Intelligent management system and method applied to radio frequency energy output device

Abstract

The invention discloses an intelligent management system and method applied to a radio frequency energy output device, and belongs to the technical field of intelligent management of the radio frequency energy output device.An integrated sensor array is deployed on a radio frequency host and a hand tool electrode, and load voltage, output current, electrode temperature and tissue impedance are synchronously collected; a dynamic thermal impedance collaborative analysis model is constructed after time synchronization alignment, a three-dimensional thermal field simulation map is generated, and a thermal accumulation trend is predicted; a double-layer control framework is constructed, a first control layer generates a dynamic power adjustment rule base based on a preset treatment target and a safety threshold value and outputs an instruction, a second control layer receives the instruction and adjusts output parameters in real time, and a strategy is adjusted by combining a model prediction result in the execution process; meanwhile, the operation state is continuously monitored, a fault classification model is constructed for anomaly detection, a fault source is positioned, and a visual report is generated; and finally, constructing a mapping relation based on historical data, and generating an energy control scheme adapted to individualized requirements through iterative optimization.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of radio frequency energy output device management, and particularly relates to an intelligent management system and method applied to a radio frequency energy output device. BACKGROUND

[0002] The radio frequency energy output device is a device for applying radio frequency energy to human skin and subcutaneous tissue to achieve a specific treatment goal through thermal effect. This kind of device usually adopts a structure of a host computer + hand electrode, and an operator needs to simultaneously pay attention to the host computer parameter setting and the electrode action position, and realizes energy output regulation and control through a traditional power control mode to achieve the treatment purpose of improving skin state.

[0003] However, with the development of the fine application requirements in clinical application, this technical path gradually shows technical limitations, and the control system of the prior art has problems of single detection dimension and insufficient fault diagnosis capability. During the operation of the device, due to the lack of real-time cooperative monitoring of the load voltage, current and temperature, it is difficult to accurately perceive the changes of the human tissue impedance and the local heat accumulation, which leads to two problems of the power output: on the one hand, in order to reduce the risk, manufacturers often set the factory preset power to be relatively conservative, which affects the treatment effect; on the other hand, if the operator tries to increase the power to improve the effect, the risk of scalding may be increased due to the untimely temperature control; in addition, due to the lack of multi-dimensional detection data support, when the device appears abnormal conditions such as open circuit caused by electrode falling off and short circuit caused by contact with metal, it is difficult to quickly locate the fault reason, and manual one-by-one checking is needed during maintenance, which not only increases the difficulty of device maintenance, but also may affect the safety and reliability of the device use. SUMMARY

[0004] The purpose of the present application is to provide an intelligent management system and method applied to a radio frequency energy output device to solve the problems raised in the background.

[0005] In order to solve the above technical problems, the present application provides the following technical scheme: an intelligent management method applied to a radio frequency energy output device, comprising:

[0006] Step S100: deploying an integrated sensor array on a radio frequency host computer and a hand electrode, synchronously collecting four types of core parameters of load voltage, output current, electrode temperature and tissue impedance; after aligning the multi-source parameters through a time synchronization algorithm, a dynamic thermal impedance cooperative analysis model is constructed; based on the model output, a three-dimensional thermal field simulation atlas of the radio frequency energy action area is generated, and the thermal accumulation trend of the radio frequency energy action area is dynamically predicted in combination with the real-time monitored tissue impedance change characteristics;

[0007] Step S200: a double-layer control architecture is constructed, including a first control layer and a second control layer; the first control layer takes preset treatment target parameters and safety threshold conditions as input quantities, generates a dynamic power adjustment rule library and outputs a control instruction; the second control layer dynamically adjusts radio frequency energy output parameters according to the control instruction; in the control execution phase, based on the prediction result of the dynamic thermal impedance cooperative analysis model, a power adjustment strategy is dynamically implemented;

[0008] Step S300: in the power adjustment process, the running state parameters of the radio frequency energy output device are continuously collected to form a multi-working condition data set, and a fault classification model is constructed; through real-time feature extraction and abnormal working condition detection of the fault classification model, the fault root cause is located in combination with multi-parameter correlation analysis, and finally a visual diagnosis report is generated;

[0009] Step S400: based on the changes of tissue impedance, temperature distribution and operation feedback information in the historical treatment data, a mapping relationship between the device running state and the power control strategy is constructed; through an iterative optimization mechanism, the treatment effect and safety indicators are evaluated, and an energy control scheme suitable for individual needs is generated.

[0010] Further, step S100 includes:

[0011] Step S101: a load voltage sampling circuit and a current transformer are integrated inside a radio frequency host, the load voltage sampling circuit is composed of a resistance and a transformer to form a voltage division and isolation network; at the same time, a thermistor temperature sensor is built-in at the contact end of a hand tool electrode to form a sensor array; when the device is started, four types of parameters of load voltage, output current, electrode temperature and tissue impedance are synchronously collected at a preset sampling frequency; the load voltage represents the voltage at the output end of the radio frequency, which is sampled through a resistance circuit after being coupled by a capacitor, and then is transmitted by a transformer; the output current represents the current output by the radio frequency, which is converted into a voltage signal by a transformer, and then is transmitted by a transformer after resistance conversion; the electrode temperature represents the temperature at the contact end of the hand tool electrode, which is detected by a temperature sensor and converted into a voltage value by a circuit; the tissue impedance represents the impedance presented by the human body tissue, which is calculated based on the phase difference between the load voltage and the output current; the collected data is subjected to low-pass filtering circuit to remove high-frequency noise, and then is aligned in sampling time by a time stamp synchronization algorithm;

