Ultrasonic cutter parameter optimization method for frozen tissue section

Abstract

The invention provides an ultrasonic cutter parameter optimization method for frozen tissue slicing, and belongs to the technical field of medical instruments. The ultrasonic cutter parameter optimization method for the frozen tissue section comprises the following steps: collecting vibration frequency, amplitude, feeding speed, load power and cut tissue type in a frozen tissue cutting process of an ultrasonic cutter as an initial parameter group, and establishing a mapping relation between the initial parameter group and frozen tissue temperature change; calculating a contribution value of the initial parameter group, comparing the contribution value with a preset threshold value, determining the parameter group greater than the preset threshold value as a key parameter group, and determining the parameter group less than or equal to the preset threshold value as a non-key parameter group; by means of the ultrasonic knife, the problem that an existing ultrasonic knife cannot achieve accurate control over temperature changes in the tissue slicing process can be solved.

Description

Technical Field

[0001] The invention belongs to the technical field of medical devices, and in particular relates to a method for optimizing ultrasonic tool parameters for frozen tissue sections. Background Art

[0002] Frozen tissue sectioning technology has important applications in the fields of pathological diagnosis and scientific research. Among them, ultrasonic cutters are widely used because of their advantages such as high cutting accuracy and little tissue damage. Ultrasonic cutters cut tissues through mechanical energy generated by high-frequency vibrations, which will generate a certain amount of heat during the cutting process. For conventional tissue sections, this heat generation does not cause a significant impact. However, during the frozen tissue sectioning process, since the sectioned specimen is in a low temperature state, the heat generated by the ultrasonic knife may cause the local tissue temperature to rise, causing the frozen specimen to melt, thereby affecting the sectioning quality and subsequent analysis results.

[0003] In the prior art, the following methods are usually used to control temperature changes during frozen tissue sectioning: first, by adjusting the working parameters of the ultrasonic tool, such as reducing the vibration frequency or reducing the amplitude to reduce heat generation; second, using an external refrigeration system to offset the heat generated by ultrasonic cutting through continuous cooling; third, optimizing the cutting process parameters, such as adjusting the feed speed and cutting depth to control heat accumulation. However, these methods have obvious limitations. Simply reducing the working parameters of the ultrasonic tool will lead to a decrease in cutting efficiency and affect surgical efficiency; the external refrigeration system is difficult to achieve precise temperature control of the cutting area and may cause excessive cooling of the tissue; and the empirical process parameter optimization lacks a scientific basis and is difficult to adapt to the cutting needs of different types of tissues.

[0004] In addition, the existing technology mostly uses single parameter adjustment or simple orthogonal test methods to optimize ultrasonic tool parameters, without considering the coupling relationship between parameters and the comprehensive impact on temperature changes. At the same time, there is a lack of real-time monitoring and feedback adjustment mechanism for the parameter optimization process, resulting in poor reliability and adaptability of the optimization results. In practical applications, different types of frozen tissues have different sensitivities to temperature changes. Existing parameter optimization methods are difficult to accurately control specific tissue characteristics, and are prone to problems such as local overheating or low cutting efficiency. Summary of the invention

[0005] In view of this, the present invention provides a method for optimizing ultrasonic tool parameters for frozen tissue slicing, which can solve the problem that the existing ultrasonic knife cannot achieve accurate control of temperature changes during tissue slicing.

[0006] The present invention is achieved in that:

[0007] The present invention provides a method for optimizing ultrasonic tool parameters for frozen tissue sections, which includes the following steps: collecting the vibration frequency, amplitude, feed speed, load power and cutting tissue type of the ultrasonic tool in the process of cutting frozen tissue as an initial parameter group, and establishing a mapping relationship between the initial parameter group and the temperature change of the frozen tissue; calculating the contribution value of the initial parameter group and comparing the contribution value with a preset threshold, determining the parameter group greater than the preset threshold as a key parameter group, and determining the parameter group less than or equal to the preset threshold as a non-key parameter group; establishing an influence value model for the key parameter group and determining a parameter adjustment sequence, and optimizing the parameters according to the parameter adjustment sequence until the temperature change of the frozen tissue meets the preset temperature range requirement.

[0008] On the basis of the above technical solution, the ultrasonic tool parameter optimization method for frozen tissue sections of the present invention can also be improved as follows:

[0009] Among them, the step of collecting the initial parameter group is specifically to use a frequency sensor to collect the vibration frequency, the sampling frequency is 1000Hz, and the collection range of the vibration frequency is 20kHz to 40kHz; use a laser displacement sensor to collect the amplitude, the measurement range is 0 microns to 100 microns, and the measurement accuracy is 0.1 microns; use a photoelectric encoder to collect the feed speed, the resolution is 0.1 mm per second; use a Hall effect power meter to collect the load power, the measurement range is 0 watts to 1000 watts, and the measurement accuracy is 1 watt.

[0010] Furthermore, the step of establishing a mapping relationship between the initial parameter group and the temperature change of frozen tissue specifically adopts a radial basis function neural network method, using the initial parameter group as the input layer, the temperature change as the output layer, the number of hidden layer nodes is determined by a cross-validation method, the activation function is a Gaussian function, the learning rate is 0.01, and the training termination condition is that the mean square error is less than 0.001 or the maximum number of iterations is 1000.

[0011] Furthermore, the step of calculating the contribution value is to normalize the initial parameter group to the maximum and minimum values, map all parameter values ​​to between 0 and 1, and then use the principal component analysis method combined with the expert scoring method to determine the weight coefficient. The principal component contribution rate threshold is 85%. The contribution value calculation introduces a random error term that obeys the normal distribution, with a mean of 0 and a standard deviation of 0.05.

[0012] Furthermore, the steps of establishing the influence value model are specifically to use the five-point numerical difference method to calculate the partial derivatives of temperature changes with respect to each parameter, with a step size of 0.1% of the parameter range; calculate the coefficient of fluctuation of the parameter, with a sample size of not less than 30; use the Pearson correlation coefficient to calculate the correlation between the parameter and temperature change; use the hierarchical analysis method combined with the grey correlation analysis to determine the weight coefficient, with a consistency ratio of less than 0.1.

