An ultrasonic horn optimization management method and system

Abstract

This invention proposes an optimization management method and system for ultrasonic amplitude transformers. The method includes: receiving a set of performance target parameters set during system initialization; dividing the amplitude transformer into an input segment, an intermediate transition segment, and an output segment to construct a structural design model, and then optimizing the structural design model inversely through a performance prediction module to generate a corresponding parameter set; collecting SLM process parameters, and inputting the parameter set into a pre-constructed error prediction module to generate a manufacturing error prediction; collecting and combining real-time amplitude transformer performance data with the manufacturing error prediction and performance offset, and generating an individual final performance value through Bayesian weighted fusion, and constructing an individual performance label for each individual amplitude transformer; and collecting tissue impedance, probe temperature, and ultrasonic echo signals in real-time during surgery, and dynamically adjusting the transducer drive voltage based on the individual performance label to perform closed-loop amplitude control. This invention provides a safer and more intelligent energy output solution for highly complex surgical scenarios.

Description

Technical Field

[0001] This invention belongs to the field of ultrasonic amplitude transformer optimization management, and particularly relates to an ultrasonic amplitude transformer optimization management method and system. Background Technology

[0002] As a core energy instrument in minimally invasive surgery, the ultrasonic scalpel relies on a high-frequency vibration chain composed of a transducer, an amplitude transformer, and a blade to achieve simultaneous tissue cutting and coagulation. The amplitude transformer not only amplifies and transmits vibrational energy but also largely determines the resonance efficiency and output stability of the entire system. Existing amplitude transformer designs often employ regular surface shapes (such as conical, exponential, or stepped structures) combined with finite element simulation for parameter optimization. While these designs possess a certain amplification capability, they still suffer from problems such as rapid amplitude decay, modal instability, and resonance drift when facing complex tissue loads and high energy output demands. Meanwhile, although additive manufacturing technology has significantly increased the freedom of structural design, it has also introduced manufacturing uncertainties such as structural mismatch, geometric deviations, and batch inconsistencies, leading to significant deviations between theoretical design and actual performance.

[0003] Furthermore, the types of tissues involved in surgery are complex and diverse, with significant differences in density, acoustic impedance, and moisture content. Existing equipment mostly uses fixed frequency and amplitude control, which cannot respond to tissue feedback specifically. This often leads to problems such as local overheating of the scalpel tip, tissue carbonization, and even adhesion, further reducing surgical precision and safety. Current technology lacks a systematic integration mechanism in the three key stages of structure generation, manufacturing adaptation, and intraoperative control, making it impossible to perform feedforward correction or intraoperative closed-loop control for performance deviations, thus hindering the intelligent and personalized development of ultrasonic scalpels. Summary of the Invention

[0004] The purpose of this invention is to propose an ultrasonic amplitude transformer optimization management method and system. Starting from the preset surgical performance target, a parametric multi-segment structure modeling mechanism is adopted, and a neural network performance regression model is combined for reverse optimization. The output is an individualized amplitude transformer structure that meets the target amplitude, frequency and modal stability, and ensures its continuity and manufacturability under additive manufacturing.

[0005] To achieve the above objectives, a first aspect of the present invention provides an ultrasonic amplitude transformer optimization management method, the method comprising the following steps:

[0006] The set of performance target parameters set during the initialization of the receiving system includes the target output amplitude multiplication factor, the target resonant frequency, and the target modal stability factor;

[0007] The amplitude transformer is divided into an input segment, an intermediate transition segment, and an output segment to construct a structural design model. The structural design model is then optimized in reverse by a performance prediction module to generate a corresponding parameter set. The parameter set includes segment length, maximum / minimum diameter, cross-sectional function type, and end cone angle.

[0008] Collect SLM process parameters, combine the parameter set with the pre-built error prediction module to generate manufacturing error prediction; calculate the performance deviation caused by the error through a digital twin model to generate the performance deviation amount;

[0009] Real-time amplitude transformer performance data is collected and combined with the manufacturing error prediction and performance offset. The final individual performance value is generated by Bayesian weighted fusion, and an individual performance label is constructed for each individual amplitude transformer. The individual performance label includes a structural parameter summary, a manufacturing error vector, the individual final performance value, and a modal stability sensitivity factor.

[0010] During the procedure, tissue electrical impedance, blade temperature, and ultrasonic echo signals are collected in real time. Combined with the individual performance tags, the transducer drive voltage is dynamically adjusted to achieve closed-loop amplitude control.