[0012] Step S102: a three-dimensional geometric model of human skin including epidermis, dermis and subcutaneous fat layer and subcutaneous tissue is established, the grid is divided, and the thermal physical parameters and electrical characteristic parameters of each layer of tissue are set; the synchronous load voltage, output current, electrode temperature and tissue impedance parameters are taken as input, and based on the electromagnetic heat coupling principle, the heat conduction equation is solved: Wherein, α is the thermal diffusion coefficient, σ is the electrical conductivity, E is the electric field intensity, ρ is the tissue density, cp Cp is the specific heat capacity, and T represents the temperature; during the model construction process, when the real-time monitored tissue impedance change rate exceeds the threshold value, it is determined that the tissue heat accumulation is accelerated, and then the heat source distribution is dynamically adjusted through the following formula: wherein Q n represents the adjusted heat source intensity, Q represents the heat source intensity before adjustment, ΔZ t represents the impedance fluctuation amplitude, and Δt represents the time interval; k is a proportional coefficient, and k ∈ [0.5, 1]; after the model is constructed, based on the temperature distribution data calculated by the model, a three-dimensional visualization algorithm is used to map the temperature values of each grid node to RGB color values, and the temperature values and temperature gradients at different positions are presented, thereby generating a heat field simulation map of the radiofrequency energy action area;

[0013] Step S103: Real-time acquisition and storage of temperature data in the recent period of time, while synchronously recording the tissue impedance change data in the corresponding time period; using the exponential smoothing method to fit the historical temperature data, and adaptively adjusting the smoothing coefficient according to the tissue impedance fluctuation: when the tissue impedance fluctuation amplitude is less than the preset value, the smoothing coefficient is set to p1; when the tissue impedance fluctuation amplitude is greater than or equal to the preset value, the smoothing coefficient is set to p2; performing first-order difference processing on the fitted temperature sequence to obtain a stationary sequence; calculating the autocorrelation function and the partial autocorrelation function of the differentiated sequence to determine the order of the ARIMA model; estimating the parameters of the ARIMA model by the maximum likelihood estimation method; based on the estimated model parameters, predicting the temperature values in the future period of time; defining the temperature rise rate threshold value as d t , when the predicted temperature rise rate in a certain period of time exceeds the threshold value, triggering the early warning mechanism and prompting the existing heat accumulation risk.

[0014] Further, step S200 includes:

[0015] Step S201: The first control layer takes the preset treatment target parameters and safety threshold conditions as inputs, wherein the treatment target parameters include the preset treatment temperature, the target energy accumulation value and the treatment time length, and the safety threshold conditions include the upper limit of the safety temperature, the tissue impedance abnormal fluctuation threshold value and the power attenuation gradient; historical treatment data is collected, including the input conditions and the corresponding power adjustment actions in various treatment scenarios, including the power adjustment amplitude, direction and duration;

[0016] For each feature in the input condition, first calculate the information gain of each feature, which is determined by calculating the entropy difference of the data set before and after the division, and the calculation formula of the entropy is:

[0017]

[0018] where D is the dataset, H(D) represents the entropy of the dataset D, C i is the i-th type of power adjustment action, |C i and |D| represent the number of samples of the i-th type of action and the total number of samples of the dataset, respectively; then, the entropy of each feature is calculated; finally, the information gain rate of each feature is obtained by dividing the information gain of each feature by its entropy, and the top x features with the highest information gain rate are selected as the key nodes of the decision tree; the dataset is divided into several subsets according to different values of the features; at each node, the feature with the maximum information gain is selected as the division condition; the above process is repeated recursively for each subset until the preset stopping condition is met; the mapping relationship between the conditions and actions is extracted from the constructed decision tree; each path from the root node to the leaf node represents a rule, and the combination of conditions on the path corresponds to a specific power adjustment action; in the actual treatment process, the appropriate rule is selected from the rule base according to the real-time monitored input conditions, and the corresponding control instruction is output; finally, the first control layer converts the generated rule into a control instruction package containing the power adjustment amplitude, direction and duration;

[0019] Step S202: receiving the control instruction package generated by the first control layer, parsing the power adjustment amplitude, direction and duration parameters; mapping the power adjustment instruction to the underlying control parameters of the radio frequency generator, including voltage / current amplitude, pulse width modulation duty cycle, working mode switching; continuously collecting the actual voltage, current and hand tool electrode temperature at the output end of the radio frequency host; combining the thermal field simulation atlas output by the dynamic thermal impedance cooperative analysis model to obtain the three-dimensional temperature distribution and temperature rise rate of the action area; using the PID control algorithm, taking the deviation between the preset treatment temperature and the actual temperature as the input, and dynamically adjusting the output parameters: where u(t) represents the control amount output by the PID controller; K p , K i , K d are the proportional coefficient, integral coefficient and differential coefficient, respectively, and e(t) represents the deviation between the preset treatment temperature and the actual temperature; when the temperature rise rate d t exceeds the threshold, the power attenuation is triggered in advance, according to the setting: ΔP=λ×d t , where ΔP represents the power attenuation amount, and λ is the attenuation coefficient.