[0013] Furthermore, the step of determining the parameter adjustment order specifically constructs a parameter importance matrix, uses an improved power iteration method to calculate the eigenvector of the matrix, selects an all-1 vector as the initial vector, has an iteration accuracy of 0.001, and has a maximum number of iterations of 1000 times; and uses a fuzzy cognitive map method to establish a parameter association network with an association strength threshold of 0.3, which is used to adjust the final adjustment order.

[0014] Furthermore, the parameter optimization step specifically adopts an improved particle swarm optimization algorithm to adjust the parameters, with a population size of 50, a maximum number of iterations of 200, a linear decrease in inertia weight from 0.9 to 0.4, and a learning factor of 2.0; an adaptive mutation strategy is adopted, the mutation probability is inversely proportional to the fitness value, and the basic mutation probability is 0.1.

[0015] Furthermore, it also includes the step of establishing a parameter optimization database, specifically using a relational database management system to establish the database, designing a three-layer table structure of a basic parameter table, an optimization result table, and a related information table, maintaining data consistency through primary and foreign key relationships; establishing a composite index for commonly used query fields; and adopting an incremental backup strategy, performing incremental backups every day and full backups once a week.

[0016] Furthermore, it also includes the step of verifying the suitability of parameters, specifically using an infrared temperature sensor array to monitor the temperature distribution of frozen tissue in real time, with a sensor spacing of 5 mm and a sampling frequency of 10 Hz; using a Kalman filter for noise filtering; and using a sliding window method to calculate the temperature change trend, with a window length of 10 seconds, a step size of 1 second, and a temperature tolerance of 0.5 degrees Celsius.

[0017] It also includes the step of optimizing the result judgment, specifically, using the fuzzy comprehensive evaluation method to establish an evaluation index system, including the maximum over-limit value, the over-limit time ratio and the temperature fluctuation intensity, and using the fuzzy hierarchical analysis method to determine the index weight; using the annealing algorithm to adaptively adjust the weight coefficient, with the initial temperature being 100, the cooling coefficient being 0.95, and the termination temperature being 0.01.

[0018] Compared with the prior art, the ultrasonic tool parameter optimization method for frozen tissue sectioning provided by the present invention has the following beneficial effects: the ultrasonic tool parameter optimization method proposed by the present invention realizes the precise control of temperature changes in the frozen tissue sectioning process by establishing a multi-parameter coupled optimization model. First, by introducing two evaluation indicators, contribution value and influence value, a quantitative evaluation system of parameter importance is established, making the parameter optimization process more scientific and controllable. Among them, the contribution value reflects the comprehensive influence of the parameter group on the temperature change, while the influence value quantifies the independent effect of a single parameter. This dual evaluation mechanism can more comprehensively grasp the direction of parameter optimization.

[0019] Secondly, an improved intelligent optimization algorithm is used for parameter adjustment. The convergence efficiency of the optimization process and the quality of the solution are improved through adaptive mutation strategy and dynamic weight adjustment. At the same time, a complete parameter optimization database is established to achieve the accumulation and reuse of optimization experience, providing data support for the optimization of slice parameters of different types of tissues.

[0020] In addition, the present invention designs a real-time monitoring system based on an infrared temperature measurement array and a feedback mechanism of fuzzy comprehensive evaluation, which can timely detect and correct temperature anomalies and ensure the reliability and stability of the optimization results. This closed-loop control strategy effectively solves the problem that the parameter optimization results in traditional methods are difficult to verify and dynamically adjust. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flow chart of a method for optimizing ultrasonic tool parameters for frozen tissue sections. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0023] like Figure 1 FIG. 1 is a flow chart of a method for optimizing ultrasonic tool parameters for frozen tissue sections provided by the present invention, comprising the following steps:

[0024] S01, collecting the vibration frequency, amplitude, feed speed, load power and cutting tissue type of the ultrasonic cutter in the process of cutting frozen tissue as an initial parameter group, and establishing a mapping relationship between the initial parameter group and the temperature change of the frozen tissue;

[0025] S02. Calculating a contribution value of each group of parameters based on the initial parameter group, wherein the contribution value is determined by the degree of influence of each parameter in the initial parameter group on the temperature change of the frozen tissue;

[0026] S03, comparing the contribution value with a preset threshold value, if the contribution value is greater than the preset threshold value, determining it as a key parameter group and executing step S04; if the contribution value is less than or equal to the preset threshold value, determining it as a non-key parameter group and executing step S08;

[0027] S04, establishing an influence value model for each parameter in the key parameter group, wherein the influence value is used to characterize the independent influence degree of a single parameter on the temperature change of frozen tissue;

[0028] S05. Prioritize the key parameter groups based on the impact values ​​to determine a parameter adjustment order;

[0029] S06. According to the parameter adjustment sequence, adjust the parameter values ​​in the key parameter group until the temperature change of the frozen tissue meets the preset temperature range requirement;

[0030] S07, recording the optimized values ​​of the key parameter group, and establishing a parameter optimization database for frozen tissue types;

[0031] S08, collecting parameter values ​​of the non-critical parameter group and storing them in the parameter optimization database as optimization reference data;

[0032] S09, performing ultrasonic tool parameter matching for frozen tissue section tasks based on the parameter optimization database;

[0033] S10, acquiring real-time temperature data of the frozen tissue section task, and verifying the applicability of the parameters in the parameter optimization database;

[0034] S11. If the real-time temperature data exceeds the preset temperature range, execute step S02 to re-optimize the parameters; if the real-time temperature data is within the preset temperature range, confirm that the parameter optimization is completed.