[0011] Furthermore, the structural design model is generated using a geometric generation function; wherein,

[0012] The process of back-optimizing the structural design model through the performance prediction module to generate a corresponding parameter set is as follows:

[0013] The parameters of the performance target parameter set are respectively input into the corresponding pre-built prediction model to generate the corresponding current prediction structure performance;

[0014] The optimization objective of the prediction model is to minimize the overall performance deviation loss; wherein the overall performance deviation loss is obtained by calculating the deviation between the current performance target parameter set and the current prediction structure performance set.

[0015] The initial structure search uses Latin hypercube sampling to generate 500 structures, and then performs gradient descent-based optimization iterations, calling the prediction model to update the performance estimate one by one until the overall performance deviation is lost to a set threshold or the number of iterations exceeds 300.

[0016] Furthermore, the pre-built prediction model is implemented by a regression network composed of a set of multilayer perceptrons, with each performance index corresponding to a model. The model structure is a 4-layer perceptron, with each layer having nodes of [64,128,128,1] and the activation function being ReLU.

[0017] Furthermore, the process of collecting SLM process parameters and combining the parameter set with a pre-built error prediction module to generate manufacturing error predictions specifically involves:

[0018] The parameter set and SLM process parameters are input into the integrated error prediction network, and the original predicted manufacturing error is output.

[0019] Combining the original predicted manufacturing error and the actual manufacturing error, the manufacturing error prediction is obtained by combining the objective function of weighted least squares method and structurally sensitive region constraints; wherein, the structurally sensitive region constraints are calculated based on the set of structural parameters most sensitive to performance; wherein the set of structural parameters most sensitive to performance includes the output end angle and the location of abrupt changes in cross-section;

[0020] The objective function is used to minimize the actual manufacturing error.

[0021] Furthermore, the process of calculating the performance offset caused by the error through the digital twin model and generating the performance offset amount specifically involves:

[0022] A performance disturbance prediction model is constructed based on the manufacturing error prediction and performance prediction module, and a performance offset is obtained by correcting it based on the deviation correction term. The performance offset includes amplitude difference, frequency difference and modal stability difference. The deviation correction term is calculated based on the output of the manufacturing error prediction and performance prediction module and the performance target parameter set.

[0023] Furthermore, the acquisition, combined with real-time amplitude transformer performance data, manufacturing error prediction, and performance offset, generates individual final performance values ​​through Bayesian weighted fusion, and constructs an individual performance label for each individual amplitude transformer, specifically as follows:

[0024] Calculate digital twin predictions of the performance offset and the structural optimization target performance; wherein the structural optimization target performance is calculated based on the performance target parameter set.

[0025] A fusion performance estimation model is designed to minimize the weighted trade-off between the digital twin prediction value and the real-time amplitude transformer performance data, thereby obtaining the individual final performance value; wherein the individual final performance value includes the final amplitude multiplication factor, the final resonant frequency, and the final modal stability factor.

[0026] The modal stability sensitivity factor is calculated based on the parameter set of the output segment and the final modal stability factor.

[0027] The individual performance label for each individual amplitude rod is generated by combining the structural parameter summary of the individual final performance value, the manufacturing error vector, the individual final performance value, and the modal stability sensitivity factor.

[0028] Furthermore, the individual performance tags are bound to the device body via RFID, laser coding, or other methods, stored in the database, and uploaded synchronously.

[0029] Furthermore, the intraoperative real-time acquisition of tissue electrical impedance, blade temperature, and ultrasonic echo signals, combined with the individual performance tags, dynamically adjusts the transducer drive voltage to achieve closed-loop amplitude control, specifically as follows:

[0030] Select any individual amplitude transformer as the target individual and collect tissue feedback signals in real time; wherein the tissue feedback signals include electrical impedance, local temperature of the cutting head, and morphological indicators of tissue feedback signals obtained by ultrasonic echo analysis.

[0031] Based on the tissue feedback signal of the target individual, the final performance value of the individual is designed by combining the thermal safety threshold and the final performance value of the individual. A target control function is designed with the goal of minimizing the weighted sum of amplitude tracking error and temperature safety deviation. This function is used to adjust the output amplitude so as to avoid excessive energy accumulation and adhesion risk while maintaining the expected performance.

[0032] According to the target control function, the controller updates the amplitude driving voltage by looking up a table to obtain the output voltage, which is used to regulate the ultrasonic energy output.

[0033] Furthermore, the modulation frequency is set to 20Hz to cover the response time of each action under typical surgery.

[0034] A second aspect of the present invention provides an ultrasonic amplitude transformer optimization management system, the system comprising:

[0035] The parameter acquisition module is used to receive the set of performance target parameters set during system initialization, including the target output amplitude multiplication factor, the target resonant frequency, and the target modal stability factor.

[0036] The amplitude transformer analysis module is used to divide the amplitude transformer into an input segment, an intermediate transition segment, and an output segment to construct a structural design model. The performance prediction module then back-optimizes the structural design model to generate a corresponding parameter set. The parameter set includes segment length, maximum / minimum diameter, cross-sectional function type, and end cone angle.

[0037] The parameter analysis module is used to collect SLM process parameters, combine the parameter set with the pre-built error prediction module to generate manufacturing error prediction; and calculate the performance deviation caused by the error through a digital twin model to generate the performance deviation amount.

[0038] The parameter adjustment module is used to collect and combine real-time amplitude bar performance data, and combine the manufacturing error prediction and performance offset to generate individual final performance values ​​through Bayesian weighted fusion, and construct individual performance labels for each individual amplitude bar; wherein the individual performance labels include structural parameter summary, manufacturing error vector, individual final performance value and modal stability sensitivity factor;

[0039] The parameter optimization module is used to collect tissue impedance, blade temperature and ultrasonic echo signals in real time during the operation. Combined with the individual performance tag, it dynamically adjusts the transducer drive voltage to perform closed-loop amplitude control.

[0040] The beneficial technical effects of the present invention are at least as follows:

[0041] This method starts from a pre-defined surgical performance target, employs a parametric multi-segment structure modeling mechanism, and combines it with a neural network performance regression model for backpropagation optimization. It outputs an individualized amplitude-variable rod structure that satisfies the target amplitude, frequency, and modal stability, while ensuring its continuity and manufacturability under additive manufacturing. During the manufacturing stage, the system constructs an error modeling framework that integrates an error prediction network with a structure-sensitive regularization term. It weights and identifies manufacturing disturbances in key areas such as the tip of the cutting tool, and further propagates the impact of these errors on the target performance through a performance model, establishing a "manufacturing error-performance offset" mapping relationship.

[0042] During the individual registration stage, the system integrates digital twin prediction results with measured data, and forms a complete performance label based on a structure-sensitive Bayesian fusion strategy. This label includes a structural summary, error estimation, fusion performance, and modal risk factors, enabling traceable individual identification and preoperative screening. Finally, during the intraoperative control stage, the system senses real-time tissue feedback (such as impedance, temperature, and ultrasound echo) and dynamically adjusts the driving voltage based on the individual target amplitude and modal risk level in the label. This achieves personalized control of ultrasound energy output, thereby reducing the risk of heat buildup and adhesion. This invention is the first to construct an integrated management method for the entire process from structure generation and manufacturing error control to intraoperative regulation, enabling the design, prediction, and dynamic adjustment of the amplitude transformer performance, providing a safer, more stable, and intelligent energy output solution for highly complex surgical scenarios. Attached Figure Description

[0043] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0044] Figure 1 This is a flowchart of an ultrasonic amplitude transformer optimization management method disclosed in an embodiment of the present invention.

[0045] Figure 2 This is a framework diagram of an ultrasonic amplitude transformer optimization management system disclosed in an embodiment of the present invention. Detailed Implementation

[0046] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0047] Example 1

[0048] like Figure 1 As shown in the figure, an embodiment of the present invention provides an ultrasonic amplitude transformer optimization management method, the method comprising:

[0049] S1. The set of performance target parameters set during the initialization of the receiving system, including the target output amplitude multiplication factor, the target resonant frequency, and the target modal stability factor.

[0050] Specifically, the set of performance target parameters set during system initialization includes:

[0051] ;

[0052] in: : Target output amplitude multiplication factor (dimensionless), set according to the energy amplification requirements of clinical use; Target resonant frequency (in Hz), preset according to the transducer frequency range (55–60kHz); The target modal stability factor (dimensionless) is defined as the modal drift coefficient within ±2% of the target frequency band, and is set based on the modal experimental database.

[0053] S2. Divide the amplitude transformer into an input segment, an intermediate transition segment, and an output segment to construct a structural design model. Then, use the performance prediction module to back-optimize the structural design model and generate a corresponding parameter set. The parameter set includes segment length, maximum / minimum diameter, cross-sectional function type, and end cone angle.

[0054] Specifically, this step aims to generate a manufacturable, controllable, and stable amplitude transformer structure design based on performance indicators set for clinical surgical needs. This design serves as the structural foundation for subsequent manufacturing simulation, digital twin modeling, and control strategy configuration. The key lies in establishing an effective mapping between target performance and specific geometric configuration, starting from the performance objectives and using parametric structural representation as a bridge. Considering that the actual manufacturing process is additive manufacturing such as selective laser melting (SLM), this approach also introduces structural continuity constraints to ensure the generated model is manufacturable.