[0020] Further, step S300 comprises:

[0021] Step S301: continuously collect the operating state parameters of the radio frequency energy output device under different working conditions of normal treatment, power attenuation and temperature anomaly, including load voltage, output current, electrode temperature, tissue impedance and radio frequency power output value, to form a multi-dimensional original data set; the collected data is mapped to the [0, 1] interval through the Min-Max normalization method; key features are extracted from the normalized data: for the load voltage, the voltage fluctuation coefficient, i.e. the ratio of the standard deviation to the mean value of the voltage within a certain time, is extracted; for the output current, the current distortion rate is calculated, which measures the distortion of the current by analyzing the proportion of harmonic components of the current; for the electrode temperature, the temperature rise rate is obtained, which is represented by the change amount of the temperature per unit time; for the tissue impedance, the impedance change gradient, i.e. the change amount of the impedance per unit time, is calculated; at the same time, the power adjustment frequency, i.e. the number of power adjustments per unit time, is extracted, to construct a feature vector set;

[0022] Step S302: calculate the correlation between each feature and the fault type through the mutual information method, and calculate the mutual information value of each feature and the fault type:

[0023] MI(X i ,Y j )=∑ xi∈Xi ∑ yj∈Yj p(x i ,y j )log[p(x i ,y j ) / p(x i )p(y j )];

[0024] Where MI(X i ,Y j ) represents the mutual information value between the i-th feature and the j-th fault, X i represents the i-th feature variable extracted from the operating state parameters, Y j represents the j-th fault type, x i represents the specific value of feature X i , y j represents the specific category value of fault type Y j , p(x i ,y j ) is the joint probability distribution of X i and Y j , p(x i ) and p(y j) is an edge probability distribution; according to the mutual information value from high to low, the top b ranked features are screened as the input of the model to form the final feature subset; wherein b is a set value; the feature subset F and the corresponding fault label Y are divided into training set and test set according to the proportion of 7:3; then the random forest algorithm is used for training, the number of decision trees and the maximum depth are set, and the fault type is taken as the output label; finally, the model parameters are optimized through 10-fold cross-validation, and the training is stopped when the accuracy and recall rate of the model reach the preset threshold;

[0025] Step S303: In the process of equipment operation, set a timing cycle to collect working condition parameters in real time, extract feature vectors and input the fault classification model; when the model outputs an abnormal classification result, start the multi-parameter correlation analysis process: based on the top 8 features ranked by mutual information value in step S302, establish a feature fault mapping table, combine the equipment hardware architecture, and map the feature anomaly to a specific component; trace the historical change trajectory of the abnormal feature, combine the feature fault mapping table, and locate to a specific fault component; after fault location is completed, the system automatically generates a visual diagnostic report: display the parameter change curve within 2 minutes before and after the fault occurs in the form of a time axis, and mark the abnormal data points in red; present the fault tracing process through a flowchart, and mark the judgment basis of each layer node.

[0026] Further, step S400 includes:

[0027] Step S401: Extract f cases of treatment data from the historical database, including tissue impedance features, temperature distribution features, and operation feedback information; perform binning processing on the temperature and impedance data, set different thresholds for the tissue impedance change rate and the temperature rise rate, and divide the temperature and impedance into three intervals of low, medium and high; use the Apriori algorithm to analyze the association rules between the device operating state and the power control strategy: the input items are tissue impedance interval, temperature distribution feature, and treatment stage; the output items are power adjustment amplitude, adjustment frequency, and duration; calculate the efficacy weight of each association rule: count the number of cases with significant efficacy marked in the operation feedback when the rule is applied in the historical treatment, divide by the total number of applications of the rule, and obtain the case proportion with significant efficacy as the quantitative value of the efficacy weight; calculate the safety weight: count the number of adverse events in the application process of the rule, subtract the adverse event rate from 1 to obtain the safety weight value, the adverse event rate = number of adverse events / total number of applications; finally, the final weight of each rule is obtained by weighted summation;

[0028] Step S402: define the quantified treatment effect and safety evaluation indicators, the treatment effect indicators include the target tissue temperature compliance rate, i.e. the proportion of time during which the target tissue temperature is maintained in the preset effective treatment temperature interval; the safety indicators include the high temperature area proportion and the number of times of exceeding the temperature rise rate; based on these indicators, the actual parameters of the historical treatment scheme are compared with the preset threshold value: if the target tissue temperature compliance rate is greater than or equal to the threshold value, it is determined that the treatment effect is effective; if the high temperature area proportion is less than or equal to the threshold value and the number of times of exceeding the temperature rise rate is less than or equal to 2, it is determined that the safety is up to standard; finally, the treatment effect and the safety result are combined to divide the scheme level; the device operating state is defined as a state space, the power adjustment strategy is used as an action space, a reward function is constructed according to the treatment effect and the safety indicators, and a corresponding reward value is given to each action, a positive reward is given if the standard is met, and a negative punishment is given if the standard is not met; during the optimization process, the scheme library is updated for a preset number of iterations, and the high-quality schemes with the top a reward values are retained, where a is a set value;

[0029] Step S403: generate an energy control scheme combined with individualized patient parameters, obtain the tissue characteristic parameters of the patient and the treatment target set by the doctor before treatment; based on these individualized inputs, the most suitable energy control strategy is matched from the optimized scheme library, and an individualized scheme including a full-process power adjustment curve, safety warning threshold adjustment and real-time feedback strategy is output.

[0030] An intelligent management system applied to a radio frequency energy output device, characterized in that: the system comprises a parameter acquisition and thermal field modeling module, a double-layer power control module, a fault detection module and an individualized scheme optimization module;

[0031] The parameter acquisition and thermal field modeling module deploys an integrated sensor array on the radio frequency host and the electrode of the handpiece, synchronously acquires four types of core parameters including load voltage, output current, electrode temperature and tissue impedance; after aligning the multi-source parameters through a time synchronization algorithm, a dynamic thermal impedance collaborative analysis model is constructed; based on the model output, a three-dimensional thermal field simulation atlas of the radio frequency energy action area is generated, and the thermal accumulation trend of the radio frequency energy action area is dynamically predicted in combination with the real-time monitored tissue impedance change characteristics;

[0032] The double-layer power control module constructs a double-layer control architecture, including a first control layer and a second control layer; the first control layer takes preset treatment target parameters and safety threshold conditions as input quantities, generates a dynamic power adjustment rule library and outputs control instructions; the second control layer dynamically adjusts the radio frequency energy output parameters according to the control instructions; in the control execution stage, the power adjustment strategy is dynamically implemented based on the prediction results of the dynamic thermal impedance collaborative analysis model;

[0033] The fault detection module continuously collects the operating state parameters of the radio frequency energy output device during power regulation, forms a multi-working condition data set, constructs a fault classification model, detects abnormal working conditions through real-time feature extraction and the fault classification model, locates the fault source in combination with multi-parameter correlation analysis, and finally generates a visual diagnosis report.