[0035] The specific implementation method of step S01 is to collect the working parameters of the ultrasonic tool during the process of cutting frozen tissue. First, determine the type of parameters that need to be collected, including vibration frequency, amplitude, feed speed, load power and cutting tissue type. The collection of vibration frequency uses a frequency sensor to monitor the working frequency of the ultrasonic tool in real time, the sampling frequency is set to 1000Hz, and the sampling time is the entire cutting process. The amplitude collection uses a displacement sensor to measure the displacement of the vibration end of the ultrasonic tool, and the sampling frequency is synchronized with the frequency sensor. The feed speed obtains the movement speed of the ultrasonic tool through an encoder and records the speed change in real time. The load power measures the real-time power consumption of the ultrasonic tool through a power meter. The cutting tissue type is determined by pre-marking. Then, a mapping relationship between these parameters and the temperature change of frozen tissue is established, and a multivariate nonlinear regression method is used to use the collected parameters as independent variables and the temperature change as the dependent variable to establish a mapping model. The purpose of this step is to obtain complete parameter information during the working process of the ultrasonic tool to provide a data basis for subsequent optimization. High-precision sensors and synchronous acquisition technology are used to ensure the accuracy and timeliness of the data. The establishment of the mapping relationship adopts a nonlinear regression method, which can better reflect the complex relationship between parameters and temperature changes.

[0036] The specific implementation method of step S02 is to calculate the parameter contribution value. First, the collected parameters are normalized to eliminate the dimensional differences between different parameters. Parameter normalization adopts the maximum and minimum normalization method to map all parameter values ​​to between 0 and 1. Then the contribution value of each group of parameters to the temperature change is calculated, and the weight coefficient is determined by the principal component analysis method using a weighted summation method. For the tissue type characteristic value, the temperature change rate of different tissues under the same cutting parameters is obtained through a comparative experiment and normalized. This step introduces a random error term to characterize the uncertainty of the model, and the value range of the error term is 0.01 to 0.1. The purpose of this step is to quantify the degree of influence of each group of parameters on the temperature change and provide a basis for parameter classification. The principal component analysis method is used to determine the weight coefficient, which can objectively reflect the relative importance of each parameter, and the normalization process ensures the comparability between different parameters.

[0037] The specific implementation method of step S03 is to perform parameter classification. First, a preset threshold is set, which is determined by statistical analysis of a large amount of experimental data, and the recommended value is 0.6. The calculated parameter contribution value is compared with the preset threshold, and the parameter group greater than the threshold is determined as the key parameter group, and the parameter group less than or equal to the threshold is determined as the non-key parameter group. The key parameter group will undergo an in-depth impact analysis, and the non-key parameter group will serve as auxiliary reference data. The purpose of this step is to identify the parameter group that has a significant impact on temperature changes and improve optimization efficiency. The setting of the threshold is based on the statistical analysis of a large amount of experimental data and has strong universality.

[0038] The specific implementation method of step S04 is to establish an influence value model. First, the partial derivatives of temperature change on each parameter are calculated, and the numerical difference method is used to select an appropriate step size for calculation. Then the coefficient of fluctuation of the parameter is calculated, and the stability of the parameter is characterized by the ratio of the standard deviation to the mean. Then the correlation coefficient between the parameter and the temperature change is calculated, and the Pearson correlation coefficient method is used. Finally, these three indicators are weighted and combined to obtain the influence value. The weight coefficient is determined by the fuzzy hierarchical analysis method, taking into account expert experience and historical data. This step introduces a random error term with a value range of 0.05 to 0.15. The purpose of this step is to comprehensively evaluate the influence of a single parameter on temperature change. The use of the fuzzy hierarchical analysis method can combine subjective experience and objective data to improve the scientific nature of the evaluation.

[0039] The specific implementation method of step S05 is to determine the order of parameter adjustment. First, a parameter importance matrix is ​​constructed, and the matrix elements are calculated by combining the influence value and the contribution value. Then the eigenvector of the matrix is ​​calculated, and the power iteration method is used, and the iteration accuracy is set to 0.001. The size of the component of the eigenvector reflects the relative importance of the parameter. Finally, the order of parameter adjustment is determined according to the order of importance from large to small. The purpose of this step is to formulate a scientific parameter optimization strategy. The use of the power iteration method improves the efficiency of eigenvector calculation, and the eigenvector can objectively reflect the relative importance of the parameter.

[0040] The specific implementation method of step S06 is to adjust the parameter values. First, set a preset temperature range. For common frozen tissues, it is recommended that the temperature change range be controlled within minus 5 degrees Celsius. Then adjust the parameter values ​​one by one according to the determined adjustment order. The parameter adjustment adopts an adaptive step search algorithm. The initial step size is set to 5% of the parameter range, and the step size is dynamically adjusted according to the temperature change trend. When the temperature change meets the preset range requirements, the current parameter value is recorded. The purpose of this step is to find a parameter combination that meets the temperature control requirements. The adaptive step search algorithm can avoid excessive parameter adjustment while ensuring search efficiency.

[0041] The specific implementation method of step S07 is to establish a parameter optimization database. First, record the values ​​of the optimized key parameter group, including the final values ​​of parameters such as vibration frequency, amplitude, and feed speed. Then establish the database table structure, design the parameter index, and adopt a relational database management system. In addition to storing parameter values, the database also contains relevant information such as tissue type and temperature change. The purpose of this step is to accumulate optimization experience and provide a reference for subsequent similar tasks. The use of relational databases facilitates data storage and query, and the design of parameter indexes improves data retrieval efficiency.

[0042] The specific implementation method of step S08 is to store non-critical parameter data. First, the parameter values ​​of the non-critical parameter group are collected, and the data format is the same as that of the critical parameter group. Then these data are stored in the parameter optimization database as optimization reference data, and an association relationship with the critical parameter group is established. The purpose of this step is to improve the database information and provide a more comprehensive reference basis for parameter optimization. The establishment of data association helps to understand the mutual influence between parameters.

