Strip line medium phase shift system for base station antenna
By constructing a closed-loop architecture and performing topology similarity analysis, assembly air gaps and material defects in base station antenna production are accurately identified, solving the problem of misjudgment and scrapping in existing technologies, reducing production costs and optimizing resource allocation.
Patent Information
- Application Number
- CN202610138959.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-02-01
- Publication Date
- 2026-05-15
AI Technical Summary
Existing technologies cannot effectively distinguish between assembly air gaps and material defects in base station antenna production, leading to misjudgments and scrapping, high production costs, and the inability to provide quantitative maintenance guidance.
A closed-loop architecture consisting of a main control unit, a data acquisition module, a virtual simulation reconstruction module, a dual-track differential extraction module, and a coupled verification and decision module is adopted. By constructing a lossless ideal transmission line model and an equivalent model of air gap perturbation, combined with topology similarity analysis, the cause of the fault is accurately identified and production control instructions are generated.
It enables accurate identification of assembly air gaps and material defects, reduces production costs, improves fault diagnosis resolution, avoids misjudgment of repairable problems, and optimizes the allocation of production resources.
Smart Images
Figure CN122051615A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of mobile communication and antenna technology, specifically to a stripline dielectric phase-shifting system for base station antennas. Background Technology
[0002] In the large-scale manufacturing of base station antennas, the stripline dielectric phase shifter is a core control component. Its electromagnetic performance is closely related to its mechanical assembly state, and strict parameter testing and quality control are required during the production process.
[0003] Existing testing solutions generally adopt a static threshold-based judgment architecture, which simply uses a vector network analyzer to collect S-parameters and judges whether a product is qualified by comparing whether the return loss or transmission phase exceeds a fixed limit. Although this solution can screen out products with abnormal performance to a certain extent, it cannot effectively capture transient faults during the movement of dielectric wafers due to the lack of in-depth correlation analysis of dynamic multidimensional data. More importantly, this type of method is difficult to analyze the fault mechanism in depth and cannot distinguish whether the root cause of performance degradation is a repairable structural assembly air gap or an irreversible intrinsic material defect. This lack of identification accuracy leads to the misjudgment of components that only require process repair as scrap products, and it cannot provide quantitative maintenance guidance, resulting in high production costs and difficulties in improving yield.
[0004] Therefore, how to accurately identify the physical causes of failures, effectively distinguish between assembly air gaps and material defects, and thereby achieve intelligent closed-loop control and diversion, has become an urgent technical problem to be solved. Summary of the Invention
[0005] To solve the above-mentioned technical problems, the present invention provides a stripline dielectric phase-shifting system for base station antennas. Specifically, the technical solution of the present invention includes:
[0006] The main control unit is connected to a data acquisition module, a virtual simulation reconstruction module, a dual-track differential extraction module, and a coupled verification and decision module.
[0007] The data acquisition module is used to acquire the vector network parameter stream and mechanical displacement feedback data of the phase shifter;
[0008] The virtual simulation reconstruction module is used to construct a lossless ideal transmission line model based on ideal geometric parameters, and to generate theoretical S-parameters with defects based on the air gap perturbation equivalent model.
[0009] The dual-track differential extraction module is used to calculate the theoretical residual feature vector and the actual residual feature vector based on the lossless ideal transmission line model, theoretical defective S-parameters and vector network parameter flow;
[0010] The coupled verification and decision module is used to calculate the topological similarity between the theoretical residual feature vector and the actual residual feature vector, and to generate fault attribution results and corresponding production control instructions based on the topological similarity.
[0011] Preferably, the process by which the data acquisition module acquires the vector network parameter stream and mechanical displacement feedback data of the phase shifter includes:
[0012] The phase shifter is scanned in real time within a preset frequency band to extract the return loss data and transmission phase data of the S-parameters, and the return loss data and transmission phase data are marked as vector network parameter streams.
[0013] The current physical position coordinates of the medium sheet are obtained by a position sensor, and these coordinates are marked as mechanical displacement feedback data.
[0014] Preferably, the process of constructing a lossless ideal transmission line model by the virtual simulation reconstruction module includes:
[0015] Obtain the stripline cavity height, conductor width, and dielectric sheet design dimensions, and mark them as ideal geometric parameters;
[0016] Using the odd-even mode analysis method, an ideal phase curve and a zero-reflection standing wave curve that vary with the current physical location coordinates are established based on ideal geometric parameters. The combination of the ideal phase curve and the zero-reflection standing wave curve is labeled as the lossless ideal transmission line model.