[0034] The personalized scheme optimization module constructs the mapping relationship between the device operating state and the power regulation strategy based on the tissue impedance change, temperature distribution and operation feedback information in the historical treatment data, evaluates the treatment effect and safety indicators through an iterative optimization mechanism, and generates an energy control scheme adapted to individual needs.

[0035] Compared with the prior art, the present application has the following beneficial effects:

[0036] The present application realizes intelligent management of the radio frequency energy output device by integrating a sensor array, a dynamic thermal impedance collaborative analysis model and a double-layer control architecture, dynamically adjusts the radio frequency energy output by real-time monitoring of core parameters such as temperature and tissue impedance, can more accurately predict and control heat distribution during treatment, optimizes treatment effect, avoids safety risks such as overheating, and improves the safety and effect of the treatment process.

[0037] The present application realizes dynamic adjustment of radio frequency energy output parameters, real-time detection of abnormal working conditions and accurate positioning of fault sources by constructing a double-layer control architecture, a fault classification model and a mapping relationship between the device operating state and the power regulation strategy. At the same time, based on the iterative optimization mechanism of historical treatment data, an energy control scheme adapted to individual needs can be generated to ensure the accuracy and individualization of the treatment process, thereby significantly improving the treatment effect and patient satisfaction. BRIEF DESCRIPTION OF DRAWINGS

[0038] The accompanying drawings are used to provide a further understanding of the present application, and constitute a part of the specification, together with the embodiments of the present application, to explain the present application, and do not constitute a limitation on the present application. In the drawings:

[0039] Figure 1 A method flowchart of an intelligent management method applied to a radio frequency energy output device. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0041] Please refer toFigure 1 The present application provides a technical solution: an intelligent management method applied to a radio frequency energy output device, comprising:

[0042] Step S100: deploying an integrated sensor array on the radio frequency host and the handpiece electrode, synchronously collecting four types of core parameters of load voltage, output current, electrode temperature and tissue impedance; after aligning the multi-source parameters through a time synchronization algorithm, a dynamic thermal impedance collaborative analysis model is constructed; based on the model output, a three-dimensional thermal field simulation atlas of the radio frequency energy action area is generated, and the thermal accumulation trend of the radio frequency energy action area is dynamically predicted in combination with the real-time monitored tissue impedance change characteristics;

[0043] Step S200: constructing a double-layer control architecture, including a first control layer and a second control layer; the first control layer takes preset treatment target parameters and safety threshold conditions as input quantities, generates a dynamic power adjustment rule library and outputs control instructions; the second control layer dynamically adjusts the radio frequency energy output parameters according to the control instructions; in the control execution stage, based on the prediction results of the dynamic thermal impedance collaborative analysis model, the power adjustment strategy is dynamically implemented;

[0044] Step S300: in the power adjustment process, continuously collecting the running state parameters of the radio frequency energy output device, forming a multi-working condition data set, constructing a fault classification model; through real-time feature extraction and abnormal working condition detection of the fault classification model, the fault root cause is located in combination with multi-parameter correlation analysis, and finally a visual diagnosis report is generated;

[0045] Step S400: based on the tissue impedance change, temperature distribution and operation feedback information in the historical treatment data, a mapping relationship between the device running state and the power control strategy is constructed; through an iterative optimization mechanism, the treatment effect and safety indicators are evaluated, and an energy control scheme suitable for individual needs is generated.

[0046] Step S100 includes:

[0047] Step S101: integrate a load voltage sampling circuit and a current transformer inside the radio frequency host, the load voltage sampling circuit is composed of a voltage divider and an isolation network by a resistor and a transformer; meanwhile, a thermistor temperature sensor is built in the hand tool electrode contact end to form a sensor array; after the device is started, four types of parameters of load voltage, output current, electrode temperature and tissue impedance are synchronously collected at a preset sampling frequency; the load voltage represents the voltage at the radio frequency output end, which is sampled through a resistor circuit after capacitive coupling, and then transmitted by a transformer; the output current represents the current output by the radio frequency, which is converted into a voltage signal by a mutual inductor, and then transmitted by a transformer; the electrode temperature represents the temperature of the hand tool electrode contact end, which is detected by a temperature sensor and converted into a voltage value by a circuit; the tissue impedance represents the impedance presented by the human body tissue, which is calculated based on the phase difference between the load voltage and the output current; the collected data is filtered by a low-pass filter circuit to remove high-frequency noise, and then the sampling time of each parameter is aligned by a timestamp synchronization algorithm;

[0048] Step S102: establish a three-dimensional geometric model of human skin and subcutaneous tissue including epidermis, dermis and subcutaneous fat layer, divide the grid and set the thermal physical parameters and electrical characteristic parameters of each layer of tissue; take the synchronized load voltage, output current, electrode temperature and tissue impedance parameters as input, based on the electromagnetic heat coupling principle, solve the heat conduction equation: Where, α is the thermal diffusion coefficient, σ is the electrical conductivity, E is the electric field strength, ρ is the tissue density, c p is the specific heat capacity, and T represents the temperature; during the model construction process, when the real-time monitored tissue impedance change rate exceeds the threshold value, it is determined that the tissue heat accumulation is accelerated, then the heat source distribution is dynamically adjusted by the following formula: Where Q n represents the adjusted heat source intensity, Q represents the heat source intensity before adjustment, ΔZ t represents the impedance fluctuation amplitude, and Δt represents the time interval; k is a proportional coefficient, and k ∈ [0.5, 1]; after the model construction is completed, based on the temperature distribution data calculated by the model, the temperature values of each grid node are mapped to RGB color values by using a three-dimensional visualization algorithm, and the temperature values and temperature gradients at different positions are presented, thereby generating a heat field simulation map of the radio frequency energy action area;