[0055] Furthermore, to achieve the reverse engineering of the structural model from performance, the design divides the amplitude transformer into three sections: an input section (connecting the transducer), an intermediate transition section (amplitude adjustment region), and an output section (connecting the cutter head). Each section is defined by a parameter set. Expressing geometric features, including cross-sectional shape (cylindrical or conical), cross-sectional variation law (linear, exponential, or Bessel function), segment length, etc. Structure generating function writing:

[0056] ;

[0057] in: The output structural model (3D parametric geometry object) is used to generate manufacturing files; : Geometric generation function, which constructs a CAD model based on input parameters; : The set of parameters for each segment, including segment length (unit: mm), cross-sectional variation law (function type), and cross-sectional dimensions.

[0058] Furthermore, to ensure the structure meets performance targets, the system introduces a loss function-driven structural backward optimization process. The performance prediction module is implemented using a regression network composed of a set of multilayer perceptrons (MLPs), with one model corresponding to each performance metric. Taking amplitude prediction as an example, the input is... The output is The model structure is a 4-layer perceptron, with each layer having nodes in the range of [64, 128, 128, 1]. The activation function is ReLU, and the training samples are from 2400 sets of structured data, covering the typical design parameter range.

[0059] The optimization objective is defined as minimizing the overall performance deviation loss:

[0060]

[0061] in: : Performance deviation loss function (dimensionless), used to quantify the difference between the current design and the target; : The current predicted structural performance calculated by the performance prediction network; Weighting coefficients are derived from user strategy settings, such as adjusting weights based on the type of surgery. : The complete set of parameters for the current design structure.

[0062] The initial structure search uses Latin hypercube sampling to generate 500 structures, followed by gradient descent-based optimization iterations. The performance estimation is updated sequentially by calling a performance network until... The value is less than the set threshold or the number of iterations exceeds 300.

[0063] For example, when the performance target is set to , kHz, The final generated structure may consist of an input section with a constant diameter cylinder (8mm), a middle section with an exponential cone shape (diameter ranging from 4.8mm to 2.2mm, length 14mm), and an output section with a linearly tapered diameter (terminal diameter 1.5mm, length 6mm). This parameter set... This is the optimal structural design.

[0064] The final output includes: Structural design model, which can be exported as an STL or STEP file; : Parameter set, used for subsequent manufacturing simulation and performance prediction modules.

[0065] S3. Collect SLM process parameters, combine the parameter set with the pre-built error prediction module to generate manufacturing error prediction; calculate the performance offset caused by the error through the digital twin model to generate the performance offset amount.

[0066] The goal of this step is to establish a closed-loop modeling system encompassing "structural parameters → manufacturing errors → performance deviations," thereby predicting the actual performance of the amplitude transformer before manufacturing. This ensures that the amplitude, frequency, and modal stability of the cutter head fluctuate within a controllable range, thus providing a robust guarantee for the stable operation of the anti-adhesion ultrasonic cutter. The input to this step is the structural model output from step one. and its parameter set The output is the manufacturing error prediction. With performance offset This serves as an important foundation for subsequent error control strategies.

[0067] The technical background of the invention dictates that the present invention must consider the following special characteristics that differ from ordinary structural digital twins: (1) Ultrasonic amplitude transformers usually have complex cross-sectional transition regions, and laser additive manufacturing is prone to stress accumulation and non-uniform shrinkage; (2) The tip of the cutter is particularly sensitive to performance changes, and even slight deviations can lead to amplitude loss, frequency drift, or even resonant mode switching, seriously affecting the anti-adhesion capability; (3) The structure itself is more functional than rigid support, so the impact of manufacturing errors cannot only be considered geometrically, but also functional disturbances must be given special attention. Therefore, this step specifically introduces structural sensitivity modeling and innovative regularization term design in the error modeling and performance propagation parts, reflecting the innovation of the present invention in manufacturing-structural performance co-design.

[0068] The steps to input are: :: Geometric models generated by the parametric modeling module GeoGen, in 3D CAD format (such as STL or STEP), containing complete structural topology; The structural parameter set, including segment length, maximum / minimum diameter, cross-sectional function type, and end cone angle, is derived from the structural optimization process in step one. In the manufacturing error modeling stage, this invention uses... and SLM process parameters Design an error prediction network for manufacturing process disturbances, using the input as the input. .

[0069] Furthermore, SLM process parameters The initial settings are based on the recommended parameters provided by the SLM equipment manufacturer (such as the default parameters for a certain model of equipment for TC4 titanium alloy).

[0070] To enhance its predictive ability for key regions, a regularization term for "structurally sensitive region constraint" is introduced into the objective function. Error weights are used to highlight the most performance-sensitive locations, such as the tip of the cutter and geometric abrupt changes in the transition section.

[0071] The error prediction formula is as follows:

[0072] ;

[0073] in: Manufacturing error prediction, representing the deviation value of each geometric parameter; : An integrated error prediction network takes structural and process parameters as input and outputs the original predicted manufacturing error; : Actual manufacturing error (from actual sample training or simulation calibration); The structural sensitivity regularization term is defined as follows:

[0074] ;

[0075] in It is the set of structural parameters that are most sensitive to performance (such as the output angle and the location of abrupt changes in the cross section). Weights, either empirically set or data-driven, are used to amplify the impact of errors at these locations; Regular weights control the trade-off between performance-sensitive constraints and overall error regression. The error prediction module is implemented using an integrated regression neural network, consisting of three sub-models, which are used to model shrinkage errors caused by thermal stress, protrusion errors caused by abnormal powder accumulation, and deformation errors caused by uneven scanning paths. Each sub-model is a four-layer perceptron network, with the number of nodes in each layer being [128, 128, 64, output dimension].

[0076] Obtaining error Next, the process moves to the digital twin performance perturbation modeling stage. This invention uses the performance prediction model from step one. However, to account for the uncertainty of error propagation and the sensitive amplification effect of local nonlinearity, an "asymmetric perturbation propagation deviation term" is specifically added. Therefore, the performance disturbance prediction model is constructed as follows:

[0077] ;

[0078] in: Performance offset, including (amplitude difference), (Frequency difference) (Poor modal stability); The deviation correction term, used to capture nonlinear amplification factors, is defined as follows:

[0079] ;

[0080] This calculation, based on the second derivative (Hessian approximation) of the performance prediction network, is used to compensate for the large performance perturbations that small errors may cause in the highly sensitive region. The scaling factor was set through experimental fitting.

[0081] The final model output includes: Manufacturing error prediction based on structural and manufacturing process conditions; Performance offset prediction results after error propagation.

[0082] This modeling strategy has the following key innovations: (1) The manufacturing error modeling introduces a structurally sensitive constraint term, which specifically improves the error modeling capability of key performance sensitive areas; (2) The digital twin performance prediction introduces a second-order partial derivative correction term for the first time. (3) The entire process is based on the company's actual manufacturing test data rather than simulation, ensuring that the model is lightweight and deployable, and has real-world feasibility.

[0083] S4. Collect and combine real-time amplitude transformer performance data, and combine the manufacturing error prediction and performance offset to generate individual final performance values ​​through Bayesian weighted fusion, and construct individual performance labels for each individual amplitude transformer; wherein the individual performance labels include structural parameter summaries, manufacturing error vectors, individual final performance values, and modal stability sensitivity factors.

[0084] The main objective of this step is to address the manufacturing errors obtained in the previous step. Performance skew prediction Combined with actual measurement data Establish individual performance profiles for the amplitude transformers and generate tags for them. The label must include structural features, manufacturing process disturbances, predictive-measured fusion performance, and risk factors to enable key functions such as personalized intraoperative configuration, preoperative grading and screening, and postoperative accuracy tracking. It is particularly suitable for surgical scenarios where "blade tip is prone to adhesion" and "amplitude deviation is sensitive to limits".

[0085] The steps to input are: : By digital twin error prediction network The output represents the parameter disturbances during the manufacturing process of each amplitude rod; : By digital twin performance propagation module The calculated structural errors affect key performance indicators, including amplitude. ,frequency With modal stability changes ; The amplitude transformer performance data were obtained by actual measurement using a vibration table, interferometer, etc. The measurement frequency was 1kHz, and the steady-state characteristics were obtained after signal smoothing.

[0086] Furthermore, to improve the accuracy and reliability of performance documentation, this step incorporates the following two innovative mechanisms:

[0087] (1) Prediction-Measurement Difference Confidence Fusion Mechanism: Considering that the prediction model is unstable in terms of boundary structure and the measured results may be affected by environmental noise, this invention constructs a Bayesian fusion model with structural sensitivity weighting and adopts a structural weight function. Enhance the influence of key parameters on the fusion results.

[0088] (2) Multidimensional performance label structure embedding mechanism: introducing risk factors It is used to quantify the sensitivity of modal stability to structural disturbances, enabling early screening of potential adhesion risks.

[0089] Specifically, for individual amplitude transformers The fusion performance estimation model is as follows:

[0090] ;

[0091] This optimization model ensures that the final performance estimate is close to the measured value. With digital twin predictions The optimal weighted compromise is achieved between these factors.

[0092] The variables are explained as follows: Actual performance (e.g.) ); : The offset after propagation of structural errors predicted by the model; : Structural optimization target performance from S2 (i.e. ); The structure-sensitive weighting function is defined as follows:

[0093] ;

[0094] in It is a set of highly sensitive performance parameters (such as the tip angle, cross-sectional change rate, etc.). For each sensitive parameter, the risk coefficient is... To amplify the index (empirically set at 1.5), Control the weighting intensity.

[0095] The fusion output This refers to the final performance value of an individual. Here... This is the target amplitude used for subsequent dynamic control. : The resonant frequency used for subsequent dynamic adjustment; : Modal stability factor used for subsequent dynamic regulation.

[0096] Subsequently, modal risk factors were extracted based on the sensitivity of modal stability to end-point perturbations. :

[0097] ;

[0098] These are the key structural parameters for the cutter head output section; The second-order sensitivity of modal stability to structural perturbations is derived from the second-order derivative estimation of the trained performance model; This represents the actual error of the key structural parameters of the cutter head output section.

[0099] Ultimately, each individual amplitude transformer generates an individual performance label. :

[0100] ;

[0101] This tag contains: : A summary of structural parameters for fast indexing; Manufacturing error vector; Individual final performance value; Modal stability sensitive factor. Tags can be bound to the device body via RFID, laser coding, etc., stored in the database and simultaneously uploaded to the hospital management system, supporting batch tracking, surgical plan matching, and postoperative review.

[0102] The final output is: Individual performance labels include structural coding, manufacturing perturbations, fusion performance values, and modal risk factors; Risk factors that serve as the basis for adjusting the weights of control strategies.

[0103] S5. During the operation, tissue electrical impedance, blade temperature and ultrasonic echo signals are collected in real time. Combined with the individual performance tags, the transducer drive voltage is dynamically adjusted to perform closed-loop amplitude control.

[0104] Specifically, this step aims to achieve personalized dynamic control based on tissue feedback to compensate for performance deviations caused by prior manufacturing errors, ensuring cutting accuracy and tissue safety during surgery. This step is the core of the entire system's closed-loop execution, and its preceding output is the performance label of the individual amplitude transformer. Includes manufacturing disturbance parameters Fusion performance indicators (including estimated amplitude) and frequency Modal risk factors .

[0105] The system collects tissue feedback signals in real time during the operation. :

[0106] The electrical impedance of the blade in contact with the tissue is obtained through the piezoelectric sensing module;

[0107] Local temperature of the cutter head, acquired by an integrated infrared thermopile array;

[0108] : Morphological indicators of tissue feedback signals obtained from ultrasound echo analysis (such as echo delay time, energy distribution, etc.).

[0109] The system goal is to use a specific amplitude transformer. During the process, based on its individual performance parameters, a set of real-time dynamic control rules are designed to control the output amplitude. Adjustments were made to avoid the risks of excessive energy buildup and adhesion while maintaining the expected performance.

[0110] Furthermore, the present invention designs the following target control function:

[0111] ;

[0112] in, The actual output amplitude at the current moment (determined by the control voltage) (derived from the drive transducer). :Depend on The target amplitude obtained by fusion; Real-time temperature of the cutter head in a localized area; Thermal safety threshold (e.g., set to) C); , This indicates that for individuals with high modal risk, the proportion of temperature control constraints should be increased; among them This is the modal risk sensitivity factor output from the previous step. The higher the value, the more sensitive the amplitude rod is to structural disturbances.

[0113] Furthermore, this control function has the following innovative features:

[0114] Combined with individual performance prediction values and risk coefficient Construct personalized regulatory targets;

[0115] The design incorporates a risk-based adjustment mechanism to ensure that high-risk devices focus more on thermal control, while low-risk devices focus on performance compensation.

[0116] It has achieved the linkage of a three-layer data closed loop of manufacturing, prediction and execution.

[0117] Furthermore, based on loss The system uses a lookup table to determine the amplitude driving voltage. The update will be implemented, and the core control rules are as follows:

[0118] ;

[0119] The base voltage is determined by the initial setting; To control the response factor (values ​​ranging from 0.01 to 0.05, experimentally calibrated), the controller pre-calculates using a quick lookup table. To achieve high-speed response; The drive module acts on the transducer to actually adjust the ultrasonic energy output.

[0120] The control frequency is set to 20Hz, which can cover the response time of each action under typical surgery. All control actions are implemented in the device through a low-power MCU and sensor chip, and the control algorithm is deployed in the edge control module.

[0121] The steps to input are: The output from step three includes: , , ; Intraoperative real-time tissue feedback information ( , , This information is obtained from the sensor module. The step output is: : Output amplitude at the current moment; : The actual excitation voltage of the transducer, used to drive the amplitude transformer to generate the target amplitude.

[0122] Example 2

[0123] like Figure 2 As shown, this embodiment of the invention also provides an ultrasonic amplitude transformer optimization management system, the system comprising:

[0124] The parameter acquisition module 101 is used to receive the set of performance target parameters set during system initialization, including the target output amplitude multiplication factor, the target resonant frequency, and the target modal stability factor.

[0125] The amplitude transformer analysis module 102 is used to divide the amplitude transformer into an input segment, an intermediate transition segment, and an output segment to construct a structural design model, and to back-optimize the structural design model through a performance prediction module to generate a corresponding parameter set; wherein, the parameter set includes segment length, maximum / minimum diameter, cross-sectional function type, and end cone angle;

[0126] The parameter analysis module 103 is used to collect SLM process parameters, combine the parameter set with the pre-constructed error prediction module to generate manufacturing error prediction; and calculate the performance deviation caused by the error through a digital twin model to generate the performance deviation amount.

[0127] The parameter adjustment module 104 is used to collect and combine real-time amplitude bar performance data, combine the manufacturing error prediction and performance offset, generate individual final performance values ​​through Bayesian weighted fusion, and construct individual performance labels for each individual amplitude bar; wherein the individual performance labels include structural parameter summary, manufacturing error vector, individual final performance value and modal stability sensitivity factor.

[0128] The parameter optimization module 105 is used to collect tissue impedance, blade temperature and ultrasonic echo signals in real time during the operation, and dynamically adjust the transducer drive voltage in combination with the individual performance tag to perform closed-loop control of the amplitude.

[0129] The foregoing has described specific embodiments of this specification; other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than those shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily have to follow the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0130] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0131] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware components.

[0132] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0133] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0134] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0135] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0136] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0137] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0138] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0139] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0140] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0141] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0142] Finally, it should be noted that the ultrasonic amplitude transformer optimization management platform disclosed in the embodiments of the present invention is only a preferred embodiment of the present invention and is only used to illustrate the technical solutions of the present invention, not to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing the management of an ultrasonic amplitude transformer, characterized in that, The method includes the following steps: The set of performance target parameters set during the initialization of the receiving system includes the target output amplitude multiplication factor, the target resonant frequency, and the target modal stability factor; The amplitude transformer is divided into an input segment, an intermediate transition segment, and an output segment to construct a structural design model. The structural design model is then optimized in reverse by a performance prediction module to generate a corresponding parameter set. The parameter set includes segment length, maximum / minimum diameter, cross-sectional function type, and end cone angle. Collect SLM process parameters, combine the parameter set with the pre-built error prediction module to generate manufacturing error prediction; calculate the performance deviation caused by the error through a digital twin model to generate the performance deviation amount; Real-time amplitude transformer performance data is collected and combined with the manufacturing error prediction and performance offset. The final individual performance value is generated by Bayesian weighted fusion, and an individual performance label is constructed for each individual amplitude transformer. The individual performance label includes a structural parameter summary, a manufacturing error vector, the individual final performance value, and a modal stability sensitivity factor. During the procedure, tissue electrical impedance, blade temperature, and ultrasonic echo signals are collected in real time. Combined with the individual performance tags, the transducer drive voltage is dynamically adjusted to achieve closed-loop amplitude control. Specifically, the acquisition of real-time amplitude transformer performance data, combined with manufacturing error prediction and performance offset, is used to generate individual final performance values ​​through Bayesian weighted fusion, and an individual performance label is constructed for each individual amplitude transformer. Calculate digital twin predictions of the performance offset and the structural optimization target performance; wherein the structural optimization target performance is calculated based on the performance target parameter set. A fusion performance estimation model is designed to minimize the weighted trade-off between the digital twin prediction value and the real-time amplitude transformer performance data, thereby obtaining the individual final performance value; wherein the individual final performance value includes the final amplitude multiplication factor, the final resonant frequency, and the final modal stability factor. The modal stability sensitivity factor is calculated based on the parameter set of the output segment and the final modal stability factor. The individual performance label for each individual amplitude rod is generated by combining the structural parameter summary of the individual final performance value, the manufacturing error vector, the individual final performance value, and the modal stability sensitivity factor.

2. The ultrasonic amplitude transformer optimization management method according to claim 1, characterized in that, The structural design model is generated using a geometric generation function; wherein... The process of back-optimizing the structural design model through the performance prediction module to generate a corresponding parameter set is as follows: The parameters of the performance target parameter set are respectively input into the corresponding pre-built prediction model to generate the corresponding current prediction structure performance; The optimization objective of the prediction model is to minimize the overall performance deviation loss; wherein the overall performance deviation loss is obtained by calculating the deviation between the current performance target parameter set and the current prediction structure performance set. The initial structure search uses Latin hypercube sampling to generate 500 sets of structures, and then performs gradient descent-based optimization iterations, calling the prediction model to update the performance estimate one by one until the overall performance deviation loss is less than a set threshold or the number of iterations exceeds 300.

3. The ultrasonic amplitude transformer optimization management method according to claim 2, characterized in that, The pre-built prediction model is implemented by a regression network consisting of a set of multilayer perceptrons. Each performance index corresponds to one model. The model structure is a 4-layer perceptron with nodes in each layer of [64,128,128,1] and the activation function is ReLU.

4. The ultrasonic amplitude transformer optimization management method according to claim 1, characterized in that, The process involves collecting SLM process parameters, combining these parameters with a pre-built error prediction module, and generating a manufacturing error prediction. Specifically: The parameter set and SLM process parameters are input into the integrated error prediction network, and the original predicted manufacturing error is output. Combining the original predicted manufacturing error and the actual manufacturing error, the manufacturing error prediction is obtained by combining the objective function of weighted least squares method and structurally sensitive region constraints; wherein, the structurally sensitive region constraints are calculated based on the set of structural parameters most sensitive to performance; wherein the set of structural parameters most sensitive to performance includes the output end angle and the location of abrupt changes in cross-section; The objective function is used to minimize the actual manufacturing error.

5. The ultrasonic amplitude transformer optimization management method according to claim 4, characterized in that, The process of calculating the performance deviation caused by the error using a digital twin model and generating the performance deviation amount specifically involves: A performance disturbance prediction model is constructed based on the manufacturing error prediction and performance prediction module, and a correction is made based on the deviation correction term to obtain the performance offset, which includes amplitude difference, frequency difference and modal stability difference. The deviation correction term is calculated based on the output of the manufacturing error prediction and performance prediction modules and the performance target parameter set.

6. The ultrasonic amplitude transformer optimization management method according to claim 1, characterized in that, The individual performance tags are bound to the device body via RFID and laser coding, stored in the database, and uploaded synchronously.

7. The ultrasonic amplitude transformer optimization management method according to claim 1, characterized in that, The intraoperative real-time acquisition of tissue electrical impedance, blade temperature, and ultrasonic echo signals, combined with the individual performance tags, dynamically adjusts the transducer drive voltage to achieve closed-loop amplitude control, specifically as follows: Select any individual amplitude transformer as the target individual and collect tissue feedback signals in real time; wherein the tissue feedback signals include electrical impedance, local temperature of the cutting head, and morphological indicators of tissue feedback signals obtained by ultrasonic echo analysis. Based on the tissue feedback signal of the target individual, combined with the thermal safety threshold and the final performance value of the individual, a target control function is designed with the goal of minimizing the weighted sum of amplitude tracking error and temperature safety deviation. This function is used to adjust the output amplitude so as to avoid excessive energy accumulation and adhesion risk while maintaining the expected performance. According to the target control function, the controller updates the amplitude driving voltage by looking up a table to obtain the output voltage, which is used to regulate the ultrasonic energy output.

8. The ultrasonic amplitude transformer optimization management method according to claim 7, characterized in that, The control frequency is set to 20Hz to cover the response time of each action under typical surgery.

9. A system for implementing the ultrasonic amplitude transformer optimization management method according to claim 1, characterized in that, The system includes: The parameter acquisition module is used to receive the set of performance target parameters set during system initialization, including the target output amplitude multiplication factor, the target resonant frequency, and the target modal stability factor. The amplitude transformer analysis module is used to divide the amplitude transformer into an input segment, an intermediate transition segment, and an output segment to construct a structural design model. The performance prediction module then back-optimizes the structural design model to generate a corresponding parameter set. The parameter set includes segment length, maximum / minimum diameter, cross-sectional function type, and end cone angle. The parameter analysis module is used to collect SLM process parameters, combine the parameter set with the pre-built error prediction module to generate manufacturing error prediction; and calculate the performance deviation caused by the error through a digital twin model to generate the performance deviation amount. The parameter adjustment module is used to collect and combine real-time amplitude bar performance data, and combine the manufacturing error prediction and performance offset to generate individual final performance values ​​through Bayesian weighted fusion, and construct individual performance labels for each individual amplitude bar; wherein the individual performance labels include structural parameter summary, manufacturing error vector, individual final performance value and modal stability sensitivity factor; The parameter optimization module is used to collect tissue impedance, blade temperature and ultrasonic echo signals in real time during the operation. Combined with the individual performance tag, it dynamically adjusts the transducer drive voltage to perform closed-loop amplitude control.

Images