[0017] Preferably, the process by which the virtual simulation reconstruction module generates theoretical S-parameters containing defects based on the air gap perturbation equivalent model includes:
[0018] A pre-defined air gap perturbation equivalent model is introduced, and the contact gap between the dielectric sheet and the conductor strip is equivalent to the perturbation of the transmission line distribution parameters.
[0019] In a lossless ideal transmission line model, a preset air gap value is forcibly set. Inverse simulation is performed in combination with the perturbation of the transmission line distribution parameters to calculate the S-parameter variation data of the lossless ideal transmission line model under the air gap value. The S-parameter variation data is marked as theoretical S-parameters with defects.
[0020] Preferably, the process by which the dual-track difference extraction module calculates the theoretical residual feature vector and the actual residual feature vector includes:
[0021] Subtracting the baseline data of the lossless ideal transmission line model from the vector network parameter stream yields a realistic residual feature vector that includes environmental noise and manufacturing tolerances.
[0022] By subtracting the baseline data of the lossless ideal transmission line model from the theoretical defective S-parameters, the signal distortion features caused only by the air gap value are extracted, and the signal distortion features are marked as theoretical residual feature vectors.
[0023] Preferably, the process by which the coupled verification decision module calculates the topological similarity between the theoretical residual feature vector and the actual residual feature vector includes:
[0024] Calculate the cross-correlation function or cosine similarity between the actual residual eigenvector and the theoretical residual eigenvector in the frequency domain;
[0025] The numerical values of cross-correlation function or cosine similarity are labeled as topological similarity.
[0026] Preferably, the process by which the coupled verification and decision module generates fault attribution results and corresponding production control instructions based on topological similarity includes:
[0027] A first similarity threshold is preset. If the topological similarity is greater than the first similarity threshold, the cause of the anomaly is determined to be an assembly air gap fault.
[0028] Generate a production control command that includes a secondary pressurization request, and send the production control command to the control terminal of the automated crimping equipment.
[0029] Preferably, the process by which the coupled verification and decision module generates fault attribution results and corresponding production control instructions based on topological similarity also includes:
[0030] A second similarity threshold and a total error threshold are preset, wherein the first similarity threshold is greater than the second similarity threshold;
[0031] If the topological similarity is less than the second similarity threshold and the magnitude of the actual residual feature vector is greater than the total error threshold, then the cause of the anomaly is determined to be a defect in the medium material.
[0032] Generate production control instructions containing defective product diversion markers and send these instructions to the control terminal of the sorting system.
[0033] Compared with the prior art, the present invention has the following beneficial effects:
[0034] 1. This invention overcomes the technical bottleneck of traditional threshold judgment methods, which cannot distinguish between structural assembly air gaps and intrinsic material defects, by constructing a closed-loop architecture of physical measurement, virtual reconstruction, and differential comparison. The system no longer simply judges whether the product performance is qualified, but actively simulates faults in virtual space and compares them with reality through synthetic analysis, thereby intelligently identifying the physical causes of the faults. This design significantly improves the resolution of fault diagnosis, can accurately identify micron-level assembly air gaps, and avoids misjudging repairable assembly problems as material scrap, thereby significantly reducing production costs in the large-scale manufacturing scenario of base station antennas.
[0035] 2. This invention uniquely binds the vector network parameter stream with the mechanical displacement feedback data using hardware timestamp technology, generating a three-dimensional data packet containing frequency, position, and performance. This ensures the spatiotemporal consistency of the data and prevents measurement ambiguity caused by the movement of the dielectric sheet. By multidimensionally fusing electromagnetic parameters with mechanical position, it ensures that subsequent analysis is based on a full-dimensional dynamic data cube. This enables the system to capture transient faults that are only triggered at specific dielectric sheet positions, greatly improving the detection system's ability to capture dynamic assembly defects and effectively solving the industry problem that static testing cannot detect dynamic hidden dangers.
[0036] 3. This invention utilizes the odd-even mode analysis method to establish a lossless ideal transmission line model that varies with the current physical location coordinates, accurately simulating the impedance step effect. Based on the air gap perturbation equivalent model, the contact gap between the dielectric sheet and the conductor strip is equivalent to the perturbation of the transmission line distributed parameters, realizing the active generation of fault features. By using a dual-track differential extraction module to calculate the theoretical residual feature vector and the actual residual feature vector in parallel, the common-mode errors of the system, such as test cable insertion loss and fixture fixation reflection, are cleverly eliminated. At the same time, by constructing an extended-dimensional feature vector containing delinearized phase and normalized amplitude and calculating topological similarity, the essence of the air gap problem is identified with extremely high sensitivity, and the fault can be accurately determined even when the amplitude of the test signal itself fluctuates greatly.