[0049] Step S103: Real-time acquisition and storage of temperature data in the recent period of time, while synchronously recording the tissue impedance change data in the corresponding time period; fitting the historical temperature data using the exponential smoothing method, and adaptively adjusting the smoothing coefficient according to the tissue impedance fluctuation: when the tissue impedance fluctuation amplitude is less than the preset value, the smoothing coefficient is set to p1; when the tissue impedance fluctuation amplitude is greater than or equal to the preset value, the smoothing coefficient is set to p2; first-order difference processing is performed on the fitted temperature sequence to obtain a smooth sequence; the autocorrelation function and the partial autocorrelation function of the differential sequence are calculated to determine the order of the ARIMA model; the parameters of the ARIMA model are estimated by the maximum likelihood estimation method; based on the estimated model parameters, the temperature value in the future period of time is predicted; the temperature rise rate threshold d is defined t When the predicted temperature rise rate in a certain period of time exceeds the threshold, the warning mechanism is triggered, and the existing heat accumulation risk is prompted.

[0050] Step S200 includes:

[0051] Step S201: The first control layer takes the preset treatment target parameters and safety threshold conditions as inputs, the treatment target parameters including the preset treatment temperature, the target energy accumulation value and the treatment time length, and the safety threshold conditions including the upper limit of the safety temperature, the tissue impedance abnormal fluctuation threshold and the power attenuation gradient; historical treatment data are collected, including the input conditions and the corresponding power adjustment actions in various treatment scenarios, including the power adjustment amplitude, direction and duration;

[0052] For each feature in the input condition, first calculate the information gain of each feature, which is determined by calculating the entropy difference of the data set before and after division, and the calculation formula of the entropy is:

[0053]

[0054] Where D is the data set, H(D) represents the entropy of the data set D, C i is the i-th power adjustment action, and |C i| and |D| represent the number of samples of the i-th action and the total number of samples of the data set, respectively; then, the entropy of each feature is calculated; finally, the information gain rate of each feature is obtained by dividing the information gain of each feature by its entropy, and the top x features with the highest information gain rate are selected as the key nodes of the decision tree; the data set is divided into several subsets according to different values of the features; at each node, the feature with the maximum information gain is selected as the division condition; the above process is repeated recursively for each subset until the preset stopping condition is met; the mapping relationship between the conditions and the actions is extracted from the constructed decision tree; each path from the root node to the leaf node represents a rule, and the combination of the conditions on the path corresponds to a specific power adjustment action; in the actual treatment process, the appropriate rule is selected from the rule base according to the real-time monitored input conditions, and the corresponding control instruction is output; finally, the first control layer converts the generated rule into a control instruction package containing the power adjustment amplitude, direction and duration;

[0055] Step S202: receiving the control instruction package generated by the first control layer, analyzing the power adjustment amplitude, direction and duration parameters; mapping the power adjustment instruction to the underlying control parameters of the radio frequency generator, including voltage / current amplitude, pulse width modulation duty cycle, working mode switching; continuously collecting the actual voltage, current and hand tool electrode temperature at the output end of the radio frequency host; combining the thermal field simulation atlas output by the dynamic thermal impedance cooperative analysis model to obtain the three-dimensional temperature distribution and temperature rise rate of the action area; using the PID control algorithm, taking the deviation between the preset treatment temperature and the actual temperature as the input, dynamically adjusting the output parameters: wherein u(t) represents the control amount output by the PID controller; K p , K i , K d are the proportional coefficient, integral coefficient and differential coefficient respectively, and e(t) represents the deviation between the preset treatment temperature and the actual temperature; when the temperature rise rate d t exceeds the threshold value, the power attenuation is triggered in advance, according to the setting: ΔP = λ × d t , wherein ΔP represents the power attenuation amount, and λ is the attenuation coefficient.

[0056] Step S300 includes:

[0057] Step S301: Continuously collect operating state parameters of the radio frequency energy output device under different working conditions, including normal treatment, power attenuation, and temperature abnormality, including load voltage, output current, electrode temperature, tissue impedance, and radio frequency power output value, to form a multi-dimensional original data set; the collected data are mapped to the interval [0,1] using the Min-Max normalization method; key features are extracted from the normalized data: for load voltage, the voltage fluctuation coefficient is extracted, that is, the ratio of the standard deviation of the voltage to the mean within a certain period of time; for output current, the current distortion rate is calculated, and the distortion of the current is measured by analyzing the proportion of harmonic components of the current; for electrode temperature, the temperature rise rate is obtained, expressed as the change in temperature per unit time; for tissue impedance, the impedance change gradient is calculated, that is, the change in impedance per unit time; at the same time, the power adjustment frequency, that is, the number of power adjustments per unit time, is extracted to construct a feature vector set;

[0058] Step S302: Calculate the correlation between each feature and the fault type using the mutual information method, and calculate the mutual information value between each feature and the fault type:

[0059] MI(X i ,Y j )=∑ xi∈Xi ∑ yj∈Yj p(x i ,y j )log[p(x i ,y j ) / p(x i )p(y j )];