[0043] The specific implementation method of step S09 is to perform parameter matching. First, obtain the basic information of the new frozen tissue section task, including tissue type, target section thickness, etc. Then retrieve similar cases in the parameter optimization database, and use a case-based reasoning method to calculate the case similarity. The similarity calculation takes into account multiple factors such as tissue type and section requirements, and uses a weighted Euclidean distance metric. Finally, select the parameter combination corresponding to the case with the highest similarity. The purpose of this step is to quickly determine the parameter combination suitable for the new task. The case-based reasoning method utilizes historical experience to improve the efficiency and reliability of parameter matching.

[0044] The specific implementation of step S10 is to verify the applicability of parameters. First, a temperature sensor is used to monitor the temperature change of frozen tissue in real time, and the sampling frequency is set to 10Hz. Then the measured real-time temperature data is compared with the preset temperature range to determine whether the parameter combination is applicable. The purpose of this step is to ensure the effectiveness of parameter optimization. Real-time temperature monitoring can detect problems in time and ensure the quality of slices.

[0045] The specific implementation of step S11 is to judge the optimization result. First, determine whether the real-time temperature data exceeds the preset temperature range. If it exceeds, return to step S02 to re-optimize the parameters; if it is within the range, confirm that the parameter optimization is completed. When re-optimization is required, the calculation weight of the parameter contribution value can be adjusted according to the degree to which the temperature exceeds the range to improve the optimization effect. The purpose of this step is to ensure that the parameter combination that meets the requirements is finally obtained. Through the feedback adjustment mechanism, the optimization results are continuously improved and the applicability of the method is improved.

[0046] The specific implementation methods of the above steps are described in detail below:

[0047] The specific implementation method of step S01 is to collect the working parameters of the ultrasonic tool in the process of cutting frozen tissue. First, determine the parameter types that need to be collected, including vibration frequency, amplitude, feed speed, load power and cutting tissue type, wherein the collection of vibration frequency uses a piezoelectric frequency sensor to monitor the working frequency of the ultrasonic tool in real time, the sampling frequency is set to 1000Hz to ensure sampling accuracy, the sampling time covers the entire cutting process, and the frequency collection range is 20kHz to 40kHz. Amplitude collection uses a laser displacement sensor to measure the displacement of the ultrasonic tool vibration end, the sampling frequency is synchronized with the frequency sensor, the amplitude measurement range is 0 micron to 100 microns, and the measurement accuracy is 0.1 micron. The feed speed obtains the movement speed of the ultrasonic tool through a photoelectric encoder, with a resolution of 0.1 mm per second, and the speed change is recorded in real time. The load power measures the real-time power consumption of the ultrasonic tool through a Hall effect power meter, with a measurement range of 0 watts to 1000 watts and a measurement accuracy of 1 watt. The cutting tissue type is determined by pre-marking, and a tissue type coding table is established. Then, the mapping relationship between these parameters and the temperature change of frozen tissue is established. The improved radial basis function neural network method is used. The collected parameters are used as the input layer, the temperature change is used as the output layer, the number of hidden layer nodes is determined by the cross-validation method, the activation function is the Gaussian function, the learning rate is set to 0.01, and the training termination condition is that the mean square error is less than 0.001 or the maximum number of iterations is 1000. The purpose of this step is to obtain complete parameter information during the working process of the ultrasonic tool to provide a data basis for subsequent optimization. High-precision sensors and synchronous acquisition technology are used to ensure the accuracy and timeliness of the data. The radial basis function neural network method is used to establish the mapping relationship, which can better reflect the nonlinear relationship between parameters and temperature changes.

[0048] The specific implementation of step S02 is to calculate the parameter contribution value. First, the collected parameters are normalized to eliminate the dimension differences between different parameters. The parameter normalization adopts the maximum and minimum value normalization method to map all parameter values ​​between 0 and 1. The normalization calculation formula adopts the F in the above formula. i , A i 、V i , P i Calculation method. Then calculate the contribution of each group of parameters to the temperature change by weighted summation. The weight coefficients α1 to α5 are determined by principal component analysis combined with expert scoring. The principal component contribution rate threshold is set to 85%. For the tissue type characteristic value T i, the temperature change rate of different tissues under the same cutting parameters was obtained through comparative experiments. The experimental temperature range was minus 40 degrees Celsius to minus 20 degrees Celsius. The temperature change rate adopted the average change rate within 5 minutes and was normalized. This step introduces a random error term ∈ 1, which obeys the normal distribution, has a mean of 0 and a standard deviation of 0.05, and is used to characterize the uncertainty of the model. The purpose of this step is to quantify the degree of influence of each group of parameters on temperature change and provide a basis for parameter classification. The principal component analysis method combined with the expert scoring method is used to determine the weight coefficient, which can combine objective data analysis with subjective experience, improve the rationality of weight determination, and normalization ensures the comparability between different parameters.

[0049] The specific implementation method of step S03 is to perform parameter classification. First, a preset threshold is set, which is determined by the Monte Carlo simulation method, and the number of simulations is 10,000. In each simulation, a parameter combination is randomly generated, and the corresponding temperature change is calculated. The threshold value obtained by statistical analysis is 0.6. The calculated parameter contribution value is compared with the preset threshold value, and the parameter group greater than the threshold value is determined as the key parameter group, and the parameter group less than or equal to the threshold value is determined as the non-key parameter group. The key parameter group will undergo an in-depth impact analysis, and the non-key parameter group will serve as auxiliary reference data. In the parameter classification process, a fuzzy clustering method is used for auxiliary judgment, and the number of clusters is set to 2, the membership threshold is 0.8, and the iteration termination condition is that the membership change is less than 0.001 or reaches the maximum number of iterations 100 times. The purpose of this step is to identify the parameter group that has a significant impact on the temperature change and improve the optimization efficiency. The setting of the threshold combines the Monte Carlo simulation and the fuzzy clustering method to improve the reliability of the classification.

[0050] The specific implementation of step S04 is to establish an influence value model. According to the given formula, first calculate the partial derivative of temperature change to each parameter The five-point numerical difference method is used, and the step size is set to 0.1% of the parameter range to improve the accuracy of derivative calculation. Then the fluctuation coefficient σ of the parameter is calculated j , using the given formula, the stability of the parameter is characterized by the ratio of the standard deviation to the mean, and the sample size is not less than 30. Then calculate the correlation coefficient ρ between the parameter and the temperature change j , using the Pearson correlation coefficient calculation formula. Finally, the three indicators are weighted and combined, and the weight coefficients β1, β2, and β3 are determined by the hierarchical analysis method, combined with the grey correlation analysis, to construct the judgment matrix, and the consistency ratio is required to be less than 0.1. This step introduces the random error term ∈2, which obeys the normal distribution, has a mean of 0, and a standard deviation of 0.1. The purpose of this step is to comprehensively evaluate the influence of a single parameter on temperature change. Combining numerical difference, statistical analysis and grey system theory, the accuracy and reliability of the impact assessment are improved.

[0051] The specific implementation of step S05 is to determine the parameter adjustment order. According to the given matrix formula, the parameter importance matrix R is first constructed. The matrix element r ij The balance coefficient γ is calculated by combining the influence value and the contribution value, and the value is 0.5. Then the improved power iteration method is used to calculate the eigenvector of the matrix. The initial vector selects the all-1 vector, the iteration accuracy is set to 0.001, and the maximum number of iterations is 1000. The size of the eigenvector component reflects the relative importance of the parameter. Finally, the quick sorting algorithm is used to sort the eigenvector components from large to small to determine the parameter adjustment order. In the sorting process, the coupling relationship between the parameters is considered, and the fuzzy cognitive map method is used to establish a parameter association network. The association strength threshold is set to 0.3 to adjust the final adjustment order. The purpose of this step is to formulate a scientific parameter optimization strategy. The improved power iteration method improves the efficiency of eigenvector calculation, and the introduction of the fuzzy cognitive map takes into account the mutual influence between the parameters.

[0052] The specific implementation of step S06 is to adjust the parameter value. First, a preset temperature range is set. For different types of frozen tissues, the temperature change range is set separately: adipose tissue is minus 5 degrees Celsius to minus 3 degrees Celsius, muscle tissue is minus 4 degrees Celsius to minus 2 degrees Celsius, and connective tissue is minus 6 degrees Celsius to minus 4 degrees Celsius. Then, according to the determined adjustment order, the improved particle swarm optimization algorithm is used to adjust the parameters. The population size is set to 50, the maximum number of iterations is 200, the inertia weight adopts a linear decreasing strategy, decreasing from 0.9 to 0.4, and the learning factor is set to 2.0 and 2.0 respectively. In order to improve the search efficiency, the parameter adjustment range is discretized, and the step size is set to 5% of the parameter range. In the optimization process, an adaptive mutation strategy is adopted, and the mutation probability is inversely proportional to the fitness value, and the basic mutation probability is set to 0.1. When the temperature change meets the preset range requirements, the current parameter value is recorded. The purpose of this step is to find a parameter combination that meets the temperature control requirements. The improved particle swarm algorithm combines the adaptive strategy to improve the efficiency and accuracy of parameter optimization.

[0053] The specific implementation method of step S07 is to establish a parameter optimization database. First, a relational database management system is used to establish a database, and a three-layer table structure is designed: a basic parameter table, an optimization result table, and an associated information table. The basic parameter table contains fields such as parameter number, parameter type, and parameter range. The optimization result table stores the optimized parameter values ​​and their effects, and the associated information table records the association between parameters. Data consistency is maintained between tables through primary and foreign key relationships. Then an index mechanism is designed to establish a composite index for commonly used query fields to improve query efficiency. The database fields are standardized in design, follow the third normal form, and avoid data redundancy. When storing data, an incremental backup strategy is adopted, incremental backup is performed every day, and a full backup is performed once a week to ensure data security. The purpose of this step is to build a systematic parameter optimization knowledge base to provide data support for subsequent optimization. The standardized design and perfect backup mechanism of the relational database ensure the reliability and maintainability of the data.

[0054] The specific implementation method of step S08 is to store non-critical parameter data. First, the non-critical parameter data is standardized and the Z value standardization method is adopted to ensure the comparability of the data. Then the processed data is stored in the parameter optimization database according to the predefined data format, and the data format includes information such as parameter value, acquisition time, and environmental conditions. During the data storage process, a batch insertion strategy is adopted to improve storage efficiency. At the same time, a version control mechanism for parameter data is established to record the change history of data. The stored data is quality evaluated, including integrity check, outlier detection and consistency verification, and the evaluation results are used as the basis for data availability. The purpose of this step is to save the complete parameter optimization history information and provide a comprehensive reference basis for parameter optimization. Standardization and quality control mechanisms ensure the availability of data, and version control facilitates tracking of data changes.

[0055] The specific implementation method of step S09 is to perform parameter matching. First, a parameter matching model based on case reasoning is established, and the case features include tissue type, target slice thickness, ambient temperature, etc. The feature weight is determined by genetic algorithm optimization, with a population size of 100, a crossover probability of 0.8, a mutation probability of 0.1, and a maximum evolutionary generation of 200. Then the case similarity is calculated, and the weighted Mahalanobis distance metric is used to consider the correlation between features. A time decay factor is introduced in the similarity calculation to reduce the influence of old cases. When retrieving similar cases in the parameter optimization database, a multi-level indexing strategy is adopted, firstly a rough screening is performed by tissue type, and then an accurate matching is performed based on similarity. Finally, the top three cases with the highest similarity are selected, and the final parameter combination is determined by weighted average, and the weight is proportional to the similarity. The purpose of this step is to achieve intelligent parameter matching. The case-based reasoning method combined with genetic algorithm optimization improves the accuracy and efficiency of parameter matching.

[0056] The specific implementation method of step S10 is to verify the applicability of parameters. First, an infrared temperature sensor array is used to monitor the temperature distribution of frozen tissue in real time. The sensor spacing is 5 mm, covering the entire cutting area, and the sampling frequency is 10 Hz. The temperature measurement data is filtered by a Kalman filter, and the state noise covariance matrix and the observation noise covariance matrix are determined by experimental calibration. Then the spatiotemporal distribution characteristics of the temperature field are calculated, including the average temperature, temperature gradient and temperature fluctuation rate. The sliding window method is used to calculate the temperature change trend, with a window length of 10 seconds and a step length of 1 second. The calculation result is fuzzy compared with the preset temperature range, and the temperature tolerance is set to 0.5 degrees Celsius. The purpose of this step is to ensure the effectiveness of parameter optimization. The infrared temperature measurement array combined with Kalman filtering provides an accurate temperature monitoring method, and the sliding window analysis helps to grasp the temperature change trend.

[0057] The specific implementation method of step S11 is to judge the optimization results. First, a temperature over-limit assessment model is established, and the real-time temperature data is compared with the preset temperature range to calculate the degree of over-limit and duration. The fuzzy comprehensive evaluation method is adopted to establish an evaluation index system, including the maximum over-limit value, the over-limit time ratio and the temperature fluctuation intensity, and the index weight is determined by the fuzzy hierarchical analysis method. Then, according to the evaluation results, it is decided whether re-optimization is needed. When re-optimization is needed, the weight coefficient in the parameter contribution value calculation is adaptively adjusted according to the temperature over-limit situation. The adjustment adopts the annealing algorithm, with the initial temperature of 100, the cooling coefficient of 0.95, and the termination temperature of 0.01. At the same time, the optimization results and their effects are stored in the parameter optimization database for the accumulation of experience for subsequent optimization.

[0058] The calculation process or equation involved in the present invention is described in detail below.

[0059] 1. Contribution value calculation process in step S02:

[0060] The calculation of the contribution value is specifically expressed as follows:

[0061] C i =α1F i +α2A i +α3V i +α4P i +α5T i +∈1;

[0062] In the formula, C i is the contribution value of the i-th group of parameters; F i is the normalized value of vibration frequency; A i is the amplitude normalized value; V i is the normalized value of feed speed; P i is the normalized value of load power; T iis the characteristic value of tissue type; α1, α2, α3, α4, α5 are the weight coefficients to be determined; ε1 is the random error term.

[0063] The parameter acquisition method is:

[0064] 1) Normalized value calculation:

[0065]

[0066]

[0067] In the formula, f i ,a i ,v i ,p i are the measured parameter values ​​respectively; f max ,a max ,v max ,p max is the maximum value of the parameter; f min ,a min ,v min ,p min is the minimum value of the parameter.

[0068] 2) The tissue type characteristic values ​​are obtained experimentally:

[0069] Step 1: Conduct ultrasonic cutting experiments on different types of tissues;

[0070] Step 2: Record the temperature change rate of different tissues under the same cutting parameters;

[0071] Step 3: Normalize the temperature change rate to get T i value.

[0072] 2. Influence value model in step S04:

[0073] The influence value model is specifically expressed as follows:

[0074]

[0075] In the formula, I j is the influence value of the jth parameter; is the partial derivative of temperature change with respect to the parameter; σ j is the parameter fluctuation coefficient; ρ j is the correlation coefficient between the parameter and temperature change; β1, β2, β3 are the weight coefficients to be determined; ∈2 is the random error term.

[0076] The parameter acquisition method is:

[0077] 1) Partial derivative calculation:

[0078]

[0079] In the formula, T(x j ) is the temperature change function; h is the small change.

[0080] 2) Calculation of volatility coefficient:

[0081]

[0082] In the formula, s j is the parameter standard deviation; is the parameter mean.

[0083] 3) Correlation coefficient calculation:

[0084]

[0085] 3. Priority sorting in step S05:

[0086] The priority ranking matrix is ​​specifically expressed as follows:

[0087]

[0088] In the formula, r ij Indicates the importance of the i-th parameter relative to the j-th parameter; n is the number of parameters.

[0089] Parameter importance calculation:

[0090]

[0091] In the formula, I i ,I j are the influence values ​​of parameters i and j respectively; C i ,C j are the contribution values ​​of parameters i and j respectively; γ is the balance coefficient.

[0092] These equations are constructed as follows:

[0093] 1. The contribution value is calculated using a linear weighted model, taking into account the relative importance of each parameter and eliminating the dimension effect through normalization;

[0094] 2. The impact value model comprehensively considers parameter sensitivity (partial derivative term), stability (fluctuation coefficient term) and correlation (correlation coefficient term), and can comprehensively evaluate the impact of parameters;

[0095] 3. The priority ranking matrix adopts the idea of ​​hierarchical analysis method, determines the relative importance by comparing parameters pairwise, and conducts a comprehensive evaluation based on the impact value and contribution value.

[0096] Advantages over existing technologies:

[0097] 1. The introduction of dual evaluation indicators of contribution value and influence value improves the accuracy of parameter importance assessment;

[0098] 2. Taking into account parameter volatility and correlation, the stability of the optimization process is enhanced;

[0099] 3. Use matrix form to sort priorities and make the decision-making process more scientific and systematic.

[0100] Specifically, the principle of the present invention is: the core principle of the scheme of the present invention is to establish a complete parameter optimization framework, which transforms the temperature control problem in the ultrasonic tool cutting process into a multi-objective optimization problem. By collecting working parameters such as vibration frequency, amplitude, and feed speed, a mapping relationship between parameters and temperature changes is established. This mapping relationship is modeled using a radial basis function neural network, which can effectively capture the nonlinear interaction between parameters.

[0101] In the parameter evaluation stage, the contribution value calculation takes into account the overall effect of the parameter group, and the weight coefficient is determined by the principal component analysis method, which can objectively reflect the relative importance of each parameter. The influence value model comprehensively evaluates the influence characteristics of a single parameter on temperature change through the combination of partial derivatives, fluctuation coefficients and correlation coefficients. This multi-dimensional parameter evaluation method conforms to the physical nature of the role of parameters in the ultrasonic cutting process.