[0037] 4. This invention achieves precise diversion of repairable assembly air gap faults and irreversible medium material defects by presetting different similarity thresholds and total error thresholds. For air gap faults, the system quantitatively calculates the target pressure holding parameters based on the projection coefficient and introduces a pressure safety fuse mechanism and process over-limit warning, solving the problem that qualitative judgment alone cannot accurately control the repair intensity, while avoiding the risk of positive feedback runaway in closed-loop control. For material defects, the combination logic of low similarity and high energy error avoids meaningless secondary pressurization of defective products, thereby optimizing the overall production resource allocation efficiency. Attached Figure Description
[0038] The present invention will be further explained below with reference to the accompanying drawings and embodiments:
[0039] Figure 1 This is a structural diagram of the system of the present invention. Detailed Implementation
[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.
[0041] Example 1:
[0042] Please see Figure 1A stripline dielectric phase-shifting system for base station antennas includes a main control unit, which is communicatively connected to a data acquisition module, a virtual simulation reconstruction module, a dual-track differential extraction module, and a coupling verification and decision module.
[0043] The data acquisition module is used to acquire the vector network parameter stream and mechanical displacement feedback data of the phase shifter;
[0044] The virtual simulation reconstruction module is used to construct a lossless ideal transmission line model based on ideal geometric parameters, and to generate theoretical S-parameters with defects based on the air gap perturbation equivalent model.
[0045] The dual-track differential extraction module is used to calculate the theoretical residual feature vector and the actual residual feature vector based on the lossless ideal transmission line model, theoretical defective S-parameters and vector network parameter flow;
[0046] The coupled verification and decision module is used to calculate the topological similarity between the theoretical residual feature vector and the actual residual feature vector, and to generate fault attribution results and corresponding production control instructions based on the topological similarity.
[0047] This embodiment details the hardware architecture and core logic flow of the system. The main control unit of the system adopts an industrial control computer equipped with a real-time operating system, which interacts with various functional modules through the PCIe bus for high-bandwidth data exchange. When the system starts the initialization program, the main control unit sends a synchronization trigger signal to the data acquisition module to ensure strict alignment of electromagnetic parameters and mechanical position on a microsecond time scale. The virtual simulation reconstruction module starts in parallel in the background. This module has a built-in analytical algorithm engine based on a cascaded transmission matrix, which uses parametric mathematical modeling technology to reconstruct the physical behavior of the phase shifter in virtual space.
[0048] Based on this, the dual-track differential extraction module constructs a differential comparator, which not only processes scalar data, but also performs phase unwrapping and amplitude normalization on vector complex data to eliminate system errors; the coupled verification and decision module performs pattern recognition-based logical operations to convert complex signal waveforms into executable production control instructions, such as release, rework, or scrapping, and sends the instructions to the production line actuators via industrial Ethernet;
[0049] This embodiment overcomes the technical bottleneck of traditional threshold judgment methods, which cannot distinguish between structural assembly air gaps and intrinsic material defects, by constructing a closed-loop architecture of physical measurement-virtual reconstruction-differential comparison. The system no longer simply judges whether the product performance is qualified, but uses synthetic analysis, that is, actively simulating faults in the virtual world and comparing them with reality, thereby intelligently identifying the physical causes of the faults. This design significantly improves the resolution of fault diagnosis, can accurately identify micron-level assembly air gaps, and avoids misjudging repairable assembly problems as material scrap, thereby significantly reducing production costs in the large-scale manufacturing scenario of base station antennas.
[0050] This embodiment defines the process by which the data acquisition module obtains the vector network parameter stream and mechanical displacement feedback data of the phase shifter: the phase shifter is scanned in real time within a preset frequency band to extract the return loss data and transmission phase data of the S-parameters, and the return loss data and transmission phase data are marked as the vector network parameter stream; the current physical position coordinates of the dielectric sheet are obtained through the position sensor, and the current physical position coordinates are marked as mechanical displacement feedback data.
[0051] This embodiment further refines the execution logic of multi-dimensional data synchronous acquisition; the data acquisition module drives the vector network analyzer to perform continuous frequency sweep within the communication frequency band of 1.7GHz to 2.7GHz, and the number of sampling points is set to a preset value, such as 201 points; this preset value is not arbitrarily selected, but is determined based on a trade-off between the Nyquist sampling theorem and hardware buffer limitations. Specifically, the determination method is as follows: calculate the ratio of the frequency band bandwidth of 1GHz to the estimated minimum resonant peak width of the phase shifter, and set an oversampling rate of not less than 3 times, while ensuring that the data volume does not exceed the buffer limit of a single scan of the vector network analyzer, so as to ensure that the frequency resolution is sufficient to capture narrowband resonant characteristics;
[0052] During this period, the module analyzes the complex data of the S11 port in real time, separating the amplitude information (return loss) and the phase information (transmission phase). These two sets of data constitute the electromagnetic fingerprint in the frequency domain. At the same time, the module reads the pulse signal of the high-precision linear grating ruler and calculates the real-time displacement of the dielectric sheet relative to the zero position. The system uses hardware timestamp technology to uniquely bind the S-parameters of each frequency point with the current physical position coordinates, generating a three-dimensional data packet containing frequency-position-performance. This process ensures the spatiotemporal consistency of the data and prevents measurement ambiguity caused by the movement of the dielectric sheet.
[0053] This embodiment integrates electromagnetic parameters (frequency and complex domains) with mechanical position (spatial domain) in a multidimensional manner, ensuring that subsequent analysis is based on a dynamic data cube with full dimensions. This enables the system to capture transient faults that are only triggered at specific dielectric sheet positions, such as when the dielectric sheet moves to the end of the slide rail and experiences a slight tilt. This greatly improves the detection system's ability to capture dynamic assembly defects and effectively solves the industry problem that static testing cannot detect dynamic hidden dangers.
[0054] This embodiment defines the process of constructing a lossless ideal transmission line model using the virtual simulation reconstruction module: obtain the stripline cavity height, conductor width, and dielectric sheet design dimensions, and mark them as ideal geometric parameters; use the odd-even mode analysis method to establish an ideal phase curve and a zero-reflection standing wave curve that change with the current physical position coordinates based on the ideal geometric parameters, and combine the ideal phase curve and the zero-reflection standing wave curve to mark the lossless ideal transmission line model.
[0055] This embodiment details the mathematical construction process of the baseline model; the virtual simulation reconstruction module retrieves the standard design parameters of the phase shifter from the CAD design database, namely the stripline cavity height b, the conductor width W, and the coupling spacing S, whose physical meanings are the horizontal distance between the edges of two parallel conductors in the stripline structure, the dielectric sheet thickness h, and the relative permittivity of the dielectric. The module applies the even-odd mode analysis method in microwave transmission line theory to calculate the characteristic impedance for the covered and uncovered areas of the dielectric sheet, respectively, in order to accurately simulate the impedance step effect.
[0056] Step 1, calculate the coverage area parameters: substitute... To the Cohn approximation formula; where the coefficients Originating from free space wave impedance A quarter approximation, i.e. To ensure dimensional consistency, the unit for this term is ohms in the calculation; the constant 0.441 in the formula is the Cohn stripline edge field capacitance correction coefficient, used to compensate for the electric field edge effect at the edge of the rectangular conduction strip; in the calculations of the following formulas, the constant term... All according to physical quantities The calculation formula for participating in the operation is as follows:
[0057]
[0058] It should be noted that the above characteristic impedance calculation formula is based on an approximate engineering solution of a specific cavity structure. Those skilled in the art can also use the precise analytical formula of Cohn theory or use commercial electromagnetic simulation software to extract the corresponding impedance value, which does not affect the implementation of the core logic of this invention.
[0059]
[0060] The composite characteristic impedance of the coverage area is obtained. and the phase constant of the coverage area ;
[0061] The second step is to calculate the parameters of the uncovered area, i.e., the air segment: substitute... To the same formula:
[0062]
[0063]
[0064] The synthesized characteristic impedance of the uncovered region is obtained. and the phase constant of the uncovered area Based on this, the ABCD transmission matrix for each transmission line segment is constructed, and its calculation formula is as follows:
[0065]
[0066] It should be noted that in all the above and subsequent calculation formulas, all physical quantities are calculated using the International System of Units (SI), specifically: frequency. The unit is Hertz, the speed of light Value meters per second, geometric dimensions Equal conversion to meters is used to ensure the phase constant. And the dimensional consistency of matrix operations;
[0067] The system uses a preset linear mapping function to map the current physical location coordinates. Convert to media cover area length
[0068]
[0069] The length of the uncovered area is calculated using the following formula:
[0070]
[0071] in, Defined as the physical length of the dielectric sheet along the signal transmission direction. Defined as the total physical length of the conductor band in the phase shifter cavity. Defined as the mechanical coordinate zero point when the front end of the dielectric sheet just enters the conductor coupling region; these three parameters are all directly extracted from the 3D CAD drawings of the product and solidified as system constants;
[0072] The calculation formula for performing segmented matrix concatenation to synthesize the overall transmission characteristics is as follows:
[0073]
[0074] The system is based on two-port network conversion theory, and converts the total transmission matrix... Converting to S-parameters, the calculation formula is as follows:
[0075]
[0076]
[0077] in, The system reference impedance is 50 ohms; based on the calculated... Phase data and Amplitude data is used to generate two reference curves. This embodiment accurately introduces the wave impedance discontinuity caused by dielectric loading by distinguishing the impedance difference between the covered and uncovered areas. This ensures that the zero-reflection standing wave curve can truly reflect the inherent return loss noise of the design, and avoids misjudging normal physical reflections as fault signals due to overly idealized models that ignore impedance mismatch.
[0078] Example 2:
[0079] This embodiment defines the process by which the virtual simulation reconstruction module generates theoretical S-parameters with defects based on the air gap perturbation equivalent model: a preset air gap perturbation equivalent model is introduced, and the gap between the contact surface of the dielectric sheet and the conductor strip is equivalent to the perturbation of the transmission line distribution parameters; a preset air gap value is forcibly set in the lossless ideal transmission line model, and inverse simulation is performed in combination with the perturbation of the transmission line distribution parameters to calculate the S-parameter variation data of the lossless ideal transmission line model under the air gap value, and the S-parameter variation data is marked as theoretical S-parameters with defects.
[0080] This embodiment elaborates on the active generation mechanism of fault characteristics. The system introduces an equivalent model of air gap perturbation, which is based on electromagnetic field perturbation theory and aims to transform the micron-level air layer, which is difficult to measure directly, into circuit parameters. Specifically, the system equates the gap between the dielectric sheet and the conductor strip caused by loose assembly to the perturbation of the characteristic impedance of the transmission line, and the calculation formula is as follows:
[0081]
[0082] At the same time, the phase constant changes:
[0083]
[0084] in, The equivalent characteristic impedance when the air gap exists, in units of: ; The characteristic impedance when ideally fitted; The preset standard air gap thickness, in units of: ; The height of the cavity; The relative permittivity of the dielectric is obtained by the following method: A standard calibration procedure is established, using a set of precision feeler gauges with thicknesses of 10µm, 30µm, and 50µm. The dielectric sheet is manually raised, and the corresponding S-parameter degradation curves are measured. The measured curves are then fitted to the theoretical curves of the model using the least squares method to solve for the optimal matching equivalent air gap thickness parameter in the model. This serves as the preset value for subsequent production, thereby eliminating the destructive testing costs and statistical errors caused by slice analysis;
[0085] The system forces the capacitance parameter to be connected in series in the transmission matrix of the lossless ideal transmission line model; specifically, the system constructs a two-port network matrix describing the air gap perturbation characteristics. The matrix is defined as follows:
[0086]
[0087] in, The angular frequency is given by , and j is the imaginary unit. This matrix is then cascaded to the transmission matrix covering the dielectric sheet area. The input and output terminals form a defective total transmission matrix. Based on the total transmission matrix, reverse circuit simulation is performed to calculate the S-parameter deterioration that the phase shifter should exhibit under a specific air gap value, thereby generating theoretical S-parameters with defects.
[0088] This embodiment transforms the mechanical air gap in the physical world into quantifiable electrical signal features through active fault injection. This is equivalent to the system mastering the fingerprint features of air gap faults, providing an accurate digital template for subsequent feature matching, thereby realizing a deep deconstruction of the air gap fault mechanism, enabling the system to identify those microscopic assembly defects that are invisible to the naked eye.
[0089] This embodiment defines the process by which the dual-track differential extraction module calculates the theoretical residual feature vector and the actual residual feature vector: the reference data of the lossless ideal transmission line model is subtracted from the vector network parameter flow to obtain the actual residual feature vector containing environmental noise and manufacturing tolerances; the reference data of the lossless ideal transmission line model is subtracted from the theoretical defective S-parameters to extract the signal distortion features caused only by the air gap value, and the signal distortion features are marked as the theoretical residual feature vector.
[0090] This embodiment details the signal feature extraction and purification process. The dual-track differential extraction module executes two parallel differential operation paths. In the real-world path, the module performs complex subtraction between the measured vector network parameter stream and the ideal reference data to obtain the real residual feature vector, which incorporates environmental noise, material discreteness, and potential fault features. Simultaneously, in the theoretical path, the module performs differential analysis between the simulated theoretical defect-containing S-parameters and the same ideal reference data to extract the signal distortion features purely caused by the air gap, i.e., the theoretical residual feature vector. The calculation formulas for these two vectors are as follows:
[0091]
[0092]
[0093] in, The feature vector of the actual residual; The theoretical residual eigenvector; This is the measured vector network parameter flow; Theoretical S-parameters containing defects; Reference data for a lossless ideal transmission line model;
[0094] This embodiment employs a dual-track differential strategy, which cleverly eliminates common-mode errors in the system, such as the basic insertion loss of the test cable and the fixed reflection of the fixture. More importantly, it simplifies the complex fault identification problem into a morphological comparison problem of two vectors, greatly improving the signal-to-noise ratio of the identification and enabling weak air gap signals to be extracted from complex background noise.
[0095] This embodiment defines the process by which the coupled verification decision module calculates the topological similarity between the theoretical residual feature vector and the actual residual feature vector: the cross-correlation function or cosine similarity value between the actual residual feature vector and the theoretical residual feature vector is calculated in the frequency domain; and the cross-correlation function or cosine similarity value is marked as the topological similarity.
[0096] This embodiment specifically illustrates the fingerprint comparison logic of fault features; the coupled verification and decision module focuses on the geometric shape of the waveform, and the system uses a weighted cosine similarity algorithm to evaluate the consistency of two feature vectors in the frequency domain; in order to fully utilize the complete information contained in the vector network parameter stream, the system constructs a high-dimensional feature vector: the amplitude sequence in the frequency domain... Normalize and convert the phase sequence Perform untangling and delinearization processing;
[0097] Among these, unwrapping refers to unfolding the phase data to avoid phase distortion. A transition occurs at this point; delinearization refers to removing the linear phase component caused by group delay; the specific steps of delinearization are: fitting a linear trend line of phase change with frequency using the least squares method, and the calculation formula is:
[0098]
[0099] in, Group delay slope, Based on the phase offset, calculate the delinearized phase sequence. This eliminates the global time delay linear component caused by fixture positioning error while retaining the nonlinear fluctuation characteristics caused by local air gaps; the system concatenates the normalized amplitude sequence with the delinearized phase sequence to form an extended-dimensional feature vector. Here, the logic for setting the weighting coefficients is as follows: calculate the standard deviation of the amplitude of the historical benchmark good product data. With phase standard deviation The calculation formula is as follows:
[0100]
[0101] in, To prevent the denominator from being zero, a preset regularization constant is used, such as To eliminate the difference in dimensions between amplitude dB and phase degree, and achieve energy balance in multidimensional feature space;
[0102] Waveform similarity is characterized by calculating the cosine of the angle between two extended-dimensional vectors, as shown in the following formula:
[0103]
[0104] in, Topological similarity is dimensionless; This is a dimension-extended vector of the real residual, which is composed of concatenated amplitude and phase. This is a dimension-extended vector of the theoretical residual, which is composed of amplitude and phase. This represents the number of samples taken at the frequency points of the sweep.
[0105] The numerical value calculated by the system directly reflects the degree of morphological agreement between the measured waveform and the theoretical air gap waveform in the multidimensional vector space, and is labeled as the topological similarity; the value is 1 in response to the waveforms being completely in phase; the value approaches 0 in response to the waveforms being orthogonal or unrelated.
[0106] This embodiment compensates for the defect that single amplitude comparison may lose phase distortion features by constructing an extended dimension feature vector containing phase information. Since the difference in air gap thickness mainly affects the amplitude of signal distortion, and the existence of air gap itself determines the periodicity of distortion, this method can identify the essence of air gap problems with extremely high sensitivity. Even when the amplitude of the test signal itself fluctuates greatly, it can accurately determine whether there is an air gap fault.
[0107] This embodiment defines the process by which the coupling verification decision module generates fault attribution results and corresponding production control instructions based on topological similarity: a first similarity threshold is preset, and if the topological similarity is greater than the first similarity threshold, the cause of the abnormality is determined to be an assembly air gap fault; a production control instruction containing a secondary pressurization request is generated, and the production control instruction is sent to the control terminal of the automated crimping equipment.
[0108] This embodiment details a closed-loop control strategy for repairable faults. The system presets a relatively high first similarity threshold, such as 0.85, which is selected based on ROC curve analysis. The coupled verification and decision module monitors the topology similarity value in real time. When the value exceeds the first similarity threshold, it determines that there is an assembly air gap fault. The system generates a production control command containing a secondary pressurization request. During this process, the system calls the process compensation parameter table pre-stored in the main control unit to calculate the projection coefficient of the actual residual feature vector onto the theoretical residual feature vector. This coefficient Characterizes the severity of the air gap;
[0109] To ensure safe equipment operation and prevent workpiece crushing due to calculation overflow, the system incorporates a pressure safety fuse mechanism; the original target pressure holding parameters are calculated, which involves fundamental process constants, among which... Defined as the basic holding pressure value set by the standard process, for example: the standard preload pressure value of 20N determined based on the defect-free product crimping test, as shown in the following formula:
[0110]
[0111] in, This is the pressure sensitivity coefficient; its source and calibration method are as follows: During the system debugging phase, a gradient pressure step is applied to the standard sample using a high-precision force gauge, with the step value strictly set to 5N, covering the elastic deformation range from 0N to 50N, and the projection coefficient is recorded simultaneously. The change in pressure value was analyzed using linear regression analysis with projection coefficients using ordinary least squares method, and the slope of the regression line was extracted as the sensitivity coefficient. The specific value; at the same time, obtain the preset equipment limit pressure threshold. This value is set to 80% of the yield strength of the medium material; the system executes the logic for generating the final instruction parameters, and its calculation formula is:
[0112]
[0113] If the calculated result In addition to issuing In addition to the pressure holding command, a process over-limit warning signal is triggered simultaneously.
[0114] Ultimately, the system will include the target holding pressure parameters that have undergone safety cutoff processing. The production control instructions for the workstation number are sent to the PLC control terminal of the automated crimping equipment via the industrial bus. This embodiment not only realizes the automated closed loop of detection-diagnosis-repair, but also solves the problem that the repair force cannot be accurately controlled by qualitative judgment alone by introducing the dual logic of quantitative calculation of projection coefficient and safety threshold cutoff. At the same time, it avoids the risk of positive feedback runaway that may occur in closed-loop control.
[0115] This embodiment further defines the process by which the coupled verification decision module generates fault attribution results and corresponding production control instructions based on topological similarity: a second similarity threshold and a total error threshold are preset, wherein the first similarity threshold is greater than the second similarity threshold; if the topological similarity is less than the second similarity threshold and the amplitude of the actual residual feature vector is greater than the total error threshold, the abnormality is determined to be due to a defect in the medium material; a production control instruction containing a defective product diversion mark is generated, and the production control instruction is sent to the control terminal of the sorting system.
[0116] This embodiment details the diversion strategy for irreversible faults. The system further presets a low second similarity threshold, such as 0.4, and a total error threshold characterizing signal energy. The total error threshold is set using statistical methods: during system initialization, M S-parameter residual data of standard good products (e.g., M=50) are collected, their energy norm distribution is calculated, and the threshold is set to the mean plus 6 times the standard deviation. The actual physical meaning of this threshold is the energy anomaly threshold.
[0117] As a specific example of experimental data, in a field calibration of this system, the mean value of the residual energy integral of 50 good samples was collected. Standard deviation Based on this, the total error threshold is set to To ensure that the false positive rate is less than one in a million;
[0118] The coupled verification decision module performs a compound logic judgment: if the topological similarity is lower than the second similarity threshold, it indicates that the measured waveform does not match the air gap characteristics at all; simultaneously, the system calculates the energy integral of the actual residual feature vector. If the value exceeds the total error threshold, it indicates a significant performance deviation rather than measurement noise. When the above two conditions are met, the system determines that the abnormality is caused by a defect in the medium material. The system generates a production control instruction containing a defective product diversion mark and drives the sorting robot arm to move the workpiece into the scrap area.
[0119] This embodiment uses a combination of low similarity and high error values, along with an adaptive threshold determined by statistics, to accurately identify non-air gap type persistent faults. This effectively prevents meaningless secondary pressurization of products with inherent material defects, avoids wasting valuable production cycles and equipment lifespan, and ensures that only truly irreparable products are scrapped, thereby optimizing the overall efficiency of production resource allocation.