[0060] Among them, MI(X i ,Y j ) represents the mutual information value between the i-th feature and the j-th type of fault, X i represents the i-th characteristic variable extracted from the operating state parameters, Y j represents the jth fault type, x i Represents feature X i The specific value of y j Indicates fault type Y j The specific category value of p(x i ,y j ) is X i and Y j The joint probability distribution of p(x i ) and p(y j) is an edge probability distribution; according to the mutual information value from high to low, the top b ranked features are screened as the input of the model to form the final feature subset; wherein b is a set value; the feature subset F and the corresponding fault label Y are divided into training set and test set according to the proportion of 7:3; then the random forest algorithm is used for training, the number of decision trees and the maximum depth are set, and the fault type is taken as the output label; finally, the model parameters are optimized through 10-fold cross-validation, and the training is stopped when the accuracy and recall rate of the model reach the preset threshold;

[0061] Step S303: In the process of equipment operation, set a timing cycle to collect working condition parameters in real time, extract feature vectors and input the fault classification model; when the model outputs an abnormal classification result, start the multi-parameter correlation analysis process: based on the top 8 features ranked by mutual information value in step S302, establish a feature fault mapping table, combine the equipment hardware architecture, and map the feature anomaly to a specific component; trace the historical change trajectory of the abnormal feature, combine the feature fault mapping table, and locate to a specific fault component; after fault location is completed, the system automatically generates a visual diagnostic report: display the parameter change curve within 2 minutes before and after the fault occurs in the form of a time axis, and mark the abnormal data points in red; present the fault tracing process through a flowchart, and mark the judgment basis of each layer node.

[0062] Step S400 includes:

[0063] Step S401: Extract f cases of treatment data from the historical database, including tissue impedance features, temperature distribution features, and operation feedback information; perform binning processing on temperature and impedance data, set different thresholds for tissue impedance change rate and temperature rise rate, and divide temperature and impedance into three intervals of low, medium and high; use the Apriori algorithm to analyze the association rules between device operating state and power control strategy: the input items are tissue impedance interval, temperature distribution feature, and treatment stage; the output items are power adjustment amplitude, adjustment frequency, and duration; calculate the efficacy weight of each association rule: count the number of cases with significant efficacy when the rule is applied in historical treatment, divide by the total number of applications of the rule, and obtain the proportion of cases with significant efficacy as the quantitative value of the efficacy weight; calculate the safety weight: count the number of adverse events during the application of the rule, subtract the adverse event rate from 1 to obtain the safety weight value, the adverse event rate = number of adverse events / total number of applications; finally, the maximum weight of each rule is obtained by weighted summation;

[0064] Step S402: define the quantified treatment effect and safety evaluation indicators, the treatment effect indicators include the target tissue temperature compliance rate, that is, the proportion of time during which the target tissue temperature is maintained in the preset effective treatment temperature interval during the treatment process; the safety indicators include the high temperature area area proportion and the temperature rise rate exceeding the standard times; based on these indicators, the actual parameters of the historical treatment scheme are compared with the preset threshold value: if the target tissue temperature compliance rate is greater than or equal to the threshold value, it is determined that the treatment effect is effective; if the high temperature area area proportion is less than or equal to the threshold value and the temperature rise rate exceeding the standard times is less than or equal to 2 times, it is determined that the safety is up to standard; finally, the treatment effect and the safety result are combined to divide the scheme level; the device running state is defined as a state space, the power adjustment strategy is regarded as an action space, the reward function is constructed according to the treatment effect and the safety indicators, and a corresponding reward value is given to each action; if the standard is reached, a positive reward is given, and if the standard is not reached, a negative punishment is given; during the optimization process, the scheme library is updated for a preset number of iterations, and the high-quality schemes with the top a reward values are retained, wherein a is a set value;

[0065] Step S403: generate an energy control scheme combined with individualized parameters of a patient, before treatment, obtain the tissue characteristic parameters of the patient and the treatment target set by the doctor; based on these individualized inputs, the most suitable energy control strategy is matched from the optimized scheme library, and an individualized scheme containing a whole-process power adjustment curve, safety warning threshold adjustment and real-time feedback strategy is output.

[0066] Embodiments of the present application:

[0067] The load voltage sampling circuit and the current transformer are integrated inside the radio frequency host, the thermistor is arranged in the contact end of the hand tool electrode, the load voltage is synchronously collected with a sampling period of 100 ms, the output current is 0-5 A, the electrode temperature is 25-60 DEG C, and the tissue impedance is calculated through the voltage-current phase difference and ranges from 50 to 500 omega; after the collected data is filtered by a low-pass filter to remove high-frequency noise, the data is aligned through a time stamp synchronization algorithm and then input into a dynamic thermal impedance collaborative analysis model;

[0068] A three-dimensional geometric model containing an epidermis (0.1 mm in thickness), a dermis (1 mm) and a subcutaneous fat layer (3 mm) is established, and the thermal physical parameters of each layer are set: the dermis thermal diffusion coefficient alpha is 0.15 mm 2 / s, and the electrical characteristic parameters: the conductivity sigma is 0.3 S / m; when the tissue impedance change rate exceeds 15% / s is monitored, the heat source distribution is dynamically adjusted through a formula, and finally a thermal field simulation atlas is generated to predict the temperature change trend of each layer of the skin in real time;

[0069] The first control layer takes the preset treatment target: dermal temperature 42-45°C, treatment duration 10 minutes, and safety threshold: upper limit of safe temperature 46°C, impedance fluctuation threshold 20% as input, collects 1000 historical treatment data, selects the top 5 key features including temperature rise rate, impedance interval, etc. by information gain rate, constructs a decision tree rule base, and generates the rule: "if temperature rise rate > 0.3°C / s and tissue impedance ∈ [100-200Ω], then power down by 10% and maintain for 30 seconds";