[0102] The parameter optimization process uses an improved particle swarm algorithm to find the optimal parameter combination through population iteration. The adaptive mutation mechanism introduced in the algorithm can effectively avoid local optimality, while the dynamic adjustment of inertia weight balances the relationship between global search and local refinement. At the same time, the introduction of the fuzzy cognitive map method takes into account the correlation between parameters, making the optimization results more in line with the actual working conditions.

[0103] The temperature monitoring solution combining infrared temperature measurement array and Kalman filter ensures the accuracy and real-time performance of the measurement data. The fuzzy comprehensive evaluation rule realizes the objective evaluation and dynamic adjustment of the optimization results by establishing a scientific evaluation index system. This systematic technical solution not only conforms to the physical process of ultrasonic cutting, but also meets the needs of engineering practice, and has a strong theoretical basis and practical value.

[0104] A specific embodiment 1 of the present invention is provided below. The specific method of each step in this embodiment 1 is described in detail as follows:

[0105] Taking the frozen sections of kidney tissue as an example, the specific implementation process of the ultrasonic tool parameter optimization method of the present invention is explained. The working environment of a general frozen section machine is -30°C. During intraoperative frozen section examination, kidney tissue needs to be quickly sectioned to determine the tumor boundary. Traditional sectioning methods are prone to tissue degeneration due to improper temperature control, affecting the diagnostic results. This embodiment adopts the parameter optimization method of the present invention to achieve precise temperature control.

[0106] First, the parameters were collected. The vibration frequency of the ultrasonic tool was collected using a KT9035 piezoelectric frequency sensor. The sampling frequency was set to 1000 Hz and the sampling time was 180 seconds. The vibration end displacement was measured using an LK8036 laser displacement sensor. The amplitude measurement range was 0 to 100 microns and the measurement accuracy was 0.1 microns. The feed speed was monitored using an E60H photoelectric encoder with a resolution of 0.1 mm per second. The load power was measured using a WT3000 power meter with a measurement range of 0 to 1000 watts and a measurement accuracy of 1 watt. The initial collected data is shown in Table 1.

[0107] Table 1 Initial parameter collection data:

[0108]

[0109] The radial basis function neural network is used to establish the parameter mapping relationship. The network adopts a three-layer structure. The input layer contains 4 nodes, the number of hidden layer nodes is determined to be 8 through cross-validation, and the output layer is 1 node. The training data contains 200 sets of historical slice records, the learning rate is set to 0.01, and after 738 iterations, the mean square error drops to 0.00098, meeting the termination condition.

[0110] Then the parameter contribution values ​​are calculated. First, the data is normalized to obtain the normalized parameter values ​​as shown in Table 2.

[0111] Table 2 Parameter normalization results:

[0112] Time point (seconds) Frequency normalized value Amplitude normalization value Speed ​​normalized value Power normalized value 0 0.85 0.72 0.65 0.58 60 0.82 0.70 0.62 0.62 120 0.78 0.74 0.68 0.64 180 0.80 0.71 0.65 0.60

[0113] The weight coefficients were determined by principal component analysis, and α1 = 0.35, α2 = 0.28, α3 = 0.22, α4 = 0.15, and α5 = 0.12 were obtained. The calculated contribution value of the parameter group was 0.73, which was greater than the preset threshold of 0.6 and was determined to be a key parameter group.

[0114] The influence value model is established for the key parameter group, and the temperature partial derivative is calculated by five-point numerical difference, with a step size of 0.1% of the parameter range. The calculated influence value results are shown in Table 3.

[0115] Table 3 Parameter influence values:

[0116] Parameter Type Partial derivative values Volatility Correlation coefficient Influence value Vibration frequency 0.45 0.12 0.85 0.52 amplitude 0.38 0.15 0.78 0.45 Feed speed 0.25 0.18 0.65 0.35 Load power 0.22 0.20 0.58 0.30

[0117] The parameter importance matrix was constructed based on the influence value, and the eigenvector was calculated using the improved power iteration method. The final parameter adjustment order was obtained as follows: vibration frequency, amplitude, feed speed, and load power.

[0118] The improved particle swarm algorithm is used for parameter optimization, the population size is set to 50, and the maximum number of iterations is 200. During the optimization process, the temperature monitoring uses the FLI R A615 infrared temperature measurement array, the sensor spacing is 5 mm, and the cutting area of ​​50 mm × 50 mm is covered. After parameter optimization, the final parameter configuration is shown in Table 4.

[0119] Table 4 Optimized parameter configuration:

[0120] Parameter Type Optimization value Tolerance range Vibration frequency 32.5kHz ±0.5kHz amplitude 42.3μm ±1.0μm Feed speed 2.2mm / s ±0.2mm / s Load power 310W ±10W

[0121] The optimized parameter configuration was used for the slice experiment, and 50 slices were performed continuously. The temperature change trend was analyzed by the sliding window method, with a window length of 10 seconds and a step length of 1 second. The temperature monitoring results showed that the temperature of the slice area was always maintained within the range of -22±0.5 degrees Celsius, which met the temperature control requirements for frozen sections of kidney tissue.

[0122] Finally, the experimental results were subjected to fuzzy comprehensive evaluation, and the evaluation indicators included the maximum over-limit value, over-limit time ratio and temperature fluctuation intensity. The analytic hierarchy process was used to determine the indicator weights to be 0.4, 0.35 and 0.25 respectively. The evaluation results showed that the optimized parameter configuration had a good temperature control effect.

[0123] Table 5 is the comparison results between the solution of the present invention and the traditional method:

[0124] Table 5 Comparison of new and old methods:

[0125] Compare Projects Traditional methods Method of the present invention Optimization basis Experience Adjustment Quantitative evaluation Parameter selection Single parameter optimization Multi-parameter coupling Temperature control Passive control Active prediction

[0126] Traditional methods mainly rely on operator experience to adjust parameters, or use a single parameter optimization method, which lacks systematicity and scientificity. The present invention, by establishing a complete parameter optimization system, realizes quantitative analysis and intelligent optimization of the coupling relationship between parameters, significantly improving temperature control accuracy and optimization efficiency. At the same time, the method of the present invention has strong adaptability and repeatability, can quickly respond to the sectioning requirements of different tissue types, and provide reliable technical support for clinical pathological diagnosis.

[0127] It should be noted that the variables involved in the present invention are explained in detail as shown in Table 6:

[0128] Table 6 Variable explanation:

[0129]

[0130] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for optimizing ultrasonic tool parameters for frozen tissue sections, characterized in that: The following steps are involved: The vibration frequency, amplitude, feed speed, load power and cutting tissue type of the ultrasonic cutter in the process of cutting frozen tissue are collected as an initial parameter group, and a mapping relationship between the initial parameter group and the temperature change of the frozen tissue is established; Calculate the contribution value of the initial parameter group and compare the contribution value with a preset threshold value, determine the parameter group greater than the preset threshold value as the key parameter group, and determine the parameter group less than or equal to the preset threshold value as the non-key parameter group; establish an influence value model for the key parameter group and determine the parameter adjustment sequence, and optimize the parameters according to the parameter adjustment sequence until the temperature change of the frozen tissue meets the preset temperature range requirements.

2. The method for optimizing ultrasonic tool parameters for frozen tissue sections according to claim 1, characterized in that: The step of collecting the initial parameter group is specifically to use a frequency sensor to collect the vibration frequency, the sampling frequency is 1000Hz, and the collection range of the vibration frequency is 20kHz to 40kHz; use a laser displacement sensor to collect the amplitude, the measurement range is 0 microns to 100 microns, and the measurement accuracy is 0.1 microns; use a photoelectric encoder to collect the feed speed, the resolution is 0.1 mm per second; use a Hall effect power meter to collect the load power, the measurement range is 0 watts to 1000 watts, and the measurement accuracy is 1 watt.

3. The method for optimizing ultrasonic tool parameters for frozen tissue sections according to claim 2, characterized in that: The step of establishing a mapping relationship between the initial parameter group and the temperature change of frozen tissue specifically adopts a radial basis function neural network method, uses the initial parameter group as the input layer, and the temperature change as the output layer. The number of hidden layer nodes is determined by a cross-validation method, and the activation function uses a Gaussian function. The learning rate is 0.01, and the training termination condition is that the mean square error is less than 0.001 or the maximum number of iterations is 1000.

4. The method for optimizing ultrasonic tool parameters for frozen tissue sections according to claim 3, characterized in that: The step of calculating the contribution value is specifically to normalize the maximum and minimum values ​​of the initial parameter group, map all parameter values ​​to between 0 and 1, and then use the principal component analysis method combined with the expert scoring method to determine the weight coefficient. The principal component contribution rate threshold is 85%. The contribution value calculation introduces a random error term that obeys the normal distribution, with a mean of 0 and a standard deviation of 0.

05.

5. The method for optimizing ultrasonic tool parameters for frozen tissue sections according to claim 4, characterized in that: The step of establishing the influence value model specifically comprises calculating the partial derivative of temperature change with respect to each parameter using a five-point numerical difference method, with a step length of 0.1% of the parameter range; calculating the coefficient of fluctuation of the parameter, with a sample size of not less than 30; The Pearson correlation coefficient was used to calculate the correlation between the parameters and temperature changes; The weight coefficient was determined by analytic hierarchy process combined with grey relational analysis, and the consistency ratio was less than 0.

1.

6. The method for optimizing ultrasonic tool parameters for frozen tissue sections according to claim 5, characterized in that: The step of determining the parameter adjustment order specifically constructs a parameter importance matrix, uses an improved power iteration method to calculate the eigenvector of the matrix, selects an all-1 vector as the initial vector, the iteration accuracy is 0.001, and the maximum number of iterations is 1000 times; uses a fuzzy cognitive map method to establish a parameter association network, and the association strength threshold is 0.3, which is used to adjust the final adjustment order.

7. The method for optimizing ultrasonic tool parameters for frozen tissue sections according to claim 6, characterized in that: The parameter optimization step specifically adopts the improved particle swarm optimization algorithm to adjust the parameters, the population size is 50, the maximum number of iterations is 200, the inertia weight decreases linearly from 0.9 to 0.4, and the learning factor is 2.0; an adaptive mutation strategy is adopted, the mutation probability is inversely proportional to the fitness value, and the basic mutation probability is 0.

1.

8. The method for optimizing ultrasonic tool parameters for frozen tissue sections according to claim 7, characterized in that: It also includes the step of establishing a parameter optimization database, specifically using a relational database management system to establish the database, designing a three-layer table structure of a basic parameter table, an optimization result table, and a related information table, and maintaining data consistency through primary and foreign key relationships; Create composite indexes for commonly used query fields; adopt an incremental backup strategy, perform incremental backup every day, and full backup once a week.

9. The method for optimizing ultrasonic tool parameters for frozen tissue sections according to claim 8, characterized in that: It also includes steps for verifying the suitability of parameters, specifically using an infrared temperature sensor array to monitor the temperature distribution of frozen tissue in real time, with a sensor spacing of 5 mm and a sampling frequency of 10 Hz; using a Kalman filter for noise filtering; and using a sliding window method to calculate the temperature change trend, with a window length of 10 seconds, a step size of 1 second, and a temperature tolerance of 0.5 degrees Celsius.

10. The method for optimizing ultrasonic tool parameters for frozen tissue sections according to claim 1, characterized in that: It also includes the steps of optimizing result judgment, specifically, using fuzzy comprehensive evaluation method to establish an evaluation index system, including the maximum over-limit value, over-limit time ratio and temperature fluctuation intensity, and using fuzzy hierarchical analysis method to determine the index weight; using annealing algorithm to adaptively adjust the weight coefficient, with the initial temperature of 100, the cooling coefficient of 0.95 and the termination temperature of 0.01.

Images