A resonant cavity pressure feedback control method for infrasound wave generation

Abstract

The application discloses a resonant cavity pressure feedback control method for infrasound wave generation and relates to the technical field of industrial control. The resonant cavity pressure feedback control method for infrasound wave generation comprises the following steps: S1, obtaining pre-wave cavity pressure observation data and performing time axis repositioning and period scale unification; S2, calculating a peak value phase difference, and evaluating pre-wave pressure window adaptation; S3, evaluating cavity pressure loss sequence deviation, and performing cavity pressure loss state feature matching; and S4, constructing a control direction matrix, writing a closed-loop control vector into a controller, and generating a lock identification. The application effectively improves the stability of resonant cavity pressure concentration, pressure relief control and frequency lowering in the infrasound wave generation process, and solves the problem that existing infrasound wave generating equipment can only rely on driving frequency or driving power adjustment, and it is difficult to determine whether an effective pressure window is formed in front of a quarter wavelength in the cavity.

Description

Technical Field

[0001] This invention relates to the field of industrial control technology, specifically to a resonant cavity pressure feedback control method for infrasound generation. Background Technology

[0002] With the development of low-frequency acoustic excitation equipment, resonant cavity structure modulation, and pressure feedback control, infrasound generation places higher demands on the sequence of pressure transmission within the cavity, the timing of outlet release, and the suppression of residual pressure. Existing equipment typically only detects external sound pressure or spectral results, lacking a closed-loop assessment of the transmission relationship between inlet pressure, central axis pressure, outlet pressure, and external release pressure. During the target infrasound frequency reduction process, existing systems are prone to insufficient inlet coupling, insufficient central axis concentration, premature outlet pressure release, intracavity hysteresis, and abnormal echo standing waves, leading to unstable infrasound output, poor reproducibility of control parameters, and insufficient energy conversion efficiency.

[0003] For example, invention patent CN114706431B discloses a pressure control method, device, and semiconductor process equipment for a reaction chamber, including: acquiring the actual gas pressure value of the reaction chamber; calculating the pressure change rate of the actual gas pressure value within a preset time in real time, and determining the corresponding hysteresis conversion coefficient based on the pressure change rate, wherein the hysteresis conversion coefficient is negative and its absolute value is positively correlated with the absolute value of the pressure change rate; multiplying the hysteresis conversion coefficient by the coefficient of the initial feedback control equation to obtain a modified feedback control equation; calculating the opening adjustment amount of the pressure regulating valve based on the modified feedback control equation; and adjusting the opening of the pressure regulating valve based on the opening adjustment amount until the actual gas pressure value reaches the target pressure. The method provided by this invention can reversely adjust the opening of the pressure regulating valve before the gas pressure reaches the target pressure, avoiding pressure overshoot and improving the pressure stability of the reaction chamber. This invention also provides a pressure control device and semiconductor process equipment for a reaction chamber.

[0004] For example, the invention patent with announcement number CN118860001B discloses a method and system for intelligent feedback control of pressure and deformation in a ceramic flattening device, including: collecting ceramic physical data of the ceramic; dividing the ceramic into m regions; obtaining the pressure values ​​required for each of the m regions during the ceramic flattening process; controlling the ceramic flattening device to flatten the ceramic according to the required pressure values ​​for each of the m regions, and collecting and feeding back deformation data of the m regions in real time during the ceramic flattening process; determining whether an adjustment command is generated based on the real-time feedback deformation data, marking the region where the adjustment command is generated as the adjustment region, calculating the adjustment pressure of the adjustment region, and controlling the pressure value of the adjustment region in the ceramic flattening device to be adjusted to the adjustment pressure; this invention can achieve individualized and refined control of pressure values, effectively reducing uneven stress on the ceramic, thereby improving the flattening quality and accuracy of the ceramic.

[0005] In existing technologies, there is a general lack of measurable pressure indicators and adjustable boundary constraints for the pressure transmission stage. It is difficult to form a clear allowable range and pressure window for intracavity pressure and to make stable control. As a result, energy is released or dissipated before the target low-frequency concentration is formed, resulting in unstable output, efficiency fluctuations and insufficient repeatability.

[0006] Therefore, in order to address the above problems, there is an urgent need for a resonant cavity pressure feedback control method for infrasound generation. Summary of the Invention

[0007] Technical problems to be solved To address the shortcomings of existing technologies, this invention provides a resonant cavity pressure feedback control method for infrasound generation, which solves the problem that existing infrasound generating devices rely solely on driving frequency or driving power adjustment, making it difficult to determine whether the compression and sparse waves in the cavity form an effective pressure window before a quarter wavelength.

[0008] Technical solution To achieve the above objectives, the present invention provides the following technical solution: a resonant cavity pressure feedback control method for infrasound generation, comprising: S1, acquiring pre-formed wave cavity pressure observation data, performing time axis repositioning and period scale unification, and generating a pre-formed wave pressure chain dataset; S2, based on the pre-formed wave pressure chain dataset, calculating the peak phase difference, evaluating the pre-formed wave pressure window adaptation, and constructing a pre-formed wave pressure window dataset; S3, based on the pre-formed wave pressure window dataset, evaluating the cavity pressure loss sequence deviation, performing cavity pressure loss state feature matching, and generating cavity pressure loss state data and boundary constraint data; S4, based on the boundary constraint data, constructing a control direction matrix, writing the closed-loop control vector into the controller, and generating a locking identifier and parameter fingerprint.

[0009] Further, the specific steps for acquiring pre-formed wave cavity pressure observation data and performing time axis repositioning and period scale unification to generate a pre-formed wave cavity pressure observation dataset are as follows: Collect the inlet pressure, central axis pressure, outlet pressure, residual pressure, and external release pressure during the infrasound generation process; set the cavity pressure loss type, target infrasound frequency, and corrected sound velocity; calibrate the resonant cavity inlet position, central axis pressure acquisition point, outlet pressure acquisition point, equivalent propagation distance of phase coordinates, and calibrate the propagation velocity; calculate the central axis pressure peak, inlet pressure peak, outlet pressure peak, and the time corresponding to the peak values; and combine all acquired, set, calibrated, and... The calculated data are aggregated to form pre-wave cavity pressure observation data. Based on the time axis relocation method of the drive cycle synchronization mark, the pressure sampling time is corrected, and the pressure data within the same target infrasound frequency period is mapped to a unified period coordinate. Through the pressure change rate constrained filtering method, the spike noise generated at the moment of drive source start-up, the airflow impact at the phase guide port, and the residual pressure fall-off stage in the cavity is dynamically denoised. Through the peak amplitude scaling and phase coordinate mapping method based on the target infrasound frequency period segmentation, the pre-wave cavity pressure observation data is normalized, and the pressure data is unified to a period scale to generate a pre-wave pressure chain dataset.

[0010] Furthermore, based on the pre-formed wave pressure chain dataset, the specific steps for calculating the peak phase difference are as follows: Based on the pre-formed wave pressure chain dataset, the structural distance from the resonant cavity inlet to the central axis pressure acquisition point is obtained through sensor installation position calibration; the allowable deviation of the structural distance is obtained by synthesizing the assembly tolerance and frequency drift equivalent distance using the root mean square method; the central axis pressure gain is obtained by the ratio of the central axis pressure peak value to the inlet pressure peak value; the actual phase difference from the inlet to the central axis is obtained by converting the time difference between the inlet pressure peak time and the central axis pressure peak time to the current target infrasonic frequency periodic phase coordinate; and the final phase difference is obtained by... The time difference between the peak pressure at the axial axis and the peak pressure at the outlet is converted to the current target infrasound frequency period phase coordinate to obtain the actual phase difference from the axial axis to the outlet; the expected phase difference from the inlet to the axial axis is determined by the target infrasound frequency, the equivalent propagation distance from the inlet to the axial axis, and the calibrated propagation speed; the expected phase difference from the axial axis to the outlet is determined by the target infrasound frequency, the equivalent propagation distance from the axial axis to the outlet, and the calibrated propagation speed; the residual pressure at the end of the cycle is obtained by comparing the residual pressure at the axial axis and the pressure at the outlet at the end of the current cycle with the peak pressure of the current cycle.

[0011] Furthermore, the specific steps for evaluating the pre-formed wave pressure window adaptation are as follows: The difference between the structural distance and the quotient of the target infrasound frequency and four times the corrected sound velocity is calculated. This difference is then divided by the sum of the allowable deviation of the structural distance and the zero-prevention term, squared, and then the negative exponent is taken to obtain the wavelength position fitting term. The central axis pressure gain is increased by one, the natural logarithm is taken, and then divided by the result of adding the natural logarithm to obtain the central axis pressure enhancement term. The difference between the actual phase difference from the inlet to the central axis and the expected phase difference from the inlet to the central axis is taken, divided by two, and then the sine value is taken and the absolute value is taken to obtain the phase deviation from the inlet to the central axis term. The phase deviation term from the centerline to the outlet is obtained by subtracting the actual phase difference from the centerline to the outlet from the expected phase difference, dividing by two, taking the sine value, and then taking the absolute value. The two phase deviation terms are added together and the negative exponent is taken to obtain the phase transmission coordination term. The residual pressure ratio at the end of the cycle is subtracted from the upper limit of the allowable residual pressure, multiplied by the residual pressure gating coefficient, and then exponentially calculated. After adding a constant one and taking the reciprocal, the pressure suppression term is obtained. The wavelength position matching term, the centerline pressure enhancement term, the phase transmission coordination term, and the pressure suppression term are multiplied in sequence to obtain the pre-formed wave pressure window adaptation value.