[0070] The second control layer receives the instructions and maps the power adjustment to the underlying parameters: voltage from 80V to 72V, while dynamically adjusting in combination with the thermal field model prediction results; when the ARIMA model predicts that the temperature rise rate will reach 0.4°C / s in the next 30 seconds, exceeding the threshold 0.3°C / s, power attenuation is triggered in advance: ΔP = 0.5×0.4 = 0.2W, ensuring that the temperature is maintained in the safe interval;

[0071] Continuous collection of parameters under normal treatment, power attenuation, etc. extracts 6 features such as voltage fluctuation coefficient and current distortion rate, selects the top 4 highly correlated features by mutual information method, trains a random forest model with 50 decision trees and a maximum depth of 8; when the model outputs the "poor electrode contact" abnormal classification, start the correlation analysis: trace back the temperature rise rate historical data, find that it rises from 0.1°C / s to 0.5°C / s in 5 minutes, and the impedance change gradient > 2Ω / s, match the rule "temperature sudden rise + impedance mutation → poor electrode contact" in the feature fault mapping table, locate to the loose contact of the hand electrode contact thermistor, generate a visual report: mark the temperature mutation point (t = 5min 40°C→45°C), and prompt "replace the electrode contact thermistor";

[0072] Extract 500 treatment data from the historical database, bin the temperature into <40°C, 40-45°C, >45°C, and bin the impedance into <150Ω, 150-300Ω, >300Ω, generate the rule "tissue impedance ∈ [150-300Ω] and treatment stage = maintenance period, recommend power adjustment amplitude -5%, frequency 1 / min" using the Apriori algorithm, efficacy weight = significant case proportion 85%, safety weight = 1 - adverse event rate (2%) = 98%, maximum weight

[0073] = 85%×0.7 + 98%×0.3 = 90.1%;

[0074] For a patient with a fat layer thickness of 2 mm and a target temperature of 43°C, a rule is matched from the optimized rule base, and an individualized scheme is output: the power adjustment curve is 50 W for a preheating period of 2 minutes and 45 W for a maintenance period of 8 minutes, the safety warning threshold is adjusted to a temperature rise rate of 0.25°C / s (sensitive skin), and the real-time feedback strategy is "automatically trigger power reduction of 10% when the temperature is greater than or equal to 44°C".

[0075] It will be obvious to a person skilled in the art that the application is not limited to the details of the foregoing exemplary embodiments and can be implemented in other concrete forms without departing from the spirit or essential characteristics of the application. The foregoing embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. No reference signs in the claims should be considered as limiting the scope of the claims with respect to the features they indicate.

Claims

1. An intelligent management method for a radio frequency energy output device, characterized in that: The method comprises: Step S100: Deploy integrated sensor arrays on the RF host and handpiece electrodes to synchronously collect four core parameters: load voltage, output current, electrode temperature, and tissue impedance. Align these multi-source parameters using a time synchronization algorithm to construct a dynamic thermal impedance collaborative analysis model. Based on the model output, a three-dimensional thermal field simulation map of the RF energy application area is generated. Combined with the real-time monitored tissue impedance variation characteristics, the heat accumulation trend in the RF energy application area is dynamically predicted. Step S200: Constructing a two-layer control architecture, including a first control layer and a second control layer; the first control layer uses preset treatment target parameters and safety threshold conditions as input, generates a dynamic power adjustment rule base, and outputs control instructions; the second control layer dynamically adjusts the RF energy output parameters according to the control instructions; during the control execution phase, the power adjustment strategy is dynamically implemented based on the prediction results of the dynamic thermal impedance collaborative analysis model; Step S300: During the power regulation process, the operating status parameters of the RF energy output device are continuously collected to form a multi-operating condition data set and construct a fault classification model. Abnormal operating conditions are detected through real-time feature extraction and the fault classification model, and the root cause of the fault is located by combining multi-parameter correlation analysis, and finally a visual diagnostic report is generated. Step S400: Based on the tissue impedance changes, temperature distribution and operation feedback information in the historical treatment data, a mapping relationship between the device operating status and the power control strategy is constructed; the treatment effect and safety indicators are evaluated through an iterative optimization mechanism to generate an energy control plan that adapts to individual needs.

2. The intelligent management method for a radio frequency energy output device according to claim 1, characterized in that: The step S100 includes: Step S101: Integrate a load voltage sampling circuit and a current transformer inside the RF host. The load voltage sampling circuit is composed of a resistor and a transformer to form a voltage division and isolation network. At the same time, a thermistor temperature sensor is built into the handpiece electrode contact end to form a sensor array. When the device is started, four parameters, namely load voltage, output current, electrode temperature and tissue impedance, are synchronously collected at a preset sampling frequency. The load voltage represents the voltage at the RF output end, which is sampled through a resistor circuit after capacitive coupling and then isolated and transmitted by a transformer. The output current represents the current output by the RF, which is converted into a voltage signal by a transformer, and then transmitted through a resistor conversion and a transformer. The electrode temperature represents the temperature of the handpiece electrode contact end, which is detected by a temperature sensor and converted into a voltage value by a circuit. The tissue impedance represents the impedance presented by human tissue and is calculated based on the phase difference between the load voltage and the output current. The collected data is filtered through a low-pass filter circuit to remove high-frequency noise, and then the sampling time of each parameter is aligned using a timestamp synchronization algorithm. Step S102: Establish a three-dimensional geometric model of human skin and subcutaneous tissue, including the epidermis, dermis, and subcutaneous fat layer. Divide the mesh and set the thermal and electrical properties of each layer of tissue. Use the synchronized load voltage, output current, electrode temperature, and tissue impedance parameters as inputs and solve the heat conduction equation based on the principle of electromagnetic thermal coupling: Where α is the thermal diffusivity, σ is the electrical conductivity, E is the electric field strength, ρ is the tissue density, and c p is the specific heat capacity, and T represents the temperature. During the model construction process, when the real-time monitored tissue impedance change rate exceeds the threshold, it is determined that tissue heat accumulation is accelerated, and the heat source distribution is dynamically adjusted using the following formula: where Q n represents the heat source intensity after adjustment, Q represents the heat source intensity before adjustment, ΔZ t represents the impedance fluctuation amplitude, Δt represents the time interval; k is the proportional coefficient, and k∈[0.5, 1]. After the model is built, based on the temperature distribution data calculated by the model, a three-dimensional visualization algorithm is used to map the temperature value of each grid node to an RGB color value, presenting the temperature values ​​and temperature gradients at different positions, thereby generating a thermal field simulation map of the RF energy action area.

3. The intelligent management method for a radio frequency energy output device according to claim 1, characterized in that: The step S100 further includes: Step S103: Real-time acquisition and storage of temperature data for the most recent period, while synchronously recording tissue impedance change data within the corresponding period; fitting the historical temperature data using the exponential smoothing method, and adaptively adjusting the smoothing coefficient according to the tissue impedance fluctuation: when the tissue impedance fluctuation amplitude is less than the preset value, the smoothing coefficient is set to p1; when the tissue impedance fluctuation amplitude is greater than or equal to the preset value, the smoothing coefficient is set to p2; performing first-order difference processing on the fitted temperature series to obtain a stationary series; calculating the autocorrelation function and partial autocorrelation function of the series after difference to determine the order of the ARIMA model; estimating the parameters of the ARIMA model using the maximum likelihood estimation method; predicting the temperature value within a period of time in the future based on the estimated model parameters; defining the temperature rise rate threshold as d t When the predicted temperature rise rate exceeds the threshold within a certain time period, the early warning mechanism is triggered to indicate the risk of heat accumulation.

4. The intelligent management method for a radio frequency energy output device according to claim 1, characterized in that: The step S200 includes: Step S201: The first control layer takes preset treatment target parameters and safety threshold conditions as input. The treatment target parameters include the preset treatment temperature, target energy accumulation value, and treatment duration. The safety threshold conditions include the upper limit of the safe temperature, the abnormal tissue impedance fluctuation threshold, and the power attenuation gradient. The first control layer collects historical treatment data, including input conditions and corresponding power adjustment actions in various treatment scenarios, including the power adjustment amplitude, direction, and duration. For each feature in the input condition, the information gain of each feature is first calculated, which is determined by calculating the entropy difference between the data set before and after the partition. The entropy calculation formula is: Where D is the data set, H(D) represents the entropy of the data set D, and C i is the i-th type of power regulation action, |C i | and |D| represent the number of samples of the i-th type of action and the total number of samples in the data set, respectively; then, the entropy of each feature is calculated; finally, the information gain of each feature is divided by its entropy to obtain the information gain rate, and the top x features with the highest information gain rate are selected as the key nodes of the decision tree; the data set is divided into several subsets according to the different values ​​of the features; at each node, the feature with the largest information gain is selected as the division condition; the above process is recursively repeated for each subset until the preset stopping condition is met; the mapping relationship between conditions and actions is extracted from the constructed decision tree; each path from the root node to the leaf node represents a rule, and the combination of conditions on the path corresponds to a specific power adjustment action; in the actual treatment process, according to the input conditions of real-time monitoring, the appropriate rules are selected from the rule base and the corresponding control instructions are output; finally, the first control layer converts the generated rules into control instruction packets containing power adjustment amplitude, direction and duration; Step S202: Receive the control instruction packet generated by the first control layer, analyze the power adjustment amplitude, direction, and duration parameters; map the power adjustment instruction to the underlying control parameters of the RF generator, including voltage / current amplitude, pulse width modulation duty cycle, and operating mode switching; continuously collect the actual voltage and current at the output of the RF host and the handpiece electrode temperature; combine the thermal field simulation map output by the dynamic thermal impedance collaborative analysis model to obtain the three-dimensional temperature distribution and temperature rise rate of the active area; use the PID control algorithm, using the deviation between the preset treatment temperature and the actual temperature as input, to dynamically adjust the output parameters: Where, u(t) represents the control quantity output by the PID controller; K p , K i , K d are proportional coefficient, integral coefficient, and differential coefficient respectively. e(t) represents the deviation between the preset treatment temperature and the actual temperature. When the temperature rise rate d t When the threshold is exceeded, power reduction is triggered in advance, according to the setting: ΔP=λ×d t , where ΔP represents the power attenuation and λ is the attenuation coefficient.

5. The intelligent management method for a radio frequency energy output device according to claim 1, characterized in that: The step S300 includes: Step S301: Continuously collect operating state parameters of the radio frequency energy output device under different working conditions, including normal treatment, power attenuation, and temperature abnormality, including load voltage, output current, electrode temperature, tissue impedance, and radio frequency power output value, to form a multi-dimensional original data set; the collected data are mapped to the interval [0,1] using the Min-Max normalization method; key features are extracted from the normalized data: for load voltage, the voltage fluctuation coefficient is extracted, that is, the ratio of the standard deviation of the voltage to the mean within a certain period of time; for output current, the current distortion rate is calculated, and the distortion of the current is measured by analyzing the proportion of harmonic components of the current; for electrode temperature, the temperature rise rate is obtained, expressed as the change in temperature per unit time; for tissue impedance, the impedance change gradient is calculated, that is, the change in impedance per unit time; at the same time, the power adjustment frequency, that is, the number of power adjustments per unit time, is extracted to construct a feature vector set; Step S302: Calculate the correlation between each feature and the fault type using the mutual information method, and calculate the mutual information value between each feature and the fault type: MI(X i ,AND j )=∑ xi∈Xi ∑ yj∈Yj p(x i ