Air conditioner terminal air outlet noise suppression adjustment method

Abstract

The application provides an air conditioner terminal air outlet noise suppression adjustment method, comprising: obtaining transmission sound pressure data and sealing member compression force distribution data from the center area of the cover plate and the gaps around the cover plate, obtaining acoustic consistency indexes of the cover plate area and frame pressure uniformity indexes; according to the acoustic consistency indexes and the frame pressure uniformity indexes, using a support vector machine algorithm to analyze the correlation between the transmission sound pressure data and the sealing member compression force distribution data, determining the potential position distribution of the gap leakage channel; after obtaining the optimized compression force distribution parameters, re-collecting the leakage sound pressure data at the gaps around the cover plate, determining the noise radiation attenuation increment; according to a further lock buckle pre-tightening reference adjustment scheme, obtaining compression force distribution update data, judging the overall consistency improvement level of the sound wave transmission coefficient of the cover plate area; extracting a noise attenuation index from the judged overall consistency improvement level, and determining an air conditioner terminal air outlet noise suppression adjustment scheme.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for suppressing and adjusting noise at the air outlet of an air conditioning terminal. Background Technology

[0002] In the field of building environmental noise control, noise suppression technology for air conditioning terminal vents is particularly important, as it directly affects the comfort of indoor spaces and the health experience of users. With increasing demands for quality of life, reducing noise generated during equipment operation has become an unavoidable issue in the industry, especially in scenarios with high requirements for quiet environments, such as offices, hotels, or residences. The effectiveness of noise control often becomes a crucial standard for evaluating the performance of air conditioning systems. However, current common methods for noise suppression at air conditioning terminal vents often neglect the complex balance between material selection and the actual installation structure. Many solutions focus solely on the sound insulation performance of the materials themselves, failing to adequately consider new problems arising from installation methods or structural design. This neglect leads to a significant reduction in noise control effectiveness in practical applications, and may even be counterproductive, failing to fundamentally improve overall sound insulation capabilities. The cover plate, as an essential component for air conditioning system maintenance, needs to maintain similar acoustic characteristics to the ceiling material to reduce noise transmission, while also being operable. This inevitably introduces mechanical structures such as frames and latches. However, when the cover material is changed to the same material as the ceiling, the increased weight directly affects the stress on the frame seals, leading to uneven sealing and the formation of tiny gaps around the cover. These gaps become the main channels for noise leakage, compromising the overall sound insulation. In a high-end office building's air conditioning system renovation project, the design team changed the access panel material to the same composite material as the ceiling to improve sound insulation. However, after installation, it was found that due to the increased weight of the cover, the frame was under unbalanced stress. The seals in some areas were over-compressed, while others became loose, forming gaps that were difficult to detect with the naked eye. Ultimately, the noise from the fan during operation was transmitted downwards through these gaps, affecting the quiet environment of the offices below. Therefore, how to solve the problem of uneven frame sealing caused by changes in material weight while ensuring the consistent acoustic characteristics of the access panel, and thus effectively reduce noise leakage from gaps, has become a key problem that this research urgently needs to address. Summary of the Invention

[0003] This invention provides a method for noise suppression and adjustment at the air outlet of an air conditioning terminal, mainly comprising: Transmitted sound pressure data and sealing force distribution data are obtained from the central area and surrounding gaps of the cover plate to obtain the acoustic consistency index and the edge pressure uniformity index of the cover plate area. Based on the acoustic consistency index and the edge pressure uniformity index, the correlation between the transmitted sound pressure data and the sealing force distribution data is analyzed using a support vector machine algorithm to determine the potential location distribution of gap leakage channels. If the potential location distribution range exceeds a preset threshold, the degree of insufficient clamping force corresponding to each leakage point is fitted using a least squares algorithm to calculate the pre-tightening reference adjustment amount of each latch and generate a pre-tightening reference update value. Based on the pre-tightening reference update value, the overall edge sealing is iteratively optimized using a gradient descent algorithm. The clamping force is distributed until the coefficient of variation of the clamping force in the circumferential direction is lower than a set limit, indicating that a uniform state has been reached, and optimized clamping force distribution parameters are obtained. Leakage sound pressure data at the gaps around the cover plate are re-collected to determine the attenuation increment of noise radiation. If the attenuation increment is lower than a preset threshold, the deviation between the collected leakage sound pressure data and the initial clamping force distribution data is re-analyzed using a support vector machine algorithm to obtain a latch pre-tightening benchmark adjustment scheme. Based on the latch pre-tightening benchmark adjustment scheme, updated clamping force distribution data is obtained to determine the overall consistency improvement level of the sound wave transmission coefficient in the cover plate area. The noise attenuation index is extracted from the overall consistency improvement level to determine the noise suppression adjustment scheme for the air conditioning terminal vent.

[0004] Furthermore, transmitted sound pressure data and sealing force distribution data are obtained from the central area and surrounding gaps of the cover plate to obtain the acoustic consistency index and the pressure uniformity index of the cover plate area. This includes: arranging sound pressure pickup units in the central area of ​​the cover plate at a preset grid spacing; arranging another set of sound pressure pickup units circumferentially in the gaps of the cover plate's surrounding edges; embedding a thin-film pressure-sensitive unit circumferentially in the sealing element of the edge; normalizing the dispersion of the difference between the two sets of sound pressure amplitudes at each sampling point to obtain the acoustic consistency index; and normalizing the offset of the average pressure force segment by segment along the circumferential direction of the edge to obtain the pressure uniformity index of the edge.

[0005] Furthermore, based on the acoustic consistency index and the uniformity of pressure on the frame, a support vector machine algorithm is used to analyze the correlation between the transmitted sound pressure data and the sealing force distribution data to determine the potential location distribution of the gap leakage channel. This includes: extracting acoustic deviation values, pressure force offset amplitudes, and frequency band deviation values ​​from each sampling point around the cover plate to form a feature vector; assigning positive and negative labels to confirmed leakage points and intact sealing points respectively; training a support vector machine with radial basis kernels to separate the hyperplane; using the directed distance from each feature vector to the hyperplane as the decision value; merging adjacent points with decision values ​​higher than a preset threshold into continuous candidate segments, and mapping them to the four sides of the cover plate to determine the potential location distribution.

[0006] Furthermore, if the potential location distribution range exceeds a preset threshold, the degree of insufficient clamping force corresponding to each leakage point is fitted using a least squares algorithm, the pre-tightening reference adjustment amount of each latch is calculated, and a pre-tightening reference update value is generated. This includes: reading the absolute difference between the measured clamping force value and the overall mean value of the circumferential side of the frame for each candidate leakage point within the potential location distribution to obtain the degree of insufficient clamping force; performing polynomial fitting with the circumferential installation position of the latch as the independent variable to obtain a continuous fitting curve; taking the corresponding function value at the circumferential coordinate of each latch; and converting it into the pre-tightening reference adjustment amount according to the preset correspondence between the sealing compression amount and the latch pre-tightening displacement.

[0007] Furthermore, based on the pre-tightening benchmark update value, a gradient descent algorithm is used to iteratively optimize the overall clamping force distribution of the frame seal until the coefficient of variation of the clamping force in the circumferential direction is lower than a set limit, indicating that a uniform state has been reached. The optimized clamping force distribution parameters are then obtained, including: using the pre-tightening displacement increment corresponding to each latch on the cover plate frame as the initial iteration value; the pre-tightening displacement values ​​of all latches constitute the pre-tightening displacement vector; using the sum of squares of the deviations of the clamping force output by the membrane pressure-sensitive unit relative to the overall circumferential mean as the objective function; in each iteration, a perturbation is applied to each component of the pre-tightening displacement vector; the clamping force is reread from the membrane pressure-sensitive unit to obtain the partial derivative of the objective function with respect to each component; after synthesizing the gradient direction, the pre-tightening displacement vector is updated along the gradient in the opposite direction with a preset step size; the coefficient of variation is obtained along the circumferential direction of the frame by the ratio of the standard deviation of the clamping force to the arithmetic mean; if the coefficient of variation is lower than a set limit, the iteration is terminated, and the pre-tightening displacement vector and the corresponding clamping force at that moment are taken as the optimized clamping force distribution parameters.

[0008] Furthermore, leakage sound pressure data at the gaps around the cover plate are re-collected to determine the attenuation increment of noise radiation. This includes: re-collecting leakage sound pressure data at the gaps around the cover plate using sound pressure pickup units under the same steady-state operating conditions of the air conditioning terminal fan to obtain the optimized leakage sound pressure amplitude; subtracting the optimized sound pressure amplitude from the initial sound pressure amplitude for each sampling point along the circumferential direction of the frame to obtain the difference, and taking the arithmetic mean of the differences for all sampling points to determine the attenuation increment of noise radiation.

[0009] Furthermore, if the attenuation increment is lower than a preset threshold, the deviation between the collected leakage sound pressure data and the initial clamping force distribution data is re-analyzed using a support vector machine algorithm to obtain a lock pre-tightening reference adjustment scheme. This includes: reading the sound pressure amplitude from the collected leakage sound pressure data at each sampling point along the circumferential direction of the gap around the cover plate, and pairing it with the measured clamping force values ​​at the same points in the initial clamping force distribution data to obtain a set of residual sound pressure clamping force deviation pairs; using radial basis function support vector machine regression to fit the residual deviation pair set to output a residual leakage tendency value, and using points higher than a preset limit as residual weak points; matching a corresponding lock for each residual weak point according to the circumferential proximity principle, and calculating the additional pre-tightening displacement increment according to the residual leakage tendency value based on the preset correspondence between the seal compression amount and the lock pre-tightening displacement, and archiving each lock alignment to form the lock pre-tightening reference adjustment scheme.

[0010] Furthermore, according to the locking pre-tightening reference adjustment scheme, the updated data of the clamping force distribution is obtained, including: synchronously applying feed to each locking buckle arranged circumferentially along the cover plate frame according to the archived additional pre-tightening displacement increment, and reading the new round of clamping force values ​​from the sampling points of the diaphragm pressure sensitive unit to obtain the updated data of the clamping force distribution.

[0011] Furthermore, a noise attenuation index is extracted from the overall consistency improvement level to determine the noise suppression adjustment scheme for the air conditioning terminal vents. This includes: taking the arithmetic mean of the sound pressure amplitude from the archived sound pressure sequences at the gaps before and after this round of adjustment, and subtracting the arithmetic mean after adjustment from the arithmetic mean before adjustment to obtain the noise attenuation index; collecting radiated sound pressure data by deploying distributed sound pressure pickup units in a preset grid in the area below the air conditioning terminal vents, taking the arithmetic mean and maximum value of the radiated sound pressure amplitude along the grid nodes, and when the arithmetic mean is lower than a preset suppression threshold and the maximum value is lower than a preset peak threshold, archiving and optimizing the pre-tightening displacement vector and the corresponding latch pre-tightening displacement increment in the clamping force distribution parameters to determine the noise suppression adjustment scheme for the air conditioning terminal vents.

[0012] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a method for noise suppression and adjustment of air conditioner terminal vents. Addressing the acoustic consistency problem caused by noise leakage and uneven sealing at air conditioner terminal vents, it innovatively integrates sound pressure data acquisition, clamping force distribution optimization, and noise attenuation analysis to construct a complete solution. First, this invention acquires transmitted sound pressure and clamping force data from the center of the cover plate and the gaps. Using a support vector machine algorithm, the correlation between these two data points is analyzed to accurately locate the potential distribution of leakage channels in the gaps. Then, the clamping force distribution along the frame is iteratively optimized using the least squares method and gradient descent algorithm to ensure circumferential uniformity. Finally, radiated sound pressure data is acquired based on a distributed sound pressure pickup unit to generate a noise suppression and adjustment scheme. This invention, through multi-algorithm collaboration and data-driven approaches, significantly improves the consistency of sound wave transmission coefficient and noise attenuation effect in the cover plate area, providing efficient and precise technical support for noise control of air conditioner terminal vents. Attached Figure Description

[0013] Figure 1 This is a flowchart of a noise suppression and adjustment method for air conditioning terminal vents according to the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and thoroughly described below with reference to the accompanying drawings. The described embodiments are merely some embodiments of the present invention.

[0015] like Figure 1 This embodiment of a method for suppressing and adjusting noise at the air conditioner terminal vent may specifically include: Step S101: Obtain transmitted sound pressure data and sealing force distribution data from the central area and surrounding gaps of the cover plate to obtain the acoustic consistency index of the cover plate area and the uniformity of pressure on the frame.

[0016] Sound pressure pickup units are arranged in a grid at equal intervals in the central area of ​​the cover plate, and another set of sound pressure pickup units are densely arranged circumferentially in the gaps of the four sides of the cover plate. For steady-state noise radiation under the operating conditions of the air conditioner terminal fan, the transmitted sound pressure data of the central area and the transmitted sound pressure data of the gaps are continuously acquired from the back cavity side of the cover plate. Simultaneously, thin-film pressure sensitive units are embedded in each sampling point of the circumferential side of the frame seal to acquire the pressure distribution data of the seal and establish the spatial correspondence between the sampling point and the sensor layout position. Based on the acquired sound pressure data transmitted from the central area and the gap, the amplitude difference between the two sets of sound pressure at the same frequency band is compared for each sampling point. The dispersion of the difference within the cover plate area is normalized to obtain the acoustic consistency index of the cover plate area. For the sealing force distribution data, the circumference of the frame is divided into several segments according to arc length. The average pressure force of each segment is statistically analyzed, and its deviation from the overall circumferential average is calculated. The dispersion of the deviation is normalized to obtain the frame pressure uniformity index. Based on the acoustic consistency index and the frame pressure uniformity index, data is paired according to the circumferential position of the cover plate area. The acoustic deviation value and the pressure force deviation amplitude at each sampling point are archived together to determine the corresponding sets of indices for the central area and gap of the cover plate in terms of both acoustic performance and mechanical pressure.

[0017] This addresses the issue of uneven sealing of the frame caused by the increased weight after the cover plate was replaced with a composite material consistent with the ceiling.

[0018] In one possible implementation, sound pressure pickup units are arranged in a grid with equal spacing in the central area of ​​the cover plate. The grid nodes are electret microphones with their diaphragms facing the back cavity side of the cover plate. The node spacing is evenly divided according to the diagonal length of the cover plate, covering the main area of ​​the central region. During installation, the microphones are fixed to the back cavity side support frame by flexible clamps to avoid rigid contact with the cover plate body, reducing the interference of structural vibration on the acquisition results.

[0019] For example, another set of sound pressure pickup units is densely arranged circumferentially in the gaps of the cover plate's perimeter frame, with a higher density than in the central area. The arc length spacing between adjacent pickup units is less than one-twelfth of the length of the cover plate's long side. Simultaneously, thin-film pressure-sensitive units are embedded at each sampling point circumferentially of the frame seal. These sensitive units are in close contact with the contact surface between the seal and the frame's metal profile, and their outputs are led out to the data acquisition host via shielded wires to continuously acquire data on the seal's clamping force distribution.

