is taken as -4 to +4 percentage points, with an interval of 1 percentage point; the candidate frequency variation is taken as -0.8Hz to +0.8Hz, with an interval of 0.2Hz. The allowable change rates for the two paths are set to 20 percentage points per second and 4Hz, respectively, with an adjacent feedback interval of 0.05s. Each cycle allows an actual increment of 1 percentage point and 0.2Hz, respectively. Each candidate total change forms a planned trajectory with at most four increments. The total trajectory change and rate are both within the calibration range, and the above target combination is retained. The controller performs a finite search on 81 allowable combinations. The selected combination can be written as: , The selected combination is a 4% reduction in the air supply valve opening and a 0.8Hz reduction in the blower frequency. The expected pressure effects for the two paths are -8kPa and -4kPa, respectively, with an expected remaining terminal pressure deviation of zero. When equal remaining terminal pressure deviations exist, the absolute values ​​of the blower frequency change, the absolute values ​​of the air supply valve opening change, and the order of these changes are compared sequentially. In this case, the target air supply valve opening is selected as 36%, and the blower frequency as 34.2Hz. The controller first saves the actual feedback of 40% and 35Hz before adjustment, and then sends the blower command according to the plan.

[0119] Step 7: Recalculate the subsequent actions based on the actual feedback from the previous actions.

[0120] The blower command is triggered at 134.1s. The inverter begins to return the actual frequency change from 134.3s, decreasing by 0.2Hz at 134.3s, 134.35s, 134.4s, and 134.45s, ultimately reaching an actual frequency of 34.2Hz. The controller receives four actual changes according to the feedback sampling sequence number, retaining only one duplicate sequence number, and continuously receiving valid feedback without changes. There were no instances of updates not exceeding two sampling cycles, so the subsequent action plan remains valid. By 134.9s, the initial terminal pressure is 222.95kPa, and the current terminal pressure deviation is 2.95kPa. The four actual frequency changes correspond to current initial unit pressure responses of 0.5kPa, 0.375kPa, 0.25kPa, and 0.125kPa per Hz, respectively, with the expected arrival times all corresponding to 5kPa per Hz. The controller calculates the execution pressure change not yet reflected based on the actual feedback, which can be written as: , The blower branch outputs -3.75 kPa, while the actual blower pressure effect is -0.25 kPa. The controller retains the already executed blower action, incorporates the unreflected pressure changes into the terminal's projected effect, and re-searches for the remaining allowable opening of the air supply valve, still selecting a reduction of 4 percentage points. The air supply valve command is sent at 134.9 s, and the actual opening decreases by 1 percentage point every 0.05 s starting from 135.1 s, reaching 36% at 135.25 s. The actual changes in both paths and their feedback times continue to be processed in the execution pressure response reconstruction, without replacing the feedback time with the command time.

[0121] Step 8: Reconstruct and strip the execution pressure response using the same pressure baseline.

[0122] The controller reads the raw unit pressure response for the corresponding action direction and operating range, and separates the pressure baseline along with the 400 zero-response samples before the action using the same online method. The raw unit pressure response of the supplementary air valve, aligned with the feedback starting point, is zero for 0.2s and earlier, linearly increasing to 2 kPa per Hz up to 1s and then holding. The raw unit pressure response of the blower is zero for 0.4s and earlier, linearly increasing to 1 kPa per Hz up to 0.8s, holding until 1s, and then linearly increasing to 5 kPa per Hz up to 2s and then holding. All readings were taken within the valid calibration duration, and readings between adjacent calibration points were obtained using linear interpolation.

[0123] The two original unit pressure responses were separated by the same 400-point average baseline to obtain the unit response matched by the current feedback increment. The controller scaled, translated, and superimposed the four actual changes in the air supply valve and the four actual changes in the blower, resulting in the response at 137.25s. The initial pressure was -11.084 kPa, and the pressure fluctuation component at the same sampling time was -2.284 kPa. After stripping, the remaining pressure fluctuation was 8.8 kPa. When simulating the original pressure, the pre-action disturbance pressure increment was set to zero. For each sample, 400 times the current remaining pressure fluctuation was added to the sum of the previous 399 unseparated disturbance pressure increments, and then divided by 399 to obtain the current unseparated disturbance pressure increment. This was then added to the 223.2 kPa static steady pressure and the two original execution pressures. The remaining pressure fluctuation outside the band was set to zero, and the sliding window was updated according to the same rules. At this moment, the original pressure at the end was 220.13344 kPa, and the pressure baseline was 222.41744 kPa. The original pressure continued to be read by the pressure control. The controller repeated the above point-by-point processing for the entire candidate interval, so that each candidate used the same execution feedback and pressure baseline. If delayed feedback was received before the confirmation output, the corresponding change was added and all affected windows were recalculated; only the final correlation value could not be modified.

[0124] Step 9: Compare all candidate error bounds and confirm the actual response time.

[0125] Based on the estimated arrival time and matching historical residuals, after adding two sampling periods, the search interval for the terminal main peak is from 135.6s to 137.5s, with a total of 39 candidate sampling positions. Each candidate is aligned with the 21-point template and 10-point main peak offset of the terminal pre-measurement point, and the latest candidate needs to read samples from 138s. The controller waits for all windows to be complete and completes the confirmation calculation at 138.05s, while continuing to perform limited feedback judgment based on the original terminal pressure. At the candidate position at 137.25s, the residual pressure fluctuations of the 21 points after stripping form an amplitude ratio of 0.80 with the template, and both the rising and falling segments fall within the calibration allowable range of 0.75 to 0.85. The controller calculates the waveform difference based on the segmented reconstruction results, obtaining... The value is 0 kPa, which is lower than the allowable upper limit of 0.055 kPa formed after the pressure calibration error following the independent verification of the maximum waveform difference of 0.045 kPa. This candidate has a unique positive main peak, and the actual main peak is later than the main peak of the measurement point before the end. The start and end interval of the band is still 1 s, which satisfies the verification of the same direction, propagation order and duration.

[0126] Within this candidate window, the unit response error envelope of the air supply valve is set to 0.0045 kPa per percentage point, and the unit response error envelope of the blower is set to 0.0045 kPa per Hz. The total of the uncovered deviation, truncation margin, and online end-point pressure measurement margin used for independent dual-action verification is set to 0.012 kPa. The controller reads the absolute value of each feedback increment to obtain the point-by-point error half-width. Within the sampling interval covering 137.25s and its adjacent candidates, the error value is 0.0336 kPa. For earlier sampling times, the error envelope is read based on each actual feedback. The controller does not narrow this range but directly selects the error endpoints according to the sign of each sample in the template, and establishes a lower and upper bound for the correlation value for all 39 candidates. This yields a nominal correlation value of 712.8 kPa² for the candidate at 137.25s. It is 708.7344 kPa². The upper bound of the correlation coefficient is 716.8656 kPa²; the maximum upper bound of other candidate correlation values ​​is 708.0656 kPa², corresponding to candidates at 137.2 s and 137.3 s, respectively. The lower bound of the former is strictly higher than the upper bound of all other candidates, and the 0.6688 kPa² interval is still retained in the calculation without rounding.

[0127] The controller, based on the verified waveforms, confirms the actual response time as 137.25 seconds and locks the event to be output only once. Maintaining the template and residual pressure fluctuations unchanged, if the error half-width of all candidate samples is uniformly increased to 0.040 kPa, the lower limit of this peak is 707.96 kPa², while the maximum upper limit of other candidates is 708.84 kPa², which cannot be uniquely confirmed. The controller freezes the time correction and continues the restricted pressure feedback. When relevant thresholds are equal or overlap, the unconfirmed state is retained, and the nominal highest peak is prohibited from being written into the propagation time correction. When the proportion of a candidate waveform exceeds the limit, the relevant threshold of that candidate is still retained for competition, but that candidate is not allowed to become the final confirmed result; if there are fewer than two comparable candidates, the actual response time is also not output. During the calculation process, all candidates use the same template and error caliber, and outward rounding is used for each multiplication and addition. The result is displayed with the required decimal places only after the comparison is completed.

[0128] Step 10: Use the effective time deviation for the next pressure disturbance.

[0129] The controller verifies that the estimated arrival time of 137s and the actual response time of 137.25s belong to the same pressure disturbance, obtaining a time deviation of +0.25s. The cumulative propagation delay from the first pressure measuring point to the last measuring point is 8s. The single update ratio is taken as 0.1, and the allowable single increase or decrease is 0.8s. Updated according to the limit relationship, the updated propagation time correction can be written as: , The propagation time correction for operating interval number 3 is updated from 0 to 0.25s, and the effective update count changes from 0 to 1. The estimated arrival time remains at 137s, and repeated confirmation messages will not increase the propagation time correction again. Subsequent new pressure disturbances in the same operating interval must re-acquire the complete propagation chain. Following the estimated arrival time formula in step S3, 0.25s is added to the sum of the newly acquired peak time of the pre-terminal measuring point and the remaining propagation time. Other operating intervals do not read the update results. When sensor response, pipeline length, or sampling period changes, the original calibration and propagation time correction are deactivated.

[0130] Step 11: Complete the recovery based on the confirmed end pressure and execution feedback.

[0131] Upon completion, the controller reads the newly acquired initial terminal pressure of 211.45254 kPa at 138.05 s, with a terminal pressure deviation of -8.54746 kPa, still within the allowable range. Aligning the complete waveband of the terminal measurement point with the actual response time of 137.25 s, the terminal waveband end position is obtained at 137.75 s. All five complete residual pressure fluctuation samples from 137.8 s to 138 s return to the static steady-state noise boundary, confirming the end of this pressure disturbance. Simultaneously, the controller checks the latest actual air supply valve opening of 36% and the actual blower frequency of 34.2 Hz, calculates the margin to the allowable upper and lower limits, and forms a recovery trajectory that satisfies the sampling change rate and calibration range.

[0132] Under the condition that the disturbance ends and the terminal pressure deviation is within acceptable limits, the controller simulates successive callbacks starting from 138.1s. Each time, the air supply valve increases by 1 percentage point and the blower increases by 0.2Hz. The actual feedback after each callback is checked before submitting the next callback. The four callback commands were issued at 138.1s, 138.35s, 138.6s, and 138.85s, respectively, and the corresponding feedback arrived at 138.3s, 138.55s, 138.8s, and 139.05s. The actual changes did not exceed the rate of change limit. At 139.05s, the two actual feedbacks returned to 40% and 35Hz, and then five consecutive valid samples remained within the feedback resolution range until the end of this recovery at 139.25s. During the recovery period, the original terminal pressure and actual feedback are still read. If the pressure exceeds the process allowable range, the feedback times out, or a new valid pressure disturbance occurs, the subsequent callbacks are immediately stopped and the current terminal pressure deviation is adjusted under limited conditions. Subsequent pressure acquisition continues.

[0133] Through the complete implementation process described above, this method confirms the propagation evidence of the same upstream pressure disturbance as two control timing sequences that can be actually triggered. It uses the feedback from the first action to correct the subsequent action, reducing redundant compensation caused by pressure effects not yet being reflected. By performing pressure response stripping and comparing all candidates within the error range to jointly constrain the actual response time, the pressure change resulting from control does not directly serve as the basis for the arrival of the upstream pressure disturbance. The time deviation after unique confirmation only enters the subsequent prediction within the same operating range, which helps reduce the transmission of errors in continuous control, improves the correspondence between the pressure propagation state of the pneumatic pipeline and the timing of end-point control, and enhances the continuity of regulation after disturbances.

[0134] like Figure 2 As shown, another embodiment of this application provides a pneumatic conveying control system based on pressure wave correlation. This system includes functional modules corresponding to each step of the aforementioned pneumatic conveying control method based on pressure wave correlation. The specific data processing procedures for each module can be found in the aforementioned method embodiments.

[0135] Based on the same concept, another embodiment of this application provides a pneumatic conveying control system based on pressure wave correlation, comprising: Pressure acquisition module 1 is used to synchronously acquire pressure time-series data from pressure measuring points in the pneumatic conveying pipeline and separate the pressure baseline to obtain pressure fluctuation components. The time delay correlation module 2 is used to perform pressure wave correlation on the pressure fluctuation components of adjacent pressure measurement points within the propagation time window, confirm the propagation delay of the same pressure disturbance, and thereby shorten the response time of the next pressure measurement point to obtain the continuous propagation delay. The arrival prediction module 3 is used to take the nearest upstream pressure measuring point as the terminal front measuring point, determine the expected arrival time based on the peak time of the terminal front measuring point, the continuous propagation delay and the remaining pipeline distance, and arrange the control sequence in advance according to the response time of the air replenishment valve and the response time of the blower. The control and distribution module 4 is used to allocate the opening degree of the air supply valve and the frequency of the blower according to the control sequence, based on the pressure fluctuation component of the end measuring point and the end pressure deviation, and to perform control before the same pressure disturbance arrives. Response verification module 5 is used to separate the execution pressure response from the pressure fluctuation component of the end pressure measurement point according to the execution feedback and calibration pressure response, associate the remaining pressure fluctuation with the execution pressure wave of the pressure fluctuation component of the end measurement point, and determine the actual response time only when the same correlation peak is always unique within the execution pressure response error range. The closed-loop correction module 6 is used to correct the expected arrival time of the next pressure disturbance according to the deviation between the actual response time and the expected arrival time after unique confirmation. If not confirmed, the correction is frozen. The opening degree of the air supply valve and the frequency of the blower are corrected and controlled according to the end pressure deviation.

[0136] The above embodiments are used to illustrate the technical solutions of this application. Equivalent substitutions, improvements, or combinations made by those skilled in the art based on the disclosure of this application shall still fall within the protection scope defined by this application.

Claims

1. A pneumatic transport control method based on pressure wave correlation, characterized in that, include: Step S1: Synchronously collect pressure time-series data from pressure measuring points in the pneumatic conveying pipeline and separate the pressure baseline to obtain the pressure fluctuation component; Step S2: Perform pressure wave correlation on the pressure fluctuation components of adjacent pressure measuring points within the propagation time window, confirm the propagation delay of the same pressure disturbance, and shorten the response time of the next pressure measuring point accordingly to obtain the continuous propagation delay; Step S3: Take the nearest pressure measuring point upstream of the terminal as the terminal front measuring point, determine the estimated arrival time based on the peak time, continuous propagation delay and remaining pipeline distance of the terminal front measuring point, and arrange the control sequence in advance according to the response time of the air replenishment valve and the response time of the blower. Step S4: According to the control sequence, allocate the opening degree of the air supply valve and the frequency of the blower based on the pressure fluctuation component of the terminal measuring point and the terminal pressure deviation, and perform control before the same pressure disturbance arrives; Step S5: Based on the execution feedback and calibration pressure response, separate the execution pressure response from the pressure fluctuation component of the end pressure measurement point, associate the remaining pressure fluctuation with the execution pressure wave of the pressure fluctuation component of the end measurement point, and determine the actual response time only when the same correlation peak is always unique within the execution pressure response error range. Step S6: After unique confirmation, the expected arrival time of the next pressure disturbance is corrected according to the deviation between the actual response time and the expected arrival time. If not confirmed, the time correction is frozen. The opening of the air supply valve and the frequency of the blower are adjusted according to the end pressure deviation and the control is executed.

2. The pneumatic conveying control method based on pressure wave correlation as described in claim 1, characterized in that, Step S2 specifically includes: Within the propagation window, cross-correlation calculations are performed on the pressure fluctuation components of the first pair of adjacent pressure measurement points. The effective propagation delay of the first segment is determined based on the relationship between the pressure change direction, main peak order, and duration of the correlation peaks. Starting from the next pair of adjacent pressure measuring points, the propagation time is calculated based on the previous effective propagation delay and the ratio of the length of the adjacent pipe segment. The propagation time is then correlated with the upstream peak time, and the response time of the next pressure measuring point is shortened within the propagation time window. Calculate the relevant peak during the response period of the next pressure measuring point, perform cross-correlation calculation on the pressure fluctuation components separated by one pressure measuring point to obtain the cross-segment propagation delay, and accumulate the propagation delay of adjacent pipe segments between the same start and end measuring points to obtain the cumulative propagation delay; When multiple related peaks appear in the same response period, candidate related peaks are selected based on the deviation between the cumulative propagation delay and the cross-segment propagation delay. Valid related peaks are determined based on the direction of pressure change, the order of the main peaks, and the duration. Continuous propagation delays are formed according to the pipeline sequence.

3. The pneumatic conveying control method based on pressure wave correlation as described in claim 2, characterized in that, When multiple related peaks appear in the same response period, candidate related peaks are screened based on the deviation between the cumulative propagation delay and the cross-segment propagation delay. Valid related peaks are determined based on the direction of pressure change, the order of the main peaks, and their duration. The steps for forming a continuous propagation delay according to the pipeline sequence include: Based on the lengths of adjacent pipe segments on both sides of the pipeline turning position, the effective propagation delay corresponding to a single historical pressure disturbance is converted into the propagation velocity before and after the turning, and a correspondence between the propagation velocity before and after the turning is established. Based on the interpolation of the current propagation speed before turning in the corresponding relationship, the propagation speed after turning and the corresponding propagation time are determined. The current upstream main peak time is shifted backward along the corresponding propagation time to determine the predicted position of the main peak after turning. Connect adjacent pipe segments on both sides of the pipeline turning position with the propagation delay corresponding to each candidate correlation peak, calculate the deviation between the cumulative propagation delay and the cross-segment propagation delay between the same start and end measurement points, and calculate the time difference between the candidate correlation peak and the predicted position. The candidate correlation peaks are selected by the deviation between the cumulative propagation delay and the cross-segment propagation delay from smallest to largest. For candidate correlation peaks with the same deviation, they are compared by the time difference from smallest to largest. Only when the direction of pressure change, the order of the main peaks and the duration match and the candidate correlation peak is unique are the effective correlation peaks confirmed and a continuous propagation delay is formed.

4. The pneumatic conveying control method based on pressure wave correlation as described in claim 1, characterized in that, Step S3 specifically includes: Pair the continuous propagation delay with the corresponding pipe segment length, calculate the propagation velocity of each pipe segment upstream of the end measuring point, and match the calibration velocity ratio of the remaining pipe segments according to the pipe turning position. The propagation time of the remaining pipeline distance is calculated based on the ratio of the propagation speed of each pipe segment to the calibration speed. The propagation time is then superimposed on the peak time of the measuring point before the end to form the estimated arrival time. Match historical single pressure disturbances to the current operating range, determine the upper limit of arrival time error by the maximum absolute value of the deviation between the historical estimated arrival time and the historical actual response time, and advance the estimated arrival time by the control lead amount greater than the upper limit of arrival time error to determine the common effective time; Based on the common effective time, the response time of the air supply valve and the response time of the blower are traced back to obtain the command time of the air supply valve and the command time of the blower. The control timing is determined by the order of the air supply valve command and the blower command. When both commands are later than the current time, the corresponding execution trigger is set.

5. The pneumatic conveying control method based on pressure wave correlation as described in claim 4, characterized in that, Based on the common effective time, the steps of tracing back along the response time of the air supply valve and the blower to obtain the command time of the air supply valve and the command time of the blower specifically include: During the historical operation period when the air supply valve changes its opening independently and the blower frequency is stable, the air supply valve operation delay is determined based on the air supply valve command time and the air supply valve feedback start point. The air supply valve pressure delay is determined based on the air supply valve feedback start point and the moment when the end pressure change first reaches the calibrated effective amplitude. During the historical operation period in which the blower frequency is changed independently and the air supply valve opening is stable, the blower operation delay is determined based on the blower command time and the blower feedback start point. The blower pressure delay is determined based on the time when the blower feedback start point and the end pressure change first reach the calibrated effective amplitude. Match historical action segments according to the current operating range, take the upper limit of the cumulative duration of the air supply valve action delay and the air supply valve pressure delay as the air supply valve response time, and take the upper limit of the cumulative duration of the blower action delay and the blower pressure delay as the blower response time. Using the common effective time as the endpoint, trace back along the response time of the air supply valve and the response time of the blower, and round forward according to the sampling period to determine the command time of the air supply valve and the command time of the blower.

6. The pneumatic conveying control method based on pressure wave correlation as described in claim 5, characterized in that, Step S4 specifically includes: Based on the historical amplitude ratio, the pressure fluctuation component of the terminal measuring point is converted into the terminal pressure change at the expected arrival time, and the opening degree of the air supply valve and the frequency of the blower are allocated in combination with the terminal pressure deviation. When the air supply valve command time is earlier than the blower command time, the terminal pressure deviation is obtained at the blower command time. The execution pressure change that has not yet been reflected is predicted according to the actual opening change of the air supply valve. The execution pressure change is included in the terminal pressure change, and the blower frequency is updated in combination with the terminal pressure deviation. When the blower command time is earlier than the air supply valve command time, the terminal pressure deviation is obtained at the air supply valve command time. The execution pressure change that has not yet been reflected is predicted according to the actual frequency change of the blower. The execution pressure change is included in the terminal pressure change, and the air supply valve opening is updated in combination with the terminal pressure deviation. The amplitude and rate of change of the air supply valve opening and the blower frequency are limited. Pneumatic conveying control is performed at the time of the air supply valve command and the time of the blower command, and is performed synchronously when the two are the same.

7. The pneumatic conveying control method based on pressure wave correlation as described in claim 1, characterized in that, Step S5 specifically includes: Read the calibration pressure response of the individual operation of the air supply valve and the blower, match the action direction and operating range according to the execution feedback, identify the unit response by execution change, process it according to the baseline separation method of pressure time series data, and obtain the response sequence corresponding to the unit execution change; Based on the feedback time of the actual change in the air supply valve opening and the actual change in the blower frequency, the response sequence is scaled according to the change amount of each execution and superimposed along the time axis to obtain the execution pressure response at each sampling time. The pressure response is extracted point by point from the pressure fluctuation components of the end pressure measurement point. The relevant search intervals on both sides of the expected arrival time are determined according to the historical arrival time deviation. Pressure wave correlation is performed on the remaining pressure fluctuation and the pressure fluctuation components of the end measurement point. Within the error range of the pressure response, relevant peaks are compared. Based on the direction of pressure change, the order of main peaks, and the duration, a unique relevant peak that conforms to the continuous propagation delay order is identified, and the corresponding terminal main peak time is determined as the actual response time.

8. The pneumatic conveying control method based on pressure wave correlation as described in claim 7, characterized in that, The steps of comparing relevant peaks within the error range of the pressure response, identifying unique relevant peaks that conform to the continuous propagation delay order based on the direction of pressure change, the order of main peaks, and their duration, and determining the corresponding terminal main peak time as the actual response time, specifically include: Each sampling time within the relevant search interval is used as the candidate relevant peak position. The remaining pressure fluctuation and the pressure fluctuation components of the terminal measurement point are aligned according to the corresponding propagation time to determine the rising segment and the falling segment. Calculate the amplitude ratio of the rising segment and the falling segment relative to the corresponding waveband of the end measuring point, reconstruct the end pressure waveform according to the amplitude ratio, and compare the reconstructed end pressure waveform with the remaining pressure fluctuation to obtain the waveform difference; The error range of the executed pressure response at each sampling time is determined by the repeated measurement difference of the calibrated pressure response. Unnormalized cross-correlation is used for calculation. The endpoints of the error range are selected according to the sign of the pressure fluctuation component at the end measurement point. The upper and lower bounds of the correlation values ​​of each candidate correlation peak are calculated. When the lower bound of the correlation value of a candidate correlation peak is higher than the upper bound of the correlation values ​​of other candidate correlation peaks and the waveform difference meets the calibration waveform deviation range, the same pressure disturbance is confirmed by combining the pressure change direction, the order of the main peaks and the duration, and the corresponding terminal main peak time is determined as the actual response time.

9. The pneumatic conveying control method based on pressure wave correlation as described in claim 8, characterized in that, Step S6 specifically includes: After the actual response time is uniquely confirmed, the propagation time correction is updated according to the time deviation between the actual response time and the expected arrival time, and the single update amplitude is limited according to the preset ratio of the propagation duration corresponding to the continuous propagation delay. Based on the peak time of the last measuring point before the next pressure disturbance, the continuous propagation delay, and the remaining pipeline distance, the estimated arrival time is calculated, and the estimated arrival time is corrected by the propagation time correction after amplitude limiting during the matching of the operating range. After the actual response time, the terminal pressure deviation is obtained, and the remaining adjustment range of the air supply valve opening and blower frequency is determined according to the execution feedback. The remaining pressure fluctuation is aligned with the peak time of the terminal front measurement point according to the actual response time to confirm the disturbance persistence status. During the ongoing disturbance, the adjustment amount corresponding to the end pressure deviation is allocated according to the remaining adjustment range. When the disturbance ends and the end pressure deviation is within the allowable range, the air supply valve opening and blower frequency are gradually adjusted back to the previous values ​​and pneumatic delivery control is executed.

10. A pneumatic conveying control system based on pressure wave correlation, characterized in that, include: The pressure acquisition module is used to synchronously acquire pressure time-series data from pressure measuring points in the pneumatic conveying pipeline and separate the pressure baseline to obtain the pressure fluctuation component. The time delay correlation module is used to perform pressure wave correlation on the pressure fluctuation components of adjacent pressure measurement points within the propagation time window, confirm the propagation time delay of the same pressure disturbance, and thereby shorten the response time of the next pressure measurement point to obtain the continuous propagation time delay. The arrival prediction module is used to determine the estimated arrival time based on the peak time of the nearest upstream pressure measurement point of the terminal, the continuous propagation delay, and the remaining pipeline distance of the terminal measurement point, and to arrange the control sequence in advance according to the response time of the air supply valve and the response time of the blower. The control and distribution module is used to allocate the opening degree of the air supply valve and the frequency of the blower according to the control sequence, based on the pressure fluctuation component of the end measuring point and the end pressure deviation, and to perform control before the same pressure disturbance arrives. The response verification module is used to separate the execution pressure response from the pressure fluctuation component of the end pressure measurement point according to the execution feedback and calibration pressure response, associate the remaining pressure fluctuation with the execution pressure wave of the pressure fluctuation component of the end measurement point, and determine the actual response time only when the same correlation peak is always unique within the execution pressure response error range. The closed-loop correction module is used to correct the expected arrival time of the next pressure disturbance by limiting the deviation between the actual response time and the expected arrival time after unique confirmation. If not confirmed, the correction is frozen. The module also corrects the opening of the air supply valve and the frequency of the blower based on the end pressure deviation and performs regulation.