[0012] Furthermore, the specific steps for constructing the preformed wave pressure window dataset are as follows: when the preformed wave pressure window adaptation value is less than the preformed wave adaptation threshold, it is determined that no effective preformed wave pressure window has been formed in the current period, and the current period is marked as an invalid pressure window period; when the preformed wave pressure window adaptation value is greater than or equal to the preformed wave adaptation threshold, it is determined that an effective preformed wave pressure window has been formed in the current period, and the current period is marked as an effective pressure window period; periodic pressure window labels are formed and summarized in order of target infrasound frequency to construct the preformed wave pressure window dataset.

[0013] Furthermore, based on the pre-formed wave pressure window dataset, the specific steps for evaluating the cavity pressure loss sequence deviation are as follows: Based on the pre-formed wave pressure window dataset, the controller divides the sampling period according to the target infrasound frequency and sequentially numbers it to obtain the target infrasound frequency period number; through the cavity pressure loss type enumeration mapping method, abnormal states are assigned state numbers to obtain candidate cavity pressure loss state numbers; the candidate cavity pressure loss state numbers and their corresponding peak amplitude ratio, phase difference, phase interval within the target infrasound frequency period, and upper and lower limits of actuator adjustment are written into the same state mapping table to obtain a candidate cavity pressure loss state set; the peak times of inlet pressure, central axis pressure, and outlet pressure in the current period are read and sorted according to time to obtain the actual occurrence order; the typical occurrence order of inlet pressure peak, central axis pressure peak, and outlet pressure peak under each type of cavity pressure loss state is calibrated to obtain the standard peak order; the number of adjacent position exchanges required to adjust the actual peak order to the standard peak order is calculated to obtain the sequence deviation value.

[0014] Further, the specific steps for matching the cavity pressure loss state characteristics are as follows: Linearly normalize the inlet pressure, central axis pressure, outlet pressure, external release pressure, and residual pressure of the current cycle to obtain state characteristic quantities; statistically analyze the sample cycles under the cavity pressure loss state to obtain standard characteristic values; extract the fluctuation amplitude of each state characteristic quantity under the same cavity pressure loss state, and use the quantile deviation of the corresponding characteristic quantity as the allowable deviation of the characteristic under the state; detect windows exceeding the anomaly judgment conditions in the inlet compression window, central axis concentration window, outlet release window, and residual pressure dissipation window respectively to obtain the pressure window that triggers the anomaly; record the allowable first-time error for each type of cavity pressure loss state using the state-window correspondence calibration method. The abnormal pressure window range is used to obtain the trigger window set. The standard characteristic value corresponding to the candidate cavity pressure loss state is subtracted from each state characteristic quantity, and the absolute value is divided by the sum of the characteristic allowable deviation and the zero-prevention term to obtain the deviation degree. The maximum value among all state characteristic quantities is selected as the maximum deviation term. It is determined whether the pressure window triggering the abnormality in the current cycle belongs to the trigger window set. If it does, the window penalty term is zero; if not, the trigger window mismatch penalty coefficient is used as the window penalty term. The sequence deviation value, the maximum deviation term, and the window penalty term are added together to obtain the total state deviation value. The candidate state with the smallest total state deviation value is selected to obtain the cavity pressure loss state matching sequence number.

[0015] Further, the specific steps for generating cavity pressure loss state data and boundary constraint data are as follows: Read the corresponding cavity pressure loss type based on the cavity pressure loss state matching number: When the cavity pressure loss state matching number equals the first state number, determine the inlet coupling insufficiency state, indicating that the drive source output has failed to be converted into inlet pressure, and initiate a review of the inlet connection, sealing state, and drive source coupling state; When the cavity pressure loss state matching number equals the second state number, determine the structural constraint insufficiency state, indicating that the inlet pressure has been established but not concentrated in the central axis region, and generate an adjustable range for the cavity angle state; When the cavity pressure loss state matching number equals the third state number, determine the outlet premature pressure relief state, indicating that the outlet... The front pressure is released earlier than the central axis pressure concentration process, causing the outlet release time to be delayed; when the cavity pressure loss state matching number is equal to the fourth state number, the cavity stagnation state is determined, indicating that the central axis pressure has been formed but has failed to be transmitted to the outside of the guide phase port, and the cavity pressure release path is opened; when the cavity pressure loss state matching number is equal to the fifth state number, the echo standing wave abnormal state is determined, and the driving amplitude range, frequency back-off range, and residual pressure dissipation waiting time are generated; if the cavity pressure loss state matching number does not belong to the above five state numbers, the state identification of the current cycle is determined to be invalid, and the generation of closed-loop regulation for the current cycle is stopped; attribution analysis is performed according to the current cavity pressure loss type, and boundary constraint data is generated simultaneously.

[0016] Further, based on boundary constraint data, the specific steps for constructing the control direction matrix are as follows: Determine the adjustment step size according to the cavity pressure loss state type and the number of abnormal cycles of the current pre-formed wave pressure window adaptation value; construct the control direction matrix using the state-to-execution mapping method; calculate the gain statistics of the peak value of the central axis pressure and the peak value of the inlet pressure, and take the median to obtain the target central axis pressure gain; calculate the difference between the peak value of the external release pressure and the peak value of the pressure before the outlet, and combine this with the time difference penalty from the peak value of the pressure before the outlet to the peak value of the external release pressure to obtain the target outlet release holding value; compare the amplitude of the peak value of the external release pressure with the peak value of the pressure before the outlet to obtain the actual outlet release holding value; read the residual amount of the central axis pressure and the residual amount of the pressure before the outlet at the end of the current cycle, and compare them with the total peak amount of the peak value of the central axis pressure and the peak value of the pressure before the outlet in the current cycle to obtain the cycle residual pressure ratio; select the pre-formed wave pressure window at the target infrasound frequency and... For periods without echo standing wave anomalies, the target residual pressure ratio is obtained through proportional calculation. The target axial pressure gain is then increased by a zero-prevention term, divided by the sum of the axial pressure gain and the zero-prevention term, and the natural logarithm is taken to obtain the axial pressure concentration error. The target outlet release holding value is increased by a zero-prevention term, divided by the sum of the outlet release holding value and the zero-prevention term, and the natural logarithm is taken to obtain the outlet release holding error. The residual pressure ratio is subtracted from the target residual pressure ratio to obtain the residual pressure error. The axial pressure concentration error, outlet release holding error, and residual pressure error are sequentially combined to form an error vector, and a hyperbolic tangent operation is performed on the error vector to obtain the error adjustment vector. The error adjustment vector is multiplied by the control direction matrix to obtain the correction direction. The correction direction is multiplied by the adjustment step size to obtain the control correction amount. The control vector of the current period is subtracted from the control correction amount to obtain the control vector of the next period. The control vector of the next period is restricted between the lower and upper control limits to obtain the closed-loop control vector.

[0017] Further, the specific steps for writing the closed-loop control vector into the controller and generating the locking identifier and parameter fingerprint are as follows: write the closed-loop control vector of the next cycle into the controller, adjust the output waveform, output intensity, cavity guide boundary position and phase guide port release section of the drive source, so that the inlet pressure peak, central axis pressure peak and outlet release peak re-enter the pre-formed wave pressure transmission sequence within the next target infrasonic frequency cycle; when the pre-formed wave pressure window adaptation value is greater than the pre-formed wave adaptation threshold, the cavity pressure loss state does not show the outlet premature pressure relief state and the echo standing wave abnormal state within m target infrasonic frequency cycles, a locking identifier is generated; after generating the locking identifier, write the parameter record according to the target infrasonic frequency cycle sequence, record the closed-loop control vector and target infrasonic frequency corresponding to each cycle in sequence, and form a parameter fingerprint.

[0018] Beneficial effects The present invention has the following beneficial effects: (1) This invention incorporates the central axis pressure gain, peak phase difference and residual pressure ratio at the end of the cycle corresponding to the target infrasound frequency into the calculation of the pre-wave pressure window adaptation value, so that the gas compression wave constraint process in the unstable wave formation stage can be quantitatively judged, thus solving the problem that existing equipment can only judge whether the infrasound output has been formed by external low-frequency peak values.

[0019] (2) This invention, by comparing the order of occurrence of the inlet pressure peak, the central axis pressure peak and the outlet pressure peak, and combining the inlet compression response, the central axis concentrated response, the outlet release response, the periodic residual pressure ratio and the pressure peak tailing time, identifies insufficient inlet coupling, insufficient structural constraints, premature outlet depressurization, intracavitary stagnation and abnormal echo standing waves, so that the cause of pressure loss can be located to the specific pressure transmission link, avoiding the generalization of different imbalance states to insufficient output.

[0020] (3) In this invention, the driving frequency, driving amplitude, cavity angle state and equivalent opening of the phase guide port are jointly corrected by the closed-loop control vector, so that the controller can generate the control parameters for the next cycle based on the central shaft pressure concentration error, outlet release holding error and residual pressure error, thereby avoiding the overpressure in the cavity, premature pressure release or pressure tailing caused by simply increasing the driving power, and improving the stability and safety of the infrasound generation process.

[0021] (4) In this invention, the stability of the pre-wave pressure window adaptation value, cavity pressure loss state, central axis pressure gain, outlet release holding value and periodic residual pressure ratio is judged by multiple consecutive target infrasound frequency cycles. After the stability conditions are met, a locking mark and parameter fingerprint are generated, so that the driving parameters, cavity boundary parameters and phase release parameters under the same target infrasound frequency can be reused, thereby improving the consistency of repeated start-up of the infrasound generating equipment and the reliability of frequency probing and retreat.

[0022] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0023] Figure 1 This is a flowchart of a resonant cavity pressure feedback control method for infrasound generation according to the present invention; Figure 2 This is the dynamic response curve of the pressure amplitude of the present invention; Figure 3 The performance index comparison curves of the present invention under different target infrasound frequency periods are shown. Figure 4 This is a heatmap showing the correlation between fault modes and state characteristic intensity in this invention. Figure 5 This is the curve showing the change in the control correction amount in this invention; Figure 6 This is a schematic diagram of the resonant cavity structure of the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Please see Figures 1-6 This invention provides a technical solution: a resonant cavity pressure feedback control method for infrasound generation, comprising: S1, acquiring pre-formed wave cavity pressure observation data, performing time axis repositioning and period scale unification, and generating a pre-formed wave pressure chain dataset; S2, based on the pre-formed wave pressure chain dataset, calculating the peak phase difference, evaluating the pre-formed wave pressure window adaptation, and constructing a pre-formed wave pressure window dataset; S3, based on the pre-formed wave pressure window dataset, evaluating the cavity pressure loss sequence deviation, performing cavity pressure loss state feature matching, and generating cavity pressure loss state data and boundary constraint data; S4, based on the boundary constraint data, constructing a control direction matrix, writing the closed-loop control vector into the controller, and generating a locking identifier and parameter fingerprint.

[0026] Specifically, the steps for acquiring pre-formed wave cavity pressure observation data, performing time axis repositioning and period scale unification, and generating a pre-formed wave cavity pressure observation dataset are as follows: The inlet pressure, central axis pressure, outlet pressure, residual pressure, and external release pressure during the infrasound generation process are collected; the cavity pressure loss type, target infrasound frequency, and corrected sound velocity are set; the resonant cavity inlet position, central axis pressure acquisition point, outlet pressure acquisition point, phase coordinate equivalent propagation distance, and calibrated propagation velocity are calibrated through sensor layout; and the peak values ​​of central axis pressure, inlet pressure, outlet pressure, and the corresponding times are calculated; all collected, set, calibrated, and calculated data are summarized to form pre-formed wave cavity pressure observation data; and time axis repositioning is performed based on the driving cycle synchronization mark. The method involves correcting the sampling time of pressure data, mapping pressure data within the same target infrasound frequency period to a unified periodic coordinate system, i.e., phase coordinates from 0 to 1. A pressure change rate constrained filtering method is used to dynamically denoise the spike noise generated during the start-up of the drive source, the airflow impact at the guide port, and the residual pressure fall-off stage within the cavity. The pre-wave cavity pressure observation data is normalized using a peak amplitude scaling and phase coordinate mapping method based on the target infrasound frequency period segmentation, normalizing the pressure amplitude within each period to the [-1,1] interval, unifying the periodic scale of the pressure data, and generating a pre-wave pressure chain dataset. The inlet pressure acquisition position is set at the resonant cavity inlet connection section, the central axis pressure acquisition position is set in the central axis region of the annular pressure concentration cavity, the outlet pressure acquisition position is set inside the guide hole, and the external release pressure acquisition position is set outside the guide hole, ensuring that the acquired pressure data can reflect the continuous process of pressure excitation entry, cavity concentration, guide release, and external output.

[0027] like Figure 2 The pressure amplitude dynamic response curve shown has the normalized target infrasound frequency period on the horizontal axis, representing the time progression within one working cycle, and the pressure amplitude on the vertical axis. It displays the pressure changes at four locations: inlet (P0), central axis (P1), outlet (P2), and external release end (P3). The three key working stages—inlet compression, central axis concentration, and outlet release—are marked with light blue windows. The curve shows that the pressure excitation responds first at the inlet, forming a compression wave. This wave then travels to the central axis, converging to form the highest pressure peak of the entire cycle. It then travels to the outlet, maintaining a secondary peak, and finally dissipates at the external release end to the lowest and delayed peak. This clearly presents the complete transmission process of the pressure wave within the system, from inlet compression to central axis concentration and outlet release. This data can be used to analyze the dynamic transmission timing and amplitude attenuation characteristics of the pressure wave, providing data support for pressure control and optimization design.

[0028] In this implementation scheme, by simultaneously acquiring, timing-correcting, dynamically denoising, and normalizing the inlet pressure, central axis pressure, outlet pressure, residual pressure, and external release pressure, cavity pressure data from different acquisition locations, pressure ranges, and sampling times can be unified within the same target infrasound frequency period for comparison. This improves the observability of the intracavitary pressure transmission process, the accuracy of pressure peak timing analysis, and the reliability of subsequent pre-formed wave pressure window adaptation calculations. It also provides a stable data foundation for identifying states such as insufficient inlet coupling, insufficient central axis concentration, premature outlet pressure release, intracavitary stagnation, and residual pressure stacking.

[0029] Specifically, based on the pre-formed wave pressure chain dataset, the specific steps for calculating the peak phase difference are as follows: Based on the pre-formed wave pressure chain dataset, the structural distance from the resonant cavity inlet to the central axis pressure acquisition point is obtained through sensor installation location calibration, which is the straight-line length along the cavity centerline; the allowable deviation of the structural distance is obtained by synthesizing the assembly tolerance and the frequency drift equivalent distance using the root mean square method, and the frequency drift equivalent distance is calculated based on the wavelength change corresponding to a fluctuation of 0.5 above or below the target infrasound frequency; the central axis pressure gain is obtained by the ratio of the central axis pressure peak value to the inlet pressure peak value; the phase coordinate of the current target infrasound frequency period is converted from the time difference between the inlet pressure peak value and the central axis pressure peak value, and divided by the current target infrasound frequency period duration, and converted to the phase coordinate within the range of [0,1), the actual phase difference from the inlet to the central axis is obtained, and the phase difference is normalized by the period length; the current target infrasound is obtained by converting the time difference between the central axis pressure peak value and the pressure peak value before the outlet. The phase coordinates of the frequency cycle are divided by the duration of the current target infrasound frequency cycle to convert them to phase coordinates within the range of [0,1), thus obtaining the actual phase difference from the central axis to the outlet. The sign of the time difference represents the order of peak values. The expected phase difference from the inlet to the central axis is determined by the target infrasound frequency, the equivalent propagation distance from the inlet to the central axis, and the calibrated propagation speed. The expected phase difference is equal to the equivalent propagation distance divided by the calibrated propagation speed and then multiplied by the target infrasound frequency. The expected phase difference from the central axis to the outlet is determined by the target infrasound frequency, the equivalent propagation distance from the central axis to the outlet, and the calibrated propagation speed. The equivalent propagation distance is calculated offline by the cavity geometric parameters. The residual pressure at the end of the current cycle is obtained by comparing the residual pressure at the central axis and the pressure before the outlet with the peak pressure of the current cycle. The residual pressure is the average pressure within 1 / 10 of the cycle duration at the end of the cycle.

[0030] In this implementation scheme, by uniformly calculating the structural distance, allowable deviation of structural distance, central axis pressure gain, actual phase difference from inlet to central axis, actual phase difference from central axis to outlet, expected phase difference, and residual pressure ratio at the end of the cycle, the transmission location, transmission time, and transmission sequence of pressure peak in the cavity can be quantitatively expressed. It can simultaneously reflect the influence of cavity installation error, frequency drift, and residual pressure at the end of the cycle on the pre-formed wave pressure window, providing a stable basis for calculating the pre-formed wave pressure window adaptation value, identifying premature outlet depressurization, and stagnation state in the cavity.

[0031] Specifically, the steps for evaluating the adaptation of the pre-formed wave pressure window are as follows: The difference between the structural distance and the target infrasound frequency and four times the corrected sound velocity is divided by the sum of the allowable deviation of the structural distance and the zero-prevention term, squared, and then the negative exponent is used to obtain the wavelength position fitting term based on the Gaussian kernel function principle; the central axis pressure gain is increased by one, the natural logarithm is taken, and then divided by the result of adding the natural logarithm to obtain the central axis pressure enhancement term; the difference between the actual phase difference from the inlet to the central axis and the expected phase difference from the inlet to the central axis is divided by two, the sine value is taken, and the absolute value is taken. The actual phase difference and the expected phase difference are both dimensionless phase coordinates in the range of [0,1) to obtain the phase deviation from the inlet to the central axis. The phase deviation term from the center axis to the outlet is obtained by subtracting the actual phase difference from the center axis to the outlet from the expected phase difference and dividing by two, then taking the sine value and the absolute value. The two phase deviation terms are added together and the negative exponent is taken to obtain the phase transmission coordination term based on the principle of exponential decay. The residual pressure ratio at the end of the cycle is subtracted from the upper limit of the allowable residual pressure, multiplied by the residual pressure gating coefficient, and then exponentially calculated. After adding a constant one and taking the reciprocal, the pressure suppression term is obtained. The wavelength position matching term, the center axis pressure enhancement term, the phase transmission coordination term, and the pressure suppression term are multiplied in sequence to obtain the pre-formed wave pressure window adaptation value.

[0032] The specific formula for calculating the pre-formed wave pressure window adaptation value is as follows: ; In the formula, The pre-formed wave pressure window adaptation value represents the frequency period of the k-th target infrasound wave, which is used to determine whether the gas compression wave forms an effective pressure concentration before a quarter wavelength in the current period. This indicates the target infrasound frequency, used to calculate the reference distance corresponding to the target infrasound frequency; This represents the corrected velocity of sound within the kth target infrasound frequency period, used to eliminate the influence of cavity temperature changes; This represents the structural distance from the resonant cavity inlet to the central pressure acquisition point, used to characterize the actual pressure concentration location of the central pressure acquisition point within the cavity; This indicates the allowable deviation of structural distance, used to scale the deviation between the central pressure acquisition positions; This indicates the zero-prevention term, which is obtained by preset a small positive number. Its value range is greater than 0 and less than the minimum effective pressure change involved in the calculation. It is used to avoid the denominator being zero or the logarithmic calculation being abnormal. This represents the central pressure gain within the kth target infrasound frequency period, used to determine whether the inlet compression pressure is enhanced in the central region. It represents the actual phase difference between the inlet pressure peak and the central axis pressure peak, and is used to determine the phase relationship of the inlet compression being concentrated and transmitted to the central axis; It represents the actual phase difference between the peak pressure at the center shaft and the peak pressure at the outlet, and is used to determine the phase relationship of the concentrated release from the center shaft to the outlet. It represents the expected phase difference between the peak inlet pressure and the peak central axis pressure at the current target infrasound frequency, and is used as a phase reference for pressure transmission from the inlet to the central axis. This represents the expected phase difference between the peak pressure at the center axis and the peak pressure at the outlet at the current target infrasound frequency, which is used as a phase reference for pressure transmission from the center axis to the outlet. It represents the percentage of residual pressure at the end of the k-th target infrasound frequency period, used to determine whether there is residual pressure accumulation or residual echo standing waves in the cavity; This indicates the upper limit of the allowable residual pressure, which is preset through equipment stable output testing or cavity safety control requirements, and is used to limit the acceptable residual pressure level at the end of the cycle. This represents the residual pressure gating coefficient, which is obtained by setting the pressure feedback control sensitivity. Its value range is greater than 0. It is used to control the suppression intensity of the pre-formed wave pressure window adaptation value when the residual pressure at the end of the cycle exceeds the allowable upper limit.

[0033] Table 1 shows the adaptation parameters for the periodic wave pressure structure. The following is a quantitative explanation of each period number: Period number K01 has a structural distance deviation of 0.46, a central axis pressure gain of 1.15, a phase deviation of 0.38, a residual pressure ratio of 0.18, and a pre-wave pressure window adaptation value of 0.48; Period number K02 has a structural distance deviation of 0.32, a central axis pressure gain of 1.21, a phase deviation of 0.31, a residual pressure ratio of 0.16, and a pre-wave pressure window adaptation value of 0.57; Period number K03 has a structural distance deviation of 0.18, a central axis pressure gain of 1.31, a phase deviation of 0.24, a residual pressure ratio of 0.14, and a pre-wave pressure window adaptation value of 0.69; Period number K04 has a structural distance deviation of... 0.08, central axis pressure gain is 1.39, phase deviation is 0.18, residual pressure ratio is 0.12, and pre-wave pressure window adaptation value is 0.8; period number K05 has a structural distance deviation of 0.04, central axis pressure gain is 1.43, phase deviation is 0.14, residual pressure ratio is 0.11, and pre-wave pressure window adaptation value is 0.87; period number K06 has a structural distance deviation of 0.16, central axis pressure gain is 1.46, phase deviation is 0.22, residual pressure ratio is 0.13, and pre-wave pressure window adaptation value is 0.76; period number K07 has a structural distance deviation of 0.39, central axis pressure gain is 1.34, phase deviation is 0.36, residual pressure ratio is 0.21, and pre-wave pressure window adaptation value is 0.52. From the overall change pattern, the structural distance deviation, phase deviation and residual pressure ratio all show a trend of first decreasing and then increasing. The central axis pressure gain shows a characteristic of first rising and then falling. The pre-formed wave pressure window adaptation value first continuously increases to reach a peak at K05 and then begins to fall. Among them, K05 has the smallest deviation index and the best adaptation value, which is the best working cycle for wave pressure structure matching. The adaptation performance of the other cycles decreases in turn, which can provide quantitative support for structural wavelength matching and pressure window parameter optimization.

[0034] Table 1. Data on Adaptation Parameters for Periodic Wave Pressure Structure like Figure 3The performance index comparison curves shown are for different target infrasound frequency periods. The horizontal axis represents the target infrasound frequency period sequence, and the vertical axis represents the normalized performance index values, including four key indicators: preformed wave pressure window adaptation value, normalized central axis pressure gain, phase coordination degree, and residual pressure suppression degree. The figure clearly shows the variation of each index with the frequency period: the normalized central axis pressure gain reaches a peak of 1.0 in period K06 and remains at a high level overall, indicating that the control method performs well in terms of central axis pressure amplification efficiency; the preformed wave pressure window adaptation value, phase coordination degree, and residual pressure suppression degree all show a trend of first increasing and then decreasing, reaching the optimal state near period K05, and then gradually decaying. This indicates that at a specific frequency period, the pressure wave window adaptation, phase coordination, and residual pressure suppression effects achieve the best match, while the control performance decreases after deviating from this period. The curve visually verifies the frequency adaptation characteristics of the infrasound generation control method, indicating that the comprehensive optimization of pressure amplification, phase coordination and residual pressure suppression can be achieved under a specific target infrasound frequency period, providing a quantitative basis for the optimization of pressure feedback control of the infrasound resonant cavity.

[0035] In this implementation scheme, by incorporating structural distance deviation, central axis pressure gain, pressure peak phase deviation, and residual pressure ratio at the end of the cycle into the pre-formed wave pressure window adaptation evaluation, the pressure concentration location, pressure amplification effect, phase transmission sequence, and residual pressure suppression state within the current target infrasound frequency cycle can be uniformly quantified.

[0036] Specifically, the steps for constructing the pre-wave pressure window dataset are as follows: When the preformed wave pressure window fit value is less than the preformed wave fit threshold, it is determined that the current period has not formed a valid preformed wave pressure window, and the current period is marked as an invalid pressure window period, marked as 0 in the binary label; when the preformed wave pressure window fit value is greater than or equal to the preformed wave fit threshold, it is determined that the current period has formed a valid preformed wave pressure window, and the current period is marked as a valid pressure window period, marked as 1 in the binary label; period-level pressure window labels are formed, each label corresponds to a target infrasound frequency period, and are sorted and summarized in ascending order according to the target infrasound frequency. Each period is accompanied by a label value and a corresponding timestamp, and a preformed wave pressure window dataset is constructed. The dataset format is a key-value pair sequence of period number and label value.

[0037] In this implementation scheme, by converting the pre-formed wave pressure window adaptation value into a binary period label, it is possible to directly determine and record whether a valid pre-formed wave pressure window is formed for each target infrasound frequency period. By arranging the period number, label value, and timestamp according to the target infrasound frequency sequence, a continuous and traceable pressure window state sequence can be formed, which facilitates the rapid screening of valid pressure window periods and eliminates the interference of invalid periods on the identification of cavity pressure loss. This improves the clarity of the pressure window determination results, the stability of continuous period analysis, and the reliability of closed-loop control parameter selection.

[0038] Specifically, the steps for evaluating the cavity pressure loss sequence deviation based on the pre-formed wave pressure window dataset are as follows: Based on the pre-formed wave pressure window dataset, the controller divides the sampling period according to the target infrasound frequency and sequentially numbers it to obtain the target infrasound frequency period number, with the numbering starting from 1 and incrementing. Using a cavity pressure loss type enumeration mapping method, the enumerated types include insufficient inlet coupling, insufficient structural constraints, premature outlet pressure relief, intracavitary hysteresis, and abnormal echo standing wave. These abnormal states are assigned state numbers to obtain candidate cavity pressure loss state numbers. The candidate cavity pressure loss state numbers, along with their corresponding peak amplitude ratio, phase difference, phase interval within the target infrasound frequency period, and upper and lower limits of actuator adjustment, are written into the same state mapping table. The mapping table uses the state number as an index, and each row records all feature parameters of the state to obtain a set of candidate cavity pressure loss states. The peak times of the inlet pressure, central axis pressure, and outlet pressure in the current cycle are read and sorted in chronological order to obtain the actual occurrence order. The typical occurrence order of the peak inlet pressure, central axis pressure, and outlet pressure under each type of cavity pressure loss state is calibrated to obtain the standard peak order. The number of adjacent position exchanges required to adjust the actual peak order to the standard peak order is calculated to obtain the order deviation value. The number of adjacent position exchanges is the number of exchanges in bubble sort, which is used to measure the degree of difference between the two sequences.

[0039] In this implementation scheme, by enumerating and numbering the chamber pressure loss states and establishing a state mapping table that includes peak amplitude ratio, phase difference, phase interval and actuator adjustment boundary, different chamber pressure loss states have a unified comparison benchmark. This can transform the differences in the order of pressure peaks into quantifiable deviation indicators, improve the objectivity and reproducibility of chamber pressure loss state identification, and provide an accurate sequence determination basis for calculating the chamber pressure loss state matching sequence number, determining the control boundary and generating closed-loop adjustment commands.

[0040] Specifically, the steps for performing cavity pressure loss state feature matching are as follows: The inlet pressure, central axis pressure, outlet pressure, external release pressure, and residual pressure of the current cycle are linearly normalized to obtain state characteristic quantities. The normalization benchmark is the difference between the current value and the minimum value within the cycle, divided by the difference between the maximum value and the minimum value within the cycle, so that each state characteristic quantity is mapped to the interval [0,1] and is dimensionless. Standard characteristic values ​​are obtained by statistically analyzing sample cycles under cavity pressure loss conditions. The standard characteristic value is the mean of the state characteristic quantities of all sample cycles under the state. The fluctuation amplitude of each state characteristic quantity under the same cavity pressure loss condition is extracted, and the quantile deviation of the corresponding characteristic quantity is used as the characteristic allowable deviation under the state. The quantile deviation is taken as the interquartile range of the sample state characteristic quantities. The pressure windows that trigger an anomaly are obtained by measuring the windows in the inlet compression window, central axis concentration window, outlet release window, and residual pressure dissipation window that exceed the anomaly judgment conditions. The inlet compression window corresponds to the phase coordinate interval [0, 0.25) within the target infrasound frequency period, the central axis concentration window corresponds to the interval [0.25, 0.5), the outlet release window corresponds to the interval [0.5, 0.75), and the residual pressure dissipation window corresponds to the interval [0.75, 1). By using the state-window correspondence calibration method, the pressure window range within which an anomaly is allowed to occur first for each type of cavity pressure loss state is recorded to obtain the trigger window set. Each element in the trigger window set is the window phase interval within which an anomaly is allowed to occur first.

[0041] Subtract the standard feature value corresponding to the candidate cavity pressure loss state from each state feature quantity, take the absolute value, and divide it by the sum of the feature allowable deviation and the zero-prevention term to obtain the deviation degree, which is dimensionless. Select the maximum value from the deviation degrees of all state feature quantities as the maximum deviation term, which reflects the feature dimension with the largest deviation between the current cycle and the candidate state. Determine whether the pressure window that triggers the anomaly in the current cycle belongs to the trigger window set. If it does, the window penalty term is zero; if it does not, the trigger window mismatch penalty coefficient is used as the window penalty term. Add the sequence deviation value, the maximum deviation term, and the window penalty term to obtain the total state deviation value, which is dimensionless. Select the candidate state with the smallest total state deviation value to obtain the cavity pressure loss state matching sequence number, which corresponds to the state number in the candidate cavity pressure loss state set.

[0042] The specific formula for calculating the cavity pressure loss state matching sequence number is as follows: ; In the formula, The cavity pressure loss state matching number represents the k-th target infrasound frequency period, used to indicate the closest cavity pressure loss state in the current period; Indicates the target infrasound frequency period number, used to distinguish different control periods; This indicates the candidate cavity pressure loss state number, which is used to participate in the matching calculation one by one; This represents the set of candidate cavity pressure loss states, obtained through a pre-established cavity pressure loss state type table, used to limit the state range for matching calculations. This indicates the actual order of occurrence of the inlet pressure peak, the central axis pressure peak, and the outlet pressure peak within the k-th cycle, which is used to determine the pressure transmission sequence in the current cycle. The standard peak order corresponding to the s-th type of cavity pressure loss state is used as a benchmark for comparison with the actual peak order. This represents the sequence deviation value, used to characterize the degree to which the current pressure peak sequence closely approximates the pressure loss state of a certain chamber; The state characteristic quantity number is obtained by sequentially numbering the inlet compression response, central axis concentrated response, outlet release response, periodic residual pressure ratio, and pressure peak tailing time, and is used to calculate the state characteristic deviation item by item. This represents the j-th state feature quantity of the k-th cycle, used to characterize the actual cavity pressure state in the corresponding feature dimension of the current cycle; This represents the j-th standard eigenvalue corresponding to the s-th type of cavity pressure loss state, which is used as a standard reference for the current state characteristic quantity; This represents the allowable deviation of the j-th feature corresponding to the s-th type of cavity pressure loss state, which is used to scale the differences in state features. This represents the zero-prevention term, which is obtained by setting a small positive number. Its value range is greater than 0 and less than the minimum effective characteristic deviation, and it is used to avoid the denominator being zero. The pressure window that triggers the anomaly in the k-th cycle is used to characterize the earliest location where the current anomaly occurs. This represents the set of trigger windows that are allowed to occur in the s-th type of cavity pressure loss state, used to determine whether the current first abnormal window conforms to the occurrence pattern of this type of state; This represents the penalty coefficient for triggering window mismatch, which is obtained by setting the sensitivity of cavity pressure state recognition. The value range is greater than 0. It is used to increase the matching cost of the state when the current first abnormal window does not belong to the set of allowed windows for the corresponding state. This indicates an indicator function, obtained by checking whether the condition within parentheses is true or false. Its value range is 0 or 1, and it is used to... Not belonging to When introducing a penalty item, belong No penalty items are introduced at times.

[0043] like Figure 4The heatmap showing the correlation between fault modes and state characteristic intensities is divided into five columns: insufficient inlet coupling, insufficient structural constraint, premature outlet depressurization, intracavity stagnation, and abnormal echo standing wave. The columns correspond to five key state characteristics: inlet compression response, insufficient central axis concentration, premature depressurization energy, residual pressure ratio, and pressure tailing. The values ​​in each cell represent the normalized state characteristic intensity, with brighter yellow indicating higher intensity. The graph clearly demonstrates the strong coupling relationship between faults and characteristics: the characteristic intensities of abnormal echo standing wave with residual pressure ratio and pressure tailing reach 0.88 and 0.92 respectively; the characteristic intensities of intracavity stagnation with inlet compression response and insufficient central axis concentration are 0.82 and 0.80 respectively; and the characteristic intensity of premature outlet depressurization with premature depressurization energy is 0.82. These high-value combinations constitute the typical characteristic fingerprints of each fault mode. Insufficient inlet coupling, on the other hand, has a relatively low overall characteristic intensity, distinguishing it from other fault modes. The heatmap intuitively quantifies the correspondence between faults and state characteristics, providing a characteristic basis for fault diagnosis and state assessment of infrasound generation systems.

[0044] This implementation improves the ability to distinguish between insufficient inlet coupling, insufficient structural constraints, premature outlet depressurization, intracavitary stagnation, and abnormal echo standing waves. By selecting the candidate state with the smallest total deviation to generate the cavity pressure loss state matching number, it avoids misjudgment caused by relying solely on a single pressure peak or a single phase index, thus improving the accuracy, stability, and reproducibility of cavity pressure loss state identification and providing a reliable basis for control boundary generation and closed-loop regulation direction determination.

[0045] Specifically, the steps for generating cavity pressure loss state data and boundary constraint data are as follows: The cavity pressure loss type is read according to the cavity pressure loss state matching sequence number: When the cavity pressure loss state matching sequence number is equal to the first state sequence number, the inlet coupling is insufficient, indicating that the drive source output has failed to be converted into inlet pressure. The upper limit of the drive amplitude is limited to 0.8 times the current value, preventing further increase in the drive amplitude, and the verification of inlet connection, sealing state, and drive source coupling state is initiated; when the cavity pressure loss state matching sequence number is equal to the second state sequence number, the structural constraint is insufficient, indicating that the inlet pressure has been established but has not been concentrated in the central axis region, generating a cavity. The adjustable range of the body angle actuator is set with a lower limit of 2 degrees below the current angle value and an upper limit of 2 degrees above the current angle value. The target infrasonic frequency is limited from further downward movement, with a lower limit set at 0.95 times the current target infrasonic frequency. When the cavity pressure loss state matching number equals the third state number, an early outlet pressure relief state is determined, indicating that the pressure before the outlet is released earlier than the central axis pressure concentration process. This generates a reduction range for the equivalent opening of the guide port, setting the upper limit of the opening to 0.7 times the current opening value and the lower limit to 0.3 times the current opening value, allowing the outlet to release pressure. The time is shifted backward; when the cavity pressure loss state matching number equals the fourth state number, the cavity stagnation state is determined, indicating that the central axis pressure has been formed but has failed to be transmitted to the outside of the phase guide port. An increase in the equivalent opening range of the phase guide port is generated, setting the lower limit of the opening to 1.3 times the current opening value and the upper limit to 1.6 times the current opening value, thus opening the cavity pressure release path; when the cavity pressure loss state matching number equals the fifth state number, the echo standing wave abnormal state is determined, generating an adjustment range for the driving amplitude of 0.6 to 0.9 times the current amplitude, generating a frequency backoff range. The range is shifted downwards by 1 to 3 from the current frequency, and the residual pressure dissipation waiting time is 5 to 10 target infrasonic frequency cycles. If the cavity pressure loss state matching number does not belong to the above five state numbers, the state identification of the current cycle is determined to be invalid, and the generation of closed-loop regulation for the current cycle is stopped, that is, no upper or lower limits of any control quantity are output. Attribution analysis is performed based on the current cavity pressure loss type to determine the main links and corresponding control objects of pressure loss occurrence. The adjustable range and limiting conditions generated above are summarized to generate cavity pressure loss state data, and boundary constraint data is generated simultaneously.

[0046] In this implementation scheme, by converting the cavity pressure loss state matching sequence number into the corresponding loss type, and generating control boundaries for different loss types, such as drive amplitude, target frequency, cavity angle, equivalent opening of the guide port, and residual pressure waiting period, the closed-loop regulation no longer adopts a uniform parameter tuning method. Instead, it implements differentiated constraints based on the different causes of insufficient inlet coupling, insufficient structural constraints, premature outlet pressure relief, cavity stagnation, and abnormal echo standing waves. This avoids direct actuator action due to erroneous state judgments, reduces the risk of misadjustment, and provides clear and reliable state data and boundary constraint basis for closed-loop control vector calculation.

[0047] Specifically, the steps for constructing the control direction matrix based on boundary constraint data are as follows: The adjustment step size is determined based on the cavity pressure loss state type and the number of abnormal cycles in the current preformed wave pressure window adaptation value. The adjustment step size is a scalar, and the more abnormal cycles in the adaptation value, the larger the step size. Using a state-to-actuator mapping method, based on five types of cavity pressure loss states—insufficient inlet coupling, insufficient structural constraints, premature outlet pressure relief, cavity hysteresis, and abnormal echo standing waves—the adjustment directions and weights of four actuators—drive frequency, drive amplitude, cavity angle state, and equivalent opening of the phase guide port—are determined respectively. A control direction matrix is ​​constructed, with each row of this matrix corresponding to the central axis pressure. The three errors—lump error, outlet release holding error, and residual pressure error—correspond to four actuated quantities: drive frequency, drive amplitude, cavity angle state, and equivalent opening of the phase guide port. Matrix elements take values ​​of +1, 0, or -1, indicating the direction of adjustment of the actuated quantity by the error. The lower and upper control limits are obtained through the superposition of equipment safety boundaries and state constraints. Both the lower and upper control limits are four-dimensional vectors, corresponding to the minimum and maximum values ​​of the four actuated quantities, respectively. At the target infrasound frequency, a condition is selected where there is no pressure relief, no echo standing wave anomaly, and the external release pressure remains constant. For the target infrasound frequency period, calculate the gain statistics of the peak value of the central axis pressure and the peak value of the inlet pressure, and take the median within the period to obtain the target central axis pressure gain, which is dimensionless. At the target infrasound frequency, select a period where the peak value of the outlet release and the peak value of the central axis pressure maintain a phase relationship, calculate the difference between the peak value of the external release pressure and the peak value of the pressure before the outlet, and combine this with the time difference penalty from the peak value of the pressure before the outlet to the peak value of the external release pressure to obtain the target outlet release holding value, which is dimensionless. Compare the amplitude of the peak value of the external release pressure and the peak value of the pressure before the outlet to obtain the actual outlet release holding value, which is dimensionless. Read the residual central axis pressure and the residual pressure before the outlet at the end of the current period, and compare them with the total peak value of the peak central axis pressure and the peak pressure before the outlet in the current period to obtain the period residual pressure ratio, which is dimensionless. At the target infrasound frequency, select a period where a pre-formed wave pressure window has been formed and no echo standing wave anomaly has appeared, and perform proportional calculations to obtain the target residual pressure ratio, which is dimensionless.

[0048] The central axis pressure concentration error is obtained by adding a zero-prevention term to the target central axis pressure gain, dividing by the sum of the central axis pressure gain and the zero-prevention term, and then taking the natural logarithm. The outlet release holding error is obtained by adding a zero-prevention term to the target outlet release holding value, dividing by the sum of the outlet release holding value and the zero-prevention term, and then taking the natural logarithm. The residual pressure error is obtained by subtracting the target residual pressure percentage from the residual pressure percentage. The central axis pressure concentration error, outlet release holding error, and residual pressure error are then combined to form an error vector, a three-dimensional column vector. A hyperbolic tangent operation is then performed on the error vector to obtain the error adjustment. The vector has elements ranging from -1 to 1. The error adjustment vector is multiplied by the control direction matrix, which has dimensions of 4 rows and 3 columns, i.e., the number of execution quantities × the number of errors, to obtain the correction direction. The correction direction is multiplied by the adjustment step size to obtain the control correction amount. The control correction amount is subtracted from the control vector of the current cycle to obtain the control vector of the next cycle. The control vector of the next cycle is restricted between the lower control limit and the upper control limit to obtain the closed-loop control vector, a four-dimensional vector, with each component corresponding to the set values ​​of drive frequency, drive amplitude, cavity angle state, and equivalent opening of the guide port.

[0049] The specific formula for calculating the closed-loop control vector is as follows: ; In the formula, Indicates the first The closed-loop control vector for each cycle is used to determine the control parameters of the drive source, cavity boundary and phase guide port for the next cycle; Indicates the first The control vector for each cycle is used as the basic parameter for closed-loop correction. This indicates the matching sequence number of the cavity pressure loss state in the k-th cycle, used to determine the adjustment direction adopted in the current cycle; Indicates the first The adjustment step size corresponding to the cavity pressure loss state is used to control the closed-loop correction amplitude; Indicates the first The control direction matrix corresponding to the cavity pressure loss state is used to convert the central axis pressure concentration error, outlet release holding error and residual pressure error into the correction direction of each control quantity; This represents the limiting function, which is the lower control limit corresponding to the current cavity pressure loss state. and control limit The calculation results are truncated to prevent the control parameters from exceeding the safe operating range in the next cycle. This represents the target axial pressure gain, which is used as a target value for the degree of axial pressure concentration. This represents the central axis pressure gain in the k-th period, used to characterize the degree of enhancement of the inlet pressure in the central axis region after it enters the resonant cavity; This represents the target export release maintenance value, which is used as the target value for export release capacity. This represents the actual outlet release holding value in the k-th cycle, used to characterize the degree of holding of the pressure before the outlet to the outside of the guide port; This represents the percentage of residual pressure in the k-th cycle, used to characterize the degree of residual pressure stacking within the cavity; This indicates the target residual pressure percentage, which is used as a residual pressure control target. This indicates a zero-prevention term, obtained by setting a small positive number. Its value range is greater than 0 and less than 0.001, and it is used to avoid abnormal logarithmic and division operations.

[0050] like Figure 5 The control correction curves shown depict the target infrasound frequency cycle sequence on the horizontal axis and the correction values ​​of four control parameters—drive frequency correction, drive amplitude correction, cavity angle correction, and phase guide opening correction—on the vertical axis. This visually reflects the adjustment patterns of each control parameter with the frequency cycle: the phase guide opening correction is -6 at K01 and reaches a peak of approximately 4.8 at K06, exhibiting the largest fluctuation and serving as the core adjustment parameter for the infrasound generation system at different frequencies; the cavity angle correction reaches a peak of approximately 2.5 at K02, then gradually declines, dropping to -2 at K07, primarily used for early phase coordination optimization; the drive amplitude correction generally shows a trend of first being negative, then positive, and then significantly decreasing, dropping to approximately -5 at K07, adapting to the pressure energy requirements under different cycles; the drive frequency correction generally exhibits smaller fluctuations, showing a slow upward trend later, used for fine calibration of the system's frequency matching. Overall, the curves clearly demonstrate the dynamic adjustment strategies and primary / secondary relationships of each control parameter in the infrasound generation system under different target infrasound frequency cycles.

[0051] In this implementation scheme, by using the cavity pressure loss state, the pre-formed wave pressure window adaptation, and the number of abnormal cycles to determine the adjustment step size, control direction matrix, and control upper and lower limits, the closed-loop control no longer relies on fixed adjustment rules. Instead, it can dynamically adjust the driving frequency, driving amplitude, cavity angle state, and equivalent opening of the phase guide port according to the current cause of cavity pressure loss. By converting the central axis pressure concentration error, outlet release holding error, and residual pressure error into error adjustment vectors and mapping them to the correction directions of four execution quantities through the control direction matrix, targeted compensation for states such as insufficient pressure concentration, abnormal outlet release, and residual pressure stacking can be achieved. By limiting the control vector of the next cycle within the boundary constraint range through the amplitude limiting function, it is possible to prevent control parameters from exceeding the limits, leading to overpressure, premature pressure release, or abnormal echo standing waves, thereby improving the stability, safety, and adaptive adjustment capability of the infrasound generation process under continuous frequency cycles.

[0052] Specifically, the steps for writing the closed-loop control vector into the controller and generating the locking identifier and parameter fingerprint are as follows: The next cycle closed-loop control vector is written into the controller, and the output waveform, output intensity, cavity guide boundary position, and phase port release section of the drive source are adjusted so that the inlet pressure peak, central axis pressure peak, and outlet release peak re-enter the pre-formed wave pressure transmission sequence within the next target infrasonic frequency cycle. When m target infrasonic frequency cycles, where m is the number of consecutive effective cycles (5 to 10), a lockout flag is generated based on the preset system stability speed and simultaneously satisfying the following conditions: the pre-formed wave pressure window adaptation value is greater than the pre-formed wave adaptation threshold, the cavity pressure loss state does not show the premature outlet pressure relief state, and the echo standing wave abnormal state. After generating the lockout flag, the parameter records are written in the order of the target infrasonic frequency cycles. The closed-loop control vector and target infrasonic frequency corresponding to each cycle are recorded sequentially. The parameter sequence formed by the records in the cycle order is the parameter fingerprint, which is used for working state reproduction or rapid parameter loading under the same working conditions in the future.

[0053] This implementation scheme can avoid accidental locking caused by conditions being met accidentally in a single cycle, thus improving the reliability of locking identifier generation. By recording the closed-loop control vector and the target infrasound frequency in cyclic order after locking, a reusable parameter fingerprint is formed, enabling the equipment to quickly load stable parameters under the same operating conditions, reducing repeated frequency sweeps and repeated adjustment processes, and improving the reproducibility of the infrasound generation state, startup efficiency, and control continuity of the frequency decline process.

[0054] like Figure 6 As shown, a resonant cavity pressure feedback control method for infrasound generation in this embodiment can be applied to an annular resonant cavity with a guide hole. The resonant cavity has an overall annular cross-sectional structure. A continuous zigzag-shaped flow guide boundary is provided in the upper part of the cavity, and an inlet connection section and an outlet release section are formed in the lower part of the cavity. The inlet connection section is used to receive the periodic pressure excitation generated by the driving source. The continuous zigzag-shaped flow guide boundary is used to change the propagation path of the gas compression wave in the cavity, so that the low-frequency pressure wave entering the cavity is reflected, superimposed and concentrated in the cavity. The outlet release section is used to release the concentrated low-frequency pressure wave in the cavity outward.

[0055] During testing, periodic low-frequency excitation was input to the inlet connecting section of the resonant cavity, and pressure responses were collected at the inlet, central axis region, outlet, and external release positions. Test results showed that the inlet pressure initially formed a peak, followed by a higher pressure peak in the central axis region. The outlet pressure peak lagged behind the central axis pressure peak. Pressure fluctuations corresponding to the target low-frequency period were detected at the external release end, indicating that the pressure wave propagated within the resonant cavity in the order of inlet compression, central axis concentration, and outlet release. Therefore, the annular resonant cavity can concentrate low-frequency compression and sparsity waves within the cavity and release them outwards through the outlet release section, providing a structural basis for infrasound generation.

[0056] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0057] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for pressure feedback control of a resonant cavity for infrasound generation, characterized in that: Includes the following steps: S1. Acquire pre-wave cavity pressure observation data, and perform time axis repositioning and period scale unification to generate pre-wave pressure chain dataset; S2, based on the pre-formed wave pressure chain dataset, calculate the peak phase difference, evaluate the pre-formed wave pressure window adaptation, and construct the pre-formed wave pressure window dataset; S3, based on the pre-formed wave pressure window dataset, evaluates the cavity pressure loss sequence deviation, performs cavity pressure loss state feature matching, and generates cavity pressure loss state data and boundary constraint data; S4, based on boundary constraint data, constructs a control direction matrix, writes the closed-loop control vector into the controller, and generates a locking identifier and parameter fingerprint.

2. The resonant cavity pressure feedback control method for infrasound generation according to claim 1, characterized in that: The specific steps for acquiring preformed wave cavity pressure observation data, performing time axis repositioning and period scale unification, and generating a preformed wave cavity pressure observation dataset are as follows: The inlet pressure, central axis pressure, outlet pressure, residual pressure, and external release pressure during the infrasound generation process are collected; the cavity pressure loss type, target infrasound frequency, and corrected sound velocity are set; the inlet position, central axis pressure acquisition point, outlet pressure acquisition point, equivalent propagation distance of phase coordinates, and propagation velocity are calibrated; and the peak values ​​of central axis pressure, inlet pressure, outlet pressure, and the corresponding times are calculated; all collected, set, calibrated, and calculated data are summarized to form pre-wave cavity pressure observation data; Based on the time axis relocation method of driving cycle synchronization marking, the pressure sampling time is corrected, and the pressure data within the same target infrasound frequency period is mapped to a unified period coordinate. The peak noise generated at the moment of driving source start-up, the airflow impact at the phase guide port and the residual pressure fall-off stage is dynamically denoised by the pressure change rate constraint filtering method. The pre-wave cavity pressure observation data is normalized by the peak amplitude scaling and phase coordinate mapping method based on the target infrasound frequency period segmentation, and the pressure data is unified to a period scale to generate a pre-wave pressure chain dataset.

3. The resonant cavity pressure feedback control method for infrasound generation according to claim 1, characterized in that: The specific steps for calculating the peak phase difference based on the pre-formed wave pressure chain dataset are as follows: Based on the pre-formed wave pressure chain dataset, the structural distance from the resonant cavity inlet to the central axis pressure acquisition point is obtained through sensor installation location calibration; the allowable deviation of the structural distance is obtained by synthesizing the equivalent distance based on assembly tolerance and frequency drift using the root mean square method; the central axis pressure gain is obtained by the ratio of the central axis pressure peak value to the inlet pressure peak value; the actual phase difference from the inlet to the central axis is obtained by converting the time difference between the inlet pressure peak value and the central axis pressure peak value to the current target infrasonic frequency periodic phase coordinate; the actual phase difference from the central axis to the outlet is obtained by converting the time difference between the central axis pressure peak value and the outlet pressure peak value to the current target infrasonic frequency periodic phase coordinate; the expected phase difference from the inlet to the central axis is determined by the target infrasonic frequency, the equivalent propagation distance of the phase coordinates from the inlet to the central axis, and the calibrated propagation speed; the expected phase difference from the central axis to the outlet is determined by the target infrasonic frequency, the equivalent propagation distance from the central axis to the outlet, and the calibrated propagation speed; the residual pressure ratio at the end of the cycle is obtained by comparing the residual amount of the central axis pressure and the outlet pressure at the end of the current cycle with the current cycle pressure peak value.

4. The resonant cavity pressure feedback control method for infrasound generation according to claim 1, characterized in that: The specific steps for evaluating the pre-wave pressure window adaptation are as follows: The difference between the structural distance and the target infrasound frequency and four times the corrected sound velocity is divided by the sum of the allowable deviation of the structural distance and the zero-prevention term, squared, and then the negative exponent is taken to obtain the wavelength position fitting term; the central axis pressure gain is added by one, the natural logarithm is taken, and then divided by the result of adding one and the natural logarithm is taken to obtain the central axis pressure enhancement term. The phase deviation term from the inlet to the central axis is obtained by subtracting the actual phase difference from the inlet to the central axis from the expected phase difference, dividing by two, taking the sine value, and then taking the absolute value. The phase deviation term from the central axis to the outlet is obtained by subtracting the actual phase difference from the central axis to the outlet from the expected phase difference, dividing by two, taking the sine value, and then taking the absolute value. The phase deviation term from the central axis to the outlet is obtained by adding the two phase deviation terms and taking the negative exponent. Subtract the upper limit of allowable residual pressure from the percentage of residual pressure at the end of the cycle, multiply by the residual pressure gating coefficient, perform an exponential operation, add a constant one, and take the reciprocal to obtain the pressure suppression term; multiply the wavelength position matching term, the central axis pressure enhancement term, the phase transmission coordination term, and the pressure suppression term in sequence to obtain the pre-formed wave pressure window adaptation value.

5. The resonant cavity pressure feedback control method for infrasound generation according to claim 1, characterized in that: The specific steps for constructing the pre-wave pressure window dataset are as follows: When the preformed wave pressure window adaptation value is less than the preformed wave adaptation threshold, it is determined that no effective preformed wave pressure window has been formed in the current period, and the current period is marked as an invalid pressure window period; when the preformed wave pressure window adaptation value is greater than or equal to the preformed wave adaptation threshold, it is determined that an effective preformed wave pressure window has been formed in the current period, and the current period is marked as an effective pressure window period; period-level pressure window labels are formed and summarized in order of target infrasound frequency to construct a preformed wave pressure window dataset.

6. The resonant cavity pressure feedback control method for infrasound generation according to claim 1, characterized in that: The specific steps for evaluating the cavity pressure loss sequence deviation based on the pre-formed wave pressure window dataset are as follows: Based on the pre-formed wave pressure window dataset, the controller divides the sampling period according to the target infrasound frequency and sequentially numbers it to obtain the target infrasound frequency period number; through the cavity pressure loss type enumeration mapping method, abnormal states are assigned state numbers to obtain candidate cavity pressure loss state numbers; the candidate cavity pressure loss state numbers, along with their corresponding peak amplitude ratio, phase difference, phase interval within the target infrasound frequency period, and upper and lower limits of actuator adjustment, are written into the same state mapping table to obtain the candidate cavity pressure loss state set; the peak times of inlet pressure, central axis pressure, and outlet pressure in the current period are read and sorted chronologically to obtain the actual occurrence order; the typical occurrence order of inlet pressure peak, central axis pressure peak, and outlet pressure peak under each type of cavity pressure loss state is calibrated to obtain the standard peak order; The order deviation value is obtained by calculating the number of adjacent position swaps required to adjust the actual peak order to the standard peak order.

7. The resonant cavity pressure feedback control method for infrasound generation according to claim 1, characterized in that: The specific steps for performing cavity pressure loss state feature matching are as follows: The inlet pressure, central axis pressure, outlet pressure, external release pressure, and residual pressure of the current cycle are linearly normalized to obtain state characteristic quantities. Standard characteristic values ​​are obtained by statistically analyzing sample cycles under cavity pressure loss states. The fluctuation amplitude of each state characteristic quantity under the same cavity pressure loss state is extracted, and the quantile deviation of the corresponding characteristic quantity is used as the characteristic allowable deviation under the state. Windows that exceed the anomaly judgment conditions in the inlet compression window, central axis concentration window, outlet release window, and residual pressure dissipation window are detected respectively to obtain the pressure windows that trigger anomalies. Through the calibration method of state-window correspondence, the range of pressure windows that allow anomalies to occur first for each type of cavity pressure loss state is recorded to obtain the set of trigger windows. Subtract the standard feature value corresponding to the candidate cavity pressure loss state from each state feature quantity, take the absolute value and divide it by the sum of the feature allowable deviation and the zero prevention term to obtain the deviation degree; select the maximum value from the deviation degrees of all state feature quantities as the maximum deviation term; determine whether the pressure window that triggers the anomaly in the current cycle belongs to the trigger window set; if it does, the window penalty term is zero; if it does not, the trigger window mismatch penalty coefficient is used as the window penalty term; add the sequence deviation value, the maximum deviation term and the window penalty term to obtain the total state deviation value, and select the candidate state with the smallest total state deviation value to obtain the cavity pressure loss state matching sequence number.

8. The resonant cavity pressure feedback control method for infrasound generation according to claim 1, characterized in that: The specific steps for generating cavity pressure loss state data and boundary constraint data are as follows: Based on the cavity pressure loss state matching sequence number, the corresponding cavity pressure loss type is read: When the cavity pressure loss state matching sequence number equals the first state sequence number, the inlet coupling insufficiency state is determined, indicating that the drive source output has failed to be converted into inlet pressure, and the inlet connection, sealing state, and drive source coupling state are checked; when the cavity pressure loss state matching sequence number equals the second state sequence number, the structural constraint insufficiency state is determined, indicating that the inlet pressure has been established but has not been concentrated in the central axis region, and the adjustable range of the cavity angle state is generated; when the cavity pressure loss state matching sequence number equals the third state sequence number, the outlet premature pressure relief state is determined, indicating that the pressure before the outlet is released earlier than the central axis pressure concentration process. Release, shifting the outlet release time later; when the cavity pressure loss state matching number equals the fourth state number, determine the cavity stagnation state, indicating that the central axis pressure has been formed but failed to be transmitted to the outside of the guide phase port, and the cavity pressure release path is opened; when the cavity pressure loss state matching number equals the fifth state number, determine the echo standing wave abnormal state, and generate the drive amplitude range, frequency back-off range, and residual pressure dissipation waiting time; if the cavity pressure loss state matching number does not belong to the above five state numbers, determine that the state identification of the current cycle is invalid, and stop generating the closed-loop regulation of the current cycle; perform attribution analysis according to the current cavity pressure loss type, and simultaneously generate boundary constraint data.

9. A resonant cavity pressure feedback control method for infrasound generation according to claim 1, characterized in that: The specific steps for constructing the control direction matrix based on boundary constraint data are as follows: The adjustment step size is determined based on the cavity pressure loss state type and the number of abnormal cycles of the current pre-formed wave pressure window adaptation value; a control direction matrix is ​​constructed using the state-to-execution mapping method; the gain statistics of the peak value of the central axis pressure and the peak value of the inlet pressure are calculated, and the median is taken to obtain the target central axis pressure gain; the difference between the peak value of the external release pressure and the peak value of the outlet pressure is calculated, and the time difference penalty from the peak value of the outlet pressure to the peak value of the external release pressure is combined to obtain the target outlet release holding value; the amplitude of the peak value of the external release pressure is compared with that of the peak value of the outlet pressure to obtain the actual outlet release holding value; the residual amount of the central axis pressure and the residual amount of the outlet pressure at the end of the current cycle are read and compared with the total peak amount of the peak value of the central axis pressure and the peak value of the outlet pressure in the current cycle to obtain the cycle residual pressure ratio; at the target infrasound frequency, a cycle in which a pre-formed wave pressure window has been formed and no echo standing wave anomaly has appeared is selected, and the target residual pressure ratio is obtained by proportional calculation; After adding the zero-prevention term to the target axial pressure gain, divide by the sum of the axial pressure gain and the zero-prevention term, and then take the natural logarithm to obtain the axial pressure concentration error. Add the zero-prevention term to the target outlet release holding value, divide by the sum of the outlet release holding value and the zero-prevention term, and take the natural logarithm to obtain the outlet release holding error; subtract the target residual pressure ratio from the residual pressure ratio to obtain the residual pressure error; combine the central axis pressure concentration error, outlet release holding error, and residual pressure error into an error vector, and perform hyperbolic tangent operation on the error vector to obtain the error adjustment vector; multiply the error adjustment vector by the control direction matrix to obtain the correction direction; multiply the correction direction by the adjustment step size to obtain the control correction amount; subtract the control correction amount from the control vector of the current cycle to obtain the control vector of the next cycle; limit the control vector of the next cycle between the lower control limit and the upper control limit to obtain the closed-loop control vector.

10. A resonant cavity pressure feedback control method for infrasound generation according to claim 1, characterized in that: The specific steps for writing the closed-loop control vector into the controller and generating the locking identifier and parameter fingerprint are as follows: The next cycle closed-loop control vector is written into the controller, and the output waveform, output intensity, cavity guide boundary position, and phase port release section of the drive source are adjusted so that the inlet pressure peak, central axis pressure peak, and outlet release peak re-enter the pre-formed wave pressure transmission sequence within the next target infrasonic frequency cycle. When the pre-formed wave pressure window adaptation value is greater than the pre-formed wave adaptation threshold, the cavity pressure loss state does not show the premature outlet pressure relief state, and the echo standing wave abnormal state are simultaneously satisfied within m target infrasonic frequency cycles, a locking flag is generated. After generating the locking flag, the parameter record is written in the order of the target infrasonic frequency cycle, and the closed-loop control vector and target infrasonic frequency corresponding to each cycle are recorded in sequence to form a parameter fingerprint.

Images