[0020] Specifically, the construction process of the acoustic consistency index for the cover plate area includes dividing the transmitted sound pressure data of the central area and the transmitted sound pressure data at the gap under the steady-state operation of the air conditioning terminal fan into three frequency band windows: low frequency, mid frequency, and high frequency, according to octave bands. Within each frequency band window, the amplitude difference between the two sets of sound pressures is calculated for each sampling point, and recorded as the frequency band deviation value at that point. The frequency band deviation values ​​at all points are then analyzed for the overall dispersion of the cover plate area. The dispersion is represented by the ratio of the range to the median, and then normalized to the zero-to-one interval to obtain the acoustic consistency index. The closer the index value is to zero, the better the consistency of the sound pressure response between the central area and the gap. Further, regarding the construction process of the edge pressure uniformity index, the circumference of the edge is divided into several segments according to the arc length, and the number of segments matches the density of the sound pressure pickup units at the gap. The output values ​​of all thin-film pressure-sensitive units within a segment are read segment by segment, and the arithmetic mean is taken as the average clamping force of the segment, denoted as Fi, where i represents the segment number. Then, the overall average value Fm of Fi for all segments is calculated. The absolute value of the difference between Fi and Fm for each segment is calculated as the proportion of Fm to the clamping force offset of that segment. The dispersion of the offset amplitudes for all segments is calculated and normalized to obtain the uniformity index of the frame pressure. The spatial correspondence between the sampling points and the sensor deployment positions is established once during the calibration phase. During calibration, the origin of the cover plate coordinate system is set at the geometric center of the central area, and the coordinate components of each sound pressure pickup unit and each thin-film pressure-sensitive unit are recorded to form a spatial correspondence table. Subsequent data archiving is indexed based on this correspondence table.

[0021] Preferably, data pairing is performed on the cover plate area according to the circumferential position. During pairing, the circumferential coordinates of the sound pressure pickup unit at the gap are used as the reference to match the nearest thin-film pressure sensitive unit. The acoustic deviation value corresponding to this point and the clamping force offset amplitude are archived in the same record entry. All record entries are summarized into the set corresponding to the index.

[0022] Step S102: Based on the acoustic consistency index and the uniformity of pressure on the frame, the support vector machine algorithm is used to analyze the correlation between the transmitted sound pressure data and the sealing force distribution data to determine the potential location distribution of the gap leakage channel.

[0023] Based on the acoustic consistency index and the uniformity of pressure on the frame, acoustic deviation values ​​and clamping force offset amplitudes are extracted as basic components for each sampling point in the circumferential direction of the cover plate. These are then superimposed with the frequency band deviation values ​​of the sampling point within the low-frequency, mid-frequency, and high-frequency windows, and concatenated in dimensional order to form the feature vectors for each sampling point. Sampling points confirmed to have leaks during the calibration phase are assigned positive class labels, while those with intact seals are assigned negative class labels, resulting in a labeled sample set. This completes the feature vector construction and sample label assignment. Based on the labeled sample set, a support vector machine algorithm is used to fit the correlation between the transmitted sound pressure data and the sealing force distribution data. A radial basis function kernel is selected to handle the nonlinear coupling of the feature vectors in the frequency band dimension and the pressure dimension. A hyperplane separating the positive and negative classes is trained on the feature vectors of each sampling point. The directed distance from the feature vector of each sampling point to be judged to the hyperplane is calculated as the correlation decision value for that point, thus determining the decision value sequence for all sampling points in the circumferential direction of the cover plate. Based on the decision value sequence, the decision values ​​are sorted according to the circumferential coordinates of the cover plate. Sampling points with decision values ​​higher than a preset threshold are marked as potential leakage candidate points. Cases where the arc length distance between adjacent candidate points is less than a preset distance threshold are merged into continuous candidate segments. The circumferential coordinate range of the candidate segments is mapped back to the four sides of the cover plate to determine the potential location distribution of the gap leakage channel.

[0024] In one possible implementation, for each sampling point in the circumference of the cover plate, the acoustic deviation value and the clamping force offset amplitude of the point are taken from the corresponding set of indicators as the first two dimensions. Then, the frequency band deviation value of the point in the three octave band windows of low frequency, mid frequency and high frequency are added in, and the components are concatenated in dimensional order to form a five-dimensional feature vector.

[0025] Preferably, the center frequencies of the frequency band window are selected as 63 Hz, 500 Hz and 4000 Hz, respectively, which correspond to three typical noise components: low-frequency eddy current noise of the fan, medium-frequency duct transmission noise and high-frequency turbulent noise of the air outlet, and can cover the main frequency range of noise from the air conditioning terminal.

[0026] It should be noted that the labels of the labeled sample set are derived from the calibration stage.

[0027] Specifically, during the factory inspection of the cover plate or the initial on-site installation and commissioning, smoke tracing or manual tightening of the seals, supplemented by auscultation, is used to confirm the presence of leaks at each point. Leaking points are assigned a positive label, and well-sealed points are assigned a negative label, thus forming a labeled sample set. The core idea of ​​the Support Vector Machine (SVM) algorithm is to find a hyperplane in the high-dimensional space of the feature vectors that can separate positive and negative samples with the maximum margin for the labeled sample set. The maximum margin refers to the maximum perpendicular distance between the nearest sample point to the hyperplane and the hyperplane itself; the sample point closest to the hyperplane is the support vector.

[0028] Specifically, in this embodiment, the eigenvectors exhibit a nonlinear coupling relationship between acoustic deviation and compressive offset. A simple linear kernel is insufficient to adequately fit this relationship; therefore, a radial basis kernel is chosen. The radial basis kernel can be represented as K(xi,xj)=exp(-γ·‖xi-xj‖). 2 ), where xi and xj represent the feature vectors of two sampling points, and ‖xi-xj‖ 2 Let represent the squared Euclidean distance between the two vectors, and γ represent the bandwidth parameter of the kernel function. The value of γ is determined within the labeled sample set through cross-validation. During the training phase, the algorithm uses the sequential minimum optimization method to solve the dual problem, obtains the Lagrange multipliers corresponding to each support vector, and then constructs the hyperplane that separates the positive and negative classes.

[0029] In one embodiment, the penalty factor is set to 1, and γ is a constant multiple of the reciprocal of the feature vector dimension, to balance overfitting control and fitting accuracy. Further, for each sampling point to be determined, its feature vector is substituted into the discriminant function of the hyperplane to obtain the directed distance from the feature vector of that point to the hyperplane. The sign of the directed distance indicates whether the point belongs to the positive or negative class; a positive sign indicates a tendency towards leakage, and a negative sign indicates a tendency towards good sealing. The absolute value of the directed distance indicates the strength of the point's deviation from the hyperplane; a larger absolute value indicates a more certain classification. After calculation point by point, a sequence of correlation decision values ​​arranged circumferentially along the cover plate is obtained.

[0030] It is understandable that, after the decision value sequence is reordered along the circumferential coordinates of the cover plate, it can intuitively reflect the fluctuation trend of leakage tendency in the circumferential direction of the frame. Sampling points with decision values ​​higher than a preset threshold are marked as potential leakage candidate points.

[0031] For example, the preset threshold is determined during the calibration phase using the receiver operating characteristic (ROC) curve, and is set to the decision value level corresponding to a positive class recognition rate of not less than 0.9 and a negative class misclassification rate of not more than 0.1. Further, if the distance between adjacent candidate points along the arc length of the border is less than a preset distance threshold, they are merged into the same continuous candidate segment.

[0032] For example, the preset spacing threshold is taken as one-twenty-fourth of the perimeter of the cover plate frame, corresponding to the adjacent span of a sound pressure pickup unit. Adjacent candidate points with a spacing less than this threshold are considered as the same weak sealing area. The start and end values ​​of the circumferential coordinates of each consecutive candidate segment after merging are mapped back to the four sides of the cover plate frame and marked on the physical location of the frame to determine the potential location distribution of the gap leakage channel. This embodiment transforms the original manual inspection process, which could only rely on point-by-point auscultation or smoke tracing, into a quantitative judgment process based on sound pressure data and clamping force data. It can objectively locate the weak sealing location of the frame under normal fan operation conditions, providing a positional basis for subsequent pre-tightening adjustment of the latches.

[0033] Step S103: If the distribution range of potential locations of the determined gap leakage channels exceeds the preset threshold, the degree of insufficient clamping force corresponding to each leakage point is fitted by the least squares algorithm, the amount of pre-tightening reference adjustment to be applied to each latch is calculated, and the updated value of pre-tightening reference for local weak locations is generated.

[0034] If the ratio of the total arc length of each consecutive candidate segment in the potential location distribution of the gap leakage channel to the perimeter of the cover plate frame exceeds a preset threshold, then for each potential leakage candidate point within the consecutive candidate segment, the measured value of the clamping force at that point is read from the sealing force distribution data, and the difference is taken from the overall mean value of the frame circumferential direction to obtain the degree of insufficient clamping force at that point. Based on the discrete data sequence formed by arranging the degree of insufficient clamping force of each potential leakage candidate point along the circumferential coordinate of the cover plate, the least squares algorithm is used to perform polynomial fitting with the circumferential installation position of the latch as the independent variable and the degree of insufficient clamping force as the dependent variable to obtain a fitting curve of the degree of insufficient clamping force that continuously changes along the circumferential direction. For each latch's circumferential coordinate, the corresponding function value is taken on the fitting curve to obtain the estimated value of the degree of insufficient clamping force at that latch. Based on the estimated degree of insufficient clamping force at each latch, the pre-tightening reference adjustment amount is calculated for each latch according to the preset correspondence between the compression amount of the seal and the pre-tightening displacement of the latch. The pre-tightening reference adjustment amount of each latch is calculated and the adjustment amount is filed for each latch to obtain the updated pre-tightening reference value for the local weak position.

[0035] In one possible implementation, the potential locations of the gap leakage channels are distributed in a circumferential direction around the edge as several consecutive candidate segments.

[0036] Specifically, the starting and ending circumferential coordinates of the candidate segments along the edge arc are read segment by segment, and the difference is the arc length of that segment. Then, the arc lengths of all candidate segments are summed, and the total arc length is divided by the total perimeter of the cover plate edge to obtain the coverage ratio of the candidate segments.

[0037] Preferably, the preset threshold is 0.15. When the coverage ratio exceeds 0.15, it indicates that the weak area is no longer limited to individual points, triggering the subsequent pre-tightening benchmark adjustment process. The determination of insufficient clamping force involves, for each potential leak candidate point, extracting the output value of the diaphragm pressure-sensitive unit at that point from the archived seal clamping force distribution data as the measured clamping force value, then taking the arithmetic mean of the output values ​​of all diaphragm pressure-sensitive units along the circumference of the frame as the overall circumferential mean, and subtracting the two values ​​and taking the absolute value. The larger the difference, the further the clamping force at that point deviates from the overall level, and the more insufficient the seal pressure.

[0038] Specifically, the principle behind the least squares algorithm's fitting process for insufficient clamping force lies in obtaining a set of polynomial coefficients that minimize the sum of squared residuals between the measured insufficient clamping force at each potential leak candidate location and the predicted value of the polynomial function. The sum of squared residuals is represented as S=Σ(yk-f(xk)). 2 , where xk represents the circumferential coordinate of the k-th candidate point, yk represents the degree of insufficient clamping force at that point, and f(xk) represents the function value of the polynomial function at xk. The summation iterates through all candidate points.

[0039] Preferably, the polynomial order is 3, corresponding to a cubic polynomial f(x) = a0 + a1·x + a2·x 2 +a3·x 3 Where a0 to a3 represent undetermined coefficients, by taking the partial derivatives of S with respect to a0 to a3 and setting the partial derivatives to zero, a system of linear equations is constructed and solved to obtain the values ​​of the undetermined coefficients, thereby determining the fitting curve for the degree of insufficient clamping force. Further, for several latches evenly distributed circumferentially on the cover plate frame, the circumferential installation coordinates of each latch are read one by one, and these coordinates are substituted into the fitting curve to obtain the function value, which is used as an estimate of the degree of insufficient clamping force at that latch.

[0040] For example, eight latches are arranged circumferentially along the edge of the cover plate, spaced apart at the four corners and the middle of the four sides of the edge, corresponding to eight circumferential coordinate values ​​respectively.

[0041] It is understandable that there is a pre-defined correspondence between the seal compression and the latch preload displacement. This pre-defined correspondence is obtained during the uniaxial pressure test of the seal during the cover plate assembly stage. For the homogeneous composite material seals used, the clamping force values ​​borne by the seal under different compression values ​​and the corresponding feed displacement values ​​of the latch preload mechanism are recorded, forming a three-column comparison table of clamping force, compression, and preload displacement. Further, for the estimated degree of insufficient clamping force at each latch, the required additional compression of the seal is obtained by looking up the clamping force difference in the comparison table, and then the required additional preload displacement of the latch preload mechanism is obtained by looking up the compression value. The resulting additional displacement is the preload reference adjustment amount for that latch. After completing the above calculation for each latch, each latch number and its corresponding preload reference adjustment amount are archived in a one-to-one correspondence to obtain the updated preload reference value for locally weak locations.

[0042] Understandably, the process of repeatedly testing and tightening by feel by assembly workers, which originally relied on assembly workers, has been transformed into a quantitative calculation process based on the relationship between the tightening force data and the geometric comparison. This process can provide differentiated pre-tightening benchmark update values ​​for each lock in areas with uneven sealing.

[0043] Step S104: Based on the pre-tightening reference update value, the gradient descent algorithm is used to iteratively optimize the overall clamping force distribution of the frame seal until the coefficient of variation of the clamping force in the circumferential direction is lower than the set limit, which is determined to be a uniform state, and the optimized clamping force distribution parameters are obtained.

[0044] Based on the pre-tightening reference update value, for several latches arranged circumferentially on the cover plate frame, the pre-tightening displacement increment corresponding to each latch number is loaded as the initial value for iteration. The pre-tightening displacement values ​​of all latches are used to form the pre-tightening displacement vector to be optimized. The sum of squares of the deviations of the clamping force values ​​output by each diaphragm pressure-sensitive unit from the overall circumferential mean is used as the objective function to complete the loading of the initial value for iteration and the construction of the objective function. Based on the pre-tightening displacement vector and the objective function, the gradient descent algorithm is used to iteratively optimize the overall clamping force distribution of the frame seal. In each iteration, a small perturbation is applied to each component of the pre-tightening displacement vector. The partial derivative values ​​of the objective function with respect to each component are obtained from the clamping force values ​​reread from the diaphragm pressure-sensitive unit. The gradient direction of the current iteration is synthesized, and the pre-tightening displacement vector is updated in the opposite direction of the gradient according to the preset step size parameter to obtain the pre-tightening displacement vector and the corresponding clamping force resampling data after this iteration. Based on the re-sampling data of the clamping force, the ratio of the standard deviation to the arithmetic mean of the clamping force value is calculated along the circumference of the frame to obtain the coefficient of variation of the clamping force in the circumference direction. If the coefficient of variation is lower than the set limit, the iteration is terminated and it is determined that the frame seal has reached a uniform state. The pre-tightening displacement vector and the corresponding clamping force value at this time are obtained as the optimized clamping force distribution parameters.

[0045] In one possible implementation, the initial value loading of the iteration is performed simultaneously for several latches arranged circumferentially along the upper edge of the cover plate frame.

[0046] Specifically, the preload displacement increment corresponding to each latch number is loaded into the main controller memory and arranged into a one-dimensional array according to the latch number order, serving as the initial value of the preload displacement vector. The dimension of the preload displacement vector is equal to the number of latches, and the i-th component ui of the vector represents the current preload displacement value of the i-th latch.

[0047] It should be noted that the objective function characterizes the degree of unevenness in the distribution of the clamping force of the frame seal.

[0048] Specifically, for the clamping force value pj output by the thin-film pressure-sensitive unit at each sampling point in the circumferential direction of the frame, the arithmetic mean pm of the clamping force values ​​at all sampling points is first calculated. Then, the squared difference between pj and pm is calculated for each sampling point and summed to obtain the sum of squared deviations J=Σ(pj-pm). 2 Here, j represents the sampling point index. The smaller the value of J, the closer the compaction force distribution is to uniformity. The core idea of ​​the gradient descent algorithm is to iteratively adjust the independent variable J, which is a target function with the pre-tightening displacement vector as its independent variable, along the direction of the fastest descent of the function at the current value of the independent variable, so that the value of the target function gradually decreases and eventually approaches the minimum value. The direction of the fastest descent is given by the opposite direction of the gradient vector formed by the partial derivatives of the target function with respect to each component of the independent variable.

[0049] Specifically, regarding the calculation of partial derivatives, since there is no explicit analytical expression between the clamping force value pj and the preload displacement vector, this implementation method uses the numerical difference method to calculate the partial derivatives component by component. In each iteration, for the i-th component ui of the preload displacement vector, the remaining components are kept unchanged, and a small disturbance Δ is temporarily added to ui to obtain the disturbed preload displacement vector. The main controller applies the disturbed preload displacement through the locking electric actuator, and the diaphragm pressure sensitive unit rereads the clamping force values ​​at each sampling point. The disturbed objective function value Jiplus is obtained according to the objective function formula. Then, ui is temporarily reduced by Δ and the above reading process is repeated to obtain Jiminus. The difference between Jiplus and Jiminus is divided by twice Δ to obtain the partial derivative of the objective function with respect to ui.

[0050] Preferably, the disturbance amount Δ is twice the minimum resolvable displacement of the locking pre-tightening mechanism, balancing differential accuracy and mechanism response capability. Further, the above differential process is performed on each component of the pre-tightening displacement vector, and the partial derivatives obtained for each component are arranged in component order to form the gradient direction of the current iteration. The pre-tightening displacement vector is updated in the opposite direction of the gradient according to a preset step size parameter α, with the update rule being... , where gi represents the i-th component of the gradient direction, α represents the step size parameter, and uinew represents the updated value of the i-th latch pre-tightening displacement.

[0051] For example, in one embodiment, the preset step size parameter α is taken as 0.05 times the full scale of the preload displacement. The main controller synchronously executes the feed through each locking electric actuator according to the updated preload displacement vector, and then the membrane pressure sensitive unit collects the clamping force re-sampling data after this iteration.

[0052] It is understood that the coefficient of variation represents a normalized measure of the dispersion of a set of numerical values ​​relative to their overall level.

[0053] Specifically, for the re-sampling data of the clamping force, the clamping force value is read at each sampling point along the circumference of the frame. The standard deviation σ is first calculated, and then the arithmetic mean μ is calculated. The coefficient of variation CV = σ / μ. The smaller the CV value, the smaller the fluctuation of the clamping force value around the overall mean, and the more uniform the circumferential distribution. Further, the set limit is set to 0.08 in one embodiment. If the coefficient of variation CV after this iteration is less than 0.08, the iteration is terminated, and it is determined that the frame seal has reached a uniform state; if CV is still greater than 0.08, the pre-tightening displacement vector of this round is retained as the initial value of the next iteration, and the differential gradient calculation and displacement update process continues until CV meets the convergence condition. After the iteration terminates, the pre-tightening displacement vector and the corresponding clamping force value at this time are obtained as optimization clamping force distribution parameters. This can adjust the clamping force distribution in the circumferential direction of the frame from the original state of coexistence of overpressure zone and loose zone due to the self-weight of the cover plate to a uniform pressure state with controllable circumferential fluctuation, providing a stable sealing baseline for subsequent leakage sound pressure re-sampling.

[0054] Step S105: After obtaining the optimized clamping force distribution parameters, re-collect the leakage sound pressure data at the gaps around the cover plate to determine the attenuation increment of noise radiation.

[0055] Based on the determination that the frame seal has reached a uniform state, the optimized clamping force distribution parameters are obtained. The sound pressure pickup units deployed around the gaps of the cover plate are then used to re-collect leakage sound pressure data under the same steady-state operating conditions of the air conditioner terminal fan, resulting in optimized leakage sound pressure data. Based on the optimized leakage sound pressure data and the aforementioned transmitted sound pressure data at the gaps, the difference between the initial sound pressure amplitude and the optimized sound pressure amplitude is calculated for each sampling point along the circumference of the frame. The arithmetic mean of the differences for all sampling points is then used to determine the noise radiation attenuation increment.

[0056] In one possible implementation, the optimized clamping force distribution parameters include the pre-tightening displacement vector at the end of the iteration and the clamping force value corresponding to each thin-film pressure-sensitive unit.

[0057] Specifically, the main controller reads the optimized clamping force distribution parameters from the memory and saves them as the current frame sealing state. Furthermore, for the sound pressure pickup units densely arranged circumferentially around the gaps around the cover plate, the air conditioner terminal fan is adjusted to the same steady-state operating condition as in the initial acquisition stage, that is, the fan speed, air outlet temperature, and air valve opening are all kept the same as before, and the optimized leakage sound pressure data is re-acquired.

[0058] It should be noted that the calculation of the attenuation increment is based on the position difference of the same sampling point.

[0059] Specifically, for each sampling point along the circumference of the cover plate edge, the initial sound pressure amplitude pinit for that point is retrieved from the storage archive, and the optimized sound pressure amplitude popt for the same point is taken from the optimized leakage sound pressure data. The difference is used to obtain the sound pressure attenuation value Δp = pinit - popt for that point. The arithmetic mean of Δp for all sampling points is then taken to obtain the noise radiation attenuation increment.

[0060] It is understandable that the larger the value of the attenuation increment, the more significant the decrease in the acoustic wave transmission intensity of the gap leakage channel after the frame pre-tightening adjustment, reflecting the suppression effect of the frame seal homogenization on the noise radiation around the cover plate.

[0061] Step S106: If the noise radiation attenuation increment is lower than the preset threshold, the deviation between the collected leakage sound pressure data and the initial clamping force distribution data is re-analyzed using the support vector machine algorithm to obtain a further locking pre-tightening reference adjustment scheme.

[0062] If the attenuation increment of the noise radiation is lower than a preset threshold, then for each sampling point along the circumferential direction at the gaps around the cover plate, the sound pressure amplitude at that point is read from the optimized leakage sound pressure data, and paired with the measured value of the clamping force at the same point recorded in the initial stage of the sealing force distribution data, to obtain a set of sound pressure-clamping force residual deviation pairs. Based on the set of sound pressure-clamping force residual deviation pairs, a support vector machine algorithm is used to perform regression fitting on the residual deviation pairs. A radial basis kernel is selected as the kernel function to handle the nonlinear correlation between the residual sound pressure and the initial clamping force. For each sampling point along the circumferential direction of the cover plate, the regression-predicted residual leakage tendency value is output. The residual leakage tendency value characterizes the estimated leakage intensity that still exists at that point. Sampling points with residual leakage tendency values ​​higher than a preset limit are taken as residual weak points, and a set of circumferential locations of residual weak points is obtained. Based on the set of circumferential locations of the remaining weak points, a corresponding latch is matched for each remaining weak point according to the principle of circumferential proximity. The current pre-tightening displacement value of the latch is retrieved from the optimized clamping force distribution parameters. The pre-tightening displacement increment is calculated and added according to the residual leakage tendency value based on the preset correspondence between the compression amount of the seal and the pre-tightening displacement of the latch. Each latch is aligned and archived to form a further latch pre-tightening reference adjustment scheme.

[0063] In one possible implementation, the preset threshold is determined during the calibration phase by the difference in noise radiation between the original sealed state of the cover plate and the fully sealed state of the ceiling.

[0064] Preferably, the preset threshold is 3 dB. When the attenuation increment of the noise radiation is less than 3 dB, it indicates that the first pre-tightening adjustment failed to suppress the gap leakage to a level close to that of the ceiling, triggering the residual deviation reanalysis process.

[0065] It should be noted that the set of residual deviations between sound pressure and clamping force is formed based on the alignment of the same sampling point.

[0066] Specifically, for each sampling point along the circumferential direction at the gaps around the cover plate, the sound pressure amplitude at that point is retrieved from the archived optimized leakage sound pressure data, and the measured value of the clamping force at the same point is retrieved from the initial stage archived sealing force distribution data. The two are written into the same record entry according to the sampling point number. All record entries are summarized to form the set of sound pressure-clamping force residual deviation pairs.

[0067] Specifically, the support vector machine (SVM) algorithm used in the aforementioned stage was used for binary classification, outputting the directed distance of the hyperplane as the classification decision value; the SVM algorithm used in this stage is used for regression fitting, outputting continuous predicted values. The training objectives of the two are different, but both follow the kernel function mapping idea of ​​SVM. In the regression fitting stage, the measured values ​​of the clamping force of each record in the set of sound pressure-compression force residual deviation pairs are used as input components, and the corresponding optimized sound pressure amplitude is used as the output target, forming training sample pairs. The objective function of the regression model is represented by obtaining a set of coefficients and biases such that the deviation between the predicted values ​​and the target values ​​of all training sample pairs falls within a preset tolerance band ε, while controlling the model complexity. The kernel function used is the radial basis function kernel K(xm,xn)=exp(-γ·‖xm-xn‖). 2 ), where xm and xn represent the measured values ​​of the clamping force at two sampling points, and γ represents the kernel function bandwidth parameter. After regression solving, the measured values ​​of the clamping force are input for each sampling point along the circumference of the cover plate, and the regression model outputs the predicted sound pressure amplitude at that point. The predicted sound pressure amplitude serves as the residual leakage tendency value at that point, representing an estimate of the leakage intensity that still exists at that point under the current clamping force level. Further, in one embodiment, the preset limit is taken as the median of the residual leakage tendency values ​​of all sampling points along the circumference of the cover plate plus one standard deviation. Sampling points with residual leakage tendency values ​​higher than the preset limit are marked as residual weak points, and the circumferential coordinates of each point are summarized into a set of circumferential positions of the residual weak points.

[0068] It is understandable that the circumferential coordinates of the remaining weak points are not necessarily perfectly aligned with the circumferential installation coordinates of the latch.

[0069] Specifically, for each residual weak point, the circumferential arc distance between each latch and that weak point is calculated, and the latch with the smallest arc distance is selected as the corresponding latch for that weak point, thus completing the circumferential proximity matching. Then, the current pre-tightening displacement value of that latch is retrieved from the optimized clamping force distribution parameters as the baseline for this round of additional adjustments. Further, based on the aforementioned preset correspondence between the seal compression and the latch pre-tightening displacement, the required additional seal compression is determined according to the strength of the residual leakage tendency value, and then the additional pre-tightening displacement increment required for the latch pre-tightening mechanism is obtained. Each latch number and its corresponding additional pre-tightening displacement increment are archived in a one-to-one correspondence to form a further latch pre-tightening baseline adjustment scheme. In another implementation, the compression increment Δc = L / 100 is first calculated, where L is the residual leakage tendency value, and Δc is in mm, and then... Where k=0.05, Δc represents the additional compression required calculated based on the residual leakage tendency value, in millimeters, and Δd is the additional preload displacement increment.

[0070] It is understandable that this implementation method performs secondary positioning and compensation for residual weak positions that have not reached the target suppression level after a first pre-tightening adjustment, and can bring the local leakage channels that remain after the original homogenization process into the scope of fine adjustment.

[0071] Step S107: Based on the further locking pre-tightening reference adjustment scheme, obtain the updated data of the clamping force distribution, and determine the overall consistency improvement level of the acoustic wave transmission coefficient in the cover plate area.

[0072] According to the further locking pre-tightening reference adjustment scheme, for each locking buckle arranged circumferentially on the cover plate frame, the feed is synchronously applied according to the archived additional pre-tightening displacement increment. The new round of clamping force values ​​are read from the sampling points of the thin-film pressure sensitive unit to obtain the clamping force distribution update data. According to the clamping force distribution update data, the sound pressure pickup units deployed in the central area and surrounding gaps of the cover plate are re-collected under the steady-state operating conditions of the air conditioner terminal fan, where the fan speed, outlet air temperature, and air valve opening are kept the same as before. For each sampling point, the sound pressure amplitude at the gap is compared with the sound pressure amplitude at the sampling point in the central area according to the spatial correspondence table to obtain the sound wave transmission coefficient at that point. The spatial correspondence table is established by radial projection of the cover plate geometric model. For example, the input gap point coordinates x1, y1, output the corresponding central area point coordinates x2, y2. The sound wave transmission coefficient distribution sequence is formed by summarizing along the circumference of the cover plate. Based on the sound wave transmission coefficient distribution sequence, the ratio of the standard deviation to the arithmetic mean of the sound wave transmission coefficient is calculated along the circumference of the cover plate to obtain the dispersion value of the sound wave transmission coefficient in the cover plate area. The dispersion value is then compared with the dispersion value of the sound wave transmission coefficient that was archived before this round of adjustment to determine the overall consistency improvement level of the sound wave transmission coefficient in the cover plate area.

[0073] In one possible implementation, the further locking preload reference adjustment scheme is stored in the main controller in the form of an archive table of locking numbers and corresponding additional preload displacement increments.

[0074] Specifically, the main controller retrieves the additional pre-tightening displacement increments item by item according to the archive table, and applies feed synchronously to each latch arranged circumferentially along the upper edge of the cover plate frame through its respective electric actuator.

[0075] Preferably, the synchronous feed adopts a pulse triggering method, and all locking electric actuators start simultaneously along the axis under the synchronous pulse of the main controller, so as to avoid the instantaneous force imbalance of the frame caused by feeding one by one.

[0076] It should be noted that after the feeding action is completed, the thin-film pressure-sensitive unit rereads the clamping force value at each sampling point, arranges them according to the sampling point number to form a new round of circumferential clamping force sequence, and archives it into the memory as clamping force distribution update data. Furthermore, for the sound pressure pickup units deployed in the central area and surrounding gaps of the cover plate, the air conditioning terminal fan is adjusted to the same steady-state operating condition as in the previous stage, including fan speed, outlet air temperature, and damper opening, all remaining the same as before. Sound pressure values ​​in the central area and gaps are collected synchronously and archived according to the sampling point number as a central area sound pressure sequence and a gap sound pressure sequence, respectively.

[0077] Specifically, regarding the calculation of the sound wave transmission coefficient, the sound wave transmission coefficient characterizes the proportion of energy attenuation of the sound wave after it passes through the cover plate and gaps from the back cavity side of the cover plate to the front side of the cover plate. In this embodiment, the transmission in the central area of ​​the cover plate is mainly controlled by the sound insulation performance of the composite material of the cover plate body, while the transmission at the gaps is simultaneously affected by the superposition of leakage from the cover plate body and the frame gaps. The ratio of the sound pressure amplitudes of the two can reflect the degree of disturbance of the gap leakage to the overall sound insulation consistency of the cover plate.