[0120] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims
1. A stripline dielectric phase-shifting system for base station antennas, characterized in that, It includes a main control unit, which is communicatively connected to a data acquisition module, a virtual simulation reconstruction module, a dual-track differential extraction module, and a coupled verification and decision module; The data acquisition module is used to acquire the vector network parameter stream and mechanical displacement feedback data of the phase shifter; The virtual simulation reconstruction module is used to construct a lossless ideal transmission line model based on ideal geometric parameters, and to generate theoretical S-parameters with defects based on the air gap perturbation equivalent model. The dual-track differential extraction module is used to calculate the theoretical residual feature vector and the actual residual feature vector based on the lossless ideal transmission line model, the theoretical defective S-parameters and the vector network parameter flow. The coupled verification decision module is used to calculate the topological similarity between the theoretical residual feature vector and the actual residual feature vector, and to generate fault attribution results and corresponding production control instructions based on the topological similarity.
2. The stripline dielectric phase-shifting system for a base station antenna according to claim 1, characterized in that, The process by which the data acquisition module acquires the vector network parameter stream and mechanical displacement feedback data of the phase shifter includes: The phase shifter is scanned in real time within a preset frequency band range to extract the return loss data and transmission phase data of the S-parameters, and the return loss data and transmission phase data are marked as the vector network parameter stream. The current physical position coordinates of the medium sheet are obtained by a position sensor, and the current physical position coordinates are marked as the mechanical displacement feedback data.
3. The stripline dielectric phase-shifting system for a base station antenna according to claim 2, characterized in that, The process of constructing a lossless ideal transmission line model by the virtual simulation reconstruction module includes: Obtain the stripline cavity height, conductor width, and dielectric sheet design dimensions, and mark them as the ideal geometric parameters; Using the odd-even mode analysis method, an ideal phase curve and a zero-reflection standing wave curve that vary with the current physical location coordinates are established based on the ideal geometric parameters. The combination of the ideal phase curve and the zero-reflection standing wave curve is labeled as the lossless ideal transmission line model.
4. A stripline dielectric phase-shifting system for a base station antenna according to claim 3, characterized in that, The virtual simulation reconstruction module generates theoretical S-parameters containing defects based on the air gap perturbation equivalent model, including the following process: By introducing the preset equivalent model of air gap perturbation, the contact gap between the dielectric sheet and the conductor strip is equivalent to the perturbation of the transmission line distribution parameters; In the lossless ideal transmission line model, a preset air gap value is forcibly set, and inverse simulation is performed in combination with the perturbation. The S-parameter variation data of the lossless ideal transmission line model under the air gap value is calculated, and the S-parameter variation data is marked as the theoretical defective S-parameter.
5. A stripline dielectric phase-shifting system for a base station antenna according to claim 4, characterized in that, The process by which the dual-track difference extraction module calculates the theoretical residual feature vector and the actual residual feature vector includes: Subtracting the baseline data of the lossless ideal transmission line model from the vector network parameter stream yields a realistic residual feature vector that includes environmental noise and manufacturing tolerances. Subtract the baseline data of the lossless ideal transmission line model from the theoretical defective S-parameters, extract the signal distortion features caused only by the air gap value, and mark the signal distortion features as the theoretical residual feature vector.
6. A stripline dielectric phase-shifting system for a base station antenna according to claim 5, characterized in that, The process by which the coupled verification decision module calculates the topological similarity between the theoretical residual feature vector and the actual residual feature vector includes: Calculate the cross-correlation function or cosine similarity value between the actual residual feature vector and the theoretical residual feature vector in the frequency domain; The cross-correlation function value or the cosine similarity value is labeled as the topological similarity.
7. A stripline dielectric phase-shifting system for a base station antenna according to claim 6, characterized in that, The process by which the coupling verification decision module generates fault attribution results and corresponding production control instructions based on the topological similarity includes: A first similarity threshold is preset. If the topological similarity is greater than the first similarity threshold, the cause of the abnormality is determined to be an assembly air gap fault. A production control command containing a secondary pressurization request is generated and sent to the control terminal of the automated pressing equipment.
8. A stripline dielectric phase-shifting system for a base station antenna according to claim 7, characterized in that, The process by which the coupling verification decision module generates fault attribution results and corresponding production control instructions based on the topological similarity also includes: A second similarity threshold and a total error threshold are preset, wherein the first similarity threshold is greater than the second similarity threshold; If the topological similarity is less than the second similarity threshold and the magnitude of the actual residual feature vector is greater than the total error threshold, then the cause of the anomaly is determined to be a defect in the dielectric material. Generate a production control instruction containing a defective product diversion mark, and send the production control instruction to the control terminal of the sorting system.