[0078] Specifically, for each sampling point in the gaps around the cover plate, the corresponding sampling point in the central area is determined according to the spatial correspondence table. Then, the sound pressure amplitude pg at the gap is taken from the sound pressure sequence at the gap, and the sound pressure amplitude pc in the central area is taken from the sound pressure sequence in the central area. The ratio τ = pg / pc is used as the sound wave transmission coefficient of that point. Further, the sound wave transmission coefficients τ of each point are arranged in order of sampling point number along the circumference of the cover plate frame, and summarized to form a sound wave transmission coefficient distribution sequence. The closer the value of τ is to 1, the closer the transmission level of that point is to the transmission level of the central area, and the smaller the disturbance of gap leakage to the sound insulation consistency.

[0079] It is understandable that the calculation of the discreteness of the sound wave transmission coefficient follows the statistical meaning of the coefficient of variation.

[0080] Specifically, for the aforementioned sound wave transmission coefficient distribution sequence, first calculate the arithmetic mean τm of all τ values ​​in the sequence, then calculate the standard deviation στ, and finally calculate the numerical value of the dispersion. The smaller the CVτ value, the smaller the fluctuation of the sound wave transmission coefficient in the circumferential direction of the cover plate, and the better the sound insulation consistency of the cover plate area. Furthermore, the determination of the overall consistency improvement level is based on the comparison of the two rounds of dispersion values.

[0081] Specifically, the previously archived acoustic wave transmission coefficient dispersion value CVτprev is retrieved from memory. The difference is obtained by subtracting the newly archived CVτ from CVτprev. The sign and magnitude of the difference represent the improvement level of the overall consistency of the acoustic wave transmission coefficient in the cover plate area brought about by this round of adjustment. A positive difference and a larger value indicate a more significant improvement in the overall consistency of the acoustic wave transmission coefficient in the cover plate area brought about by this round of secondary pre-tightening benchmark adjustment.

[0082] Step S108: Extract the noise attenuation index from the overall consistency improvement level and determine the noise suppression adjustment scheme for the air conditioning terminal vent.

[0083] Based on the overall consistency improvement level of the sound wave transmission coefficient in the cover plate area, the arithmetic mean of the sound pressure amplitude of all sampling points is taken from the sound pressure sequence of the gap archived before and after this round of adjustment. The arithmetic mean before this round of adjustment is subtracted from the arithmetic mean after this round of adjustment to obtain the noise attenuation index. According to the noise attenuation index, distributed sound pressure pickup units are deployed in a preset grid in the area below the air conditioner terminal vent. Under the same steady-state operating conditions of the fan, the radiated sound pressure data of each grid node in the area below is collected. The radiated sound pressure amplitude is calculated node by node along the grid to obtain the radiated sound pressure distribution in the area below. According to the radiated sound pressure distribution in the area below, the arithmetic mean and maximum value of the radiated sound pressure amplitude are calculated along the grid nodes. If the arithmetic mean is lower than the preset suppression threshold and the maximum value is lower than the preset peak threshold, the pre-tightening displacement vector in the optimized clamping force distribution parameters and the corresponding latch pre-tightening displacement increment are merged and archived to determine the noise suppression adjustment scheme of the air conditioner terminal vent.

[0084] In one possible implementation, the noise attenuation index is extracted based on two sets of sound pressure sequences at the gaps archived before and after this round of adjustment.

[0085] Specifically, the sound pressure sequence at the gap archived before and after this round of adjustment are retrieved from the memory. The arithmetic mean of the sound pressure amplitude at all sampling points is calculated for each set of sequences, and denoted as Abefore and Aafter, respectively. The difference between Abefore and Aafter is used as the noise attenuation index; the larger the difference, the more significant the suppression effect of this round of adjustment on the leakage sound pressure at the gap.

[0086] It should be noted that the distributed sound pressure pickup units in the area below the air conditioner terminal vents are deployed according to a preset grid pattern.

[0087] Specifically, the lower area refers to the horizontal projection area extending along the floor surface directly below the air conditioner terminal vent. The grid is divided into orthogonal rectangles, and the row and column spacings are scaled proportionally to the length of the long side of the vent.

[0088] Preferably, the row and column spacing of the grid is 0.5 times the length of the longer side of the air outlet. Grid nodes are located at the intersection of the row and column spacing, and an electret microphone is installed at each node, with the microphone diaphragm facing upwards and perpendicular to the floor. Furthermore, the microphones are led out via independent shielded cables to the data acquisition host at the edge of the lower area, and all microphones collect data in parallel under the unified clock synchronization of the main controller. Under steady-state operating conditions consistent with the previous stages, including fan speed, outlet air temperature, and damper opening, radiated sound pressure data are collected for each grid node in the lower area.

[0089] Specifically, for each grid node, the sound pressure time series within the steady-state segment of that node is read from the storage of the acquisition host. The root mean square value of that time series is obtained according to the commonly used acoustic processing method as the radiated sound pressure amplitude of that node. The radiated sound pressure amplitude of each node is archived according to the grid row and column number, and summarized to form the radiated sound pressure distribution of the area below.

[0090] It is understandable that the radiated sound pressure distribution in the lower region simultaneously exhibits both average level and local peak characteristics.

[0091] Specifically, firstly, the arithmetic mean Bmean of the radiated sound pressure amplitude of all grid nodes is calculated to represent the overall noise level of the area below; then, the maximum value Bmax of the radiated sound pressure amplitude of all grid nodes is calculated to represent the noise level at the loudest local location in the area below. Further, in one embodiment, the preset suppression threshold is 35 dB, corresponding to the conventional upper limit of background noise in an office space; the preset peak threshold is 40 dB in one embodiment, corresponding to the local allowable peak value in an office space. If Bmean is lower than 35 dB and Bmax is lower than 40 dB, it is determined that the overall radiated sound pressure level and local peak value of the area below after this round of adjustment both meet the requirements for a quiet office environment. Further, when both thresholds are satisfied, the pre-tightening displacement vector in the optimized clamping force distribution parameters and the corresponding pre-tightening displacement increment in the further pre-tightening reference adjustment scheme are merged item by item according to the pre-tightening number. The merged result is the final pre-tightening displacement value of each pre-tightening element. Each pre-tightening element is aligned and archived to form the final air conditioning terminal vent noise suppression adjustment scheme.

[0092] It is understandable that this implementation method extends the pre-tightening adjustment effect of the frame from the local sound pressure characterization at the gaps around the cover plate to the overall radiated sound pressure characterization of the area under the air conditioner terminal vent. This can realize the local results of sealing uniformity as the noise suppression results of the entire area under use, and provide a feasible pre-tightening benchmark final configuration for places with high requirements for quiet environment, such as high-end office buildings, hotel rooms and residential bedrooms.

[0093] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for suppressing and adjusting noise at the air outlet of an air conditioning terminal, characterized in that, The method includes: acquiring transmitted sound pressure data and sealing force distribution data from the central area and surrounding gaps of the cover plate to obtain an acoustic consistency index and a frame pressure uniformity index for the cover plate area; based on the acoustic consistency index and the frame pressure uniformity index, using a support vector machine algorithm to analyze the correlation between the transmitted sound pressure data and the sealing force distribution data to determine the potential location distribution of gap leakage channels; if the potential location distribution range exceeds a preset threshold, using a least squares algorithm to fit the degree of insufficient pressure corresponding to each leakage point, calculating the pre-tightening reference adjustment amount for each latch, and generating a pre-tightening reference update value; and based on the pre-tightening reference update value, using a gradient descent algorithm to iteratively optimize the frame seal. The overall clamping force distribution is adjusted until the coefficient of variation of the clamping force in the circumferential direction is lower than a set limit, at which point a uniform state is determined, and optimized clamping force distribution parameters are obtained. Leakage sound pressure data at the gaps around the cover plate is re-collected to determine the noise radiation attenuation increment. If the attenuation increment is lower than a preset threshold, the deviation between the collected leakage sound pressure data and the initial clamping force distribution data is re-analyzed using a support vector machine algorithm to obtain a latch pre-tightening benchmark adjustment scheme. Based on the latch pre-tightening benchmark adjustment scheme, updated clamping force distribution data is obtained, and the overall consistency improvement level of the sound wave transmission coefficient in the cover plate area is determined. Noise attenuation indicators are extracted from the overall consistency improvement level to determine the air conditioning terminal vent noise suppression adjustment scheme.

2. The method for suppressing and adjusting noise at the air outlet of an air conditioner terminal as described in claim 1, characterized in that, The process of acquiring transmitted sound pressure data and sealing force distribution data from the central area and surrounding gaps of the cover plate to obtain the acoustic consistency index and the pressure uniformity index of the cover plate area includes: arranging sound pressure pickup units in the central area of ​​the cover plate at a preset grid spacing; arranging another set of sound pressure pickup units circumferentially at the gaps around the cover plate's perimeter; embedding a thin-film pressure-sensitive unit circumferentially in the sealing element of the perimeter; normalizing the dispersion of the difference between the two sets of sound pressure amplitudes at each sampling point to obtain the acoustic consistency index; and normalizing the offset amplitude of the average pressure force segment by segment along the perimeter of the perimeter to obtain the pressure uniformity index of the perimeter.

3. The method for suppressing and adjusting noise at the air outlet of an air conditioner terminal as described in claim 1, characterized in that, The step of analyzing the correlation between transmitted sound pressure data and sealing force distribution data using a support vector machine algorithm based on the acoustic consistency index and the uniformity of pressure on the frame to determine the potential location distribution of gap leakage channels includes: extracting acoustic deviation values, pressure offset amplitudes, and frequency band deviation values ​​from each sampling point around the cover plate to form feature vectors; assigning positive and negative labels to confirmed leakage points and intact sealing points respectively; training a support vector machine with radial basis kernels to separate the hyperplane; using the directed distance from each feature vector to the hyperplane as the decision value; merging adjacent points with decision values ​​higher than a preset threshold into continuous candidate segments, and mapping them to the four sides of the cover plate to determine the potential location distribution.

4. The method for suppressing and adjusting noise at the air outlet of an air conditioner terminal as described in claim 1, characterized in that, If the potential location distribution range exceeds a preset threshold, the degree of insufficient clamping force corresponding to each leakage point is fitted using a least squares algorithm, the pre-tightening reference adjustment amount of each latch is calculated, and a pre-tightening reference update value is generated. This includes: reading the absolute difference between the measured clamping force value and the overall mean value of the circumferential side of the frame for each candidate leakage point within the potential location distribution to obtain the degree of insufficient clamping force; performing polynomial fitting with the circumferential installation position of the latch as the independent variable to obtain a continuous fitting curve; taking the corresponding function value for the circumferential coordinate of each latch; and converting it into the pre-tightening reference adjustment amount according to the preset correspondence between the compression amount of the seal and the pre-tightening displacement of the latch.

5. The method for suppressing and adjusting noise at the air outlet of an air conditioner terminal as described in claim 1, characterized in that, The process involves updating the pre-tightening reference value and iteratively optimizing the overall clamping force distribution of the frame seal using a gradient descent algorithm until the coefficient of variation of the clamping force in the circumferential direction is lower than a set limit, indicating that a uniform state has been reached. The optimized clamping force distribution parameters are then obtained. This includes: using the pre-tightening displacement increment corresponding to each latch on the cover frame as the initial iteration value; the pre-tightening displacement values ​​of all latches forming a pre-tightening displacement vector; using the sum of squares of the deviations of the clamping force output by the membrane pressure-sensitive unit relative to the overall circumferential mean as the objective function; applying perturbation to each component of the pre-tightening displacement vector in each iteration; rereading the clamping force from the membrane pressure-sensitive unit to obtain the partial derivatives of the objective function with respect to each component; synthesizing the gradient direction; updating the pre-tightening displacement vector along the gradient in the opposite direction with a preset step size; obtaining the coefficient of variation along the circumferential direction of the frame by the ratio of the standard deviation of the clamping force to the arithmetic mean; terminating the iteration when the coefficient of variation is lower than a set limit; and taking the pre-tightening displacement vector and the corresponding clamping force at that moment as the optimized clamping force distribution parameters.

6. The method for suppressing and adjusting noise at the air outlet of an air conditioner terminal as described in claim 1, characterized in that, The process of re-collecting leakage sound pressure data at the gaps around the cover plate to determine the attenuation increment of noise radiation includes: re-collecting leakage sound pressure data at the gaps around the cover plate using sound pressure pickup units under the same steady-state operating conditions of the air conditioning terminal fan to obtain the optimized leakage sound pressure amplitude; subtracting the optimized sound pressure amplitude from the initial sound pressure amplitude for each sampling point along the circumferential direction of the frame to obtain the difference; and taking the arithmetic mean of the differences for all sampling points to determine the attenuation increment of noise radiation.

7. The method for suppressing and adjusting noise at the air outlet of an air conditioner terminal as described in claim 1, characterized in that, If the attenuation increment is lower than a preset threshold, the deviation between the collected leakage sound pressure data and the initial clamping force distribution data is re-analyzed using a support vector machine algorithm to obtain a lock pre-tightening reference adjustment scheme. This includes: reading the sound pressure amplitude from the collected leakage sound pressure data at each sampling point along the circumferential direction of the gap around the cover plate, and pairing it with the measured clamping force values ​​at the same points in the initial clamping force distribution data to obtain a set of residual sound pressure clamping force deviation pairs; using radial basis kernel support vector machine regression to fit the residual deviation pair set to output a residual leakage tendency value, and using points higher than a preset limit as residual weak points; matching a corresponding lock for each residual weak point according to the circumferential proximity principle, and calculating the additional pre-tightening displacement increment according to the residual leakage tendency value based on the preset correspondence between the seal compression amount and the lock pre-tightening displacement, and archiving each lock alignment to form the lock pre-tightening reference adjustment scheme.

8. The method for suppressing and adjusting noise at the air outlet of an air conditioner terminal as described in claim 1, characterized in that, The step of obtaining the updated data of the clamping force distribution according to the adjustment scheme of the pre-tightening reference of the latch includes: synchronously applying feed to each latch arranged circumferentially along the upper edge of the cover plate according to the archived additional pre-tightening displacement increment, and reading the new round of clamping force values ​​from the sampling points of the membrane pressure sensitive unit to obtain the updated data of the clamping force distribution.

9. The method for suppressing and adjusting noise at the air outlet of an air conditioner terminal as described in claim 1, characterized in that, The step of extracting the noise attenuation index from the overall consistency improvement level and determining the noise suppression adjustment scheme for the air conditioning terminal vent includes: taking the arithmetic mean of the sound pressure amplitude from the sound pressure sequences archived before and after this round of adjustment, and subtracting the arithmetic mean after adjustment from the arithmetic mean before adjustment to obtain the noise attenuation index; collecting radiated sound pressure data by deploying distributed sound pressure pickup units in a preset grid in the area below the air conditioning terminal vent, taking the arithmetic mean and maximum value of the radiated sound pressure amplitude along the grid nodes, and when the arithmetic mean is lower than a preset suppression threshold and the maximum value is lower than a preset peak threshold, archiving the pre-tightening displacement vector and the corresponding locking pre-tightening displacement increment in the optimized clamping force distribution parameters to determine the noise suppression adjustment scheme for the air conditioning terminal vent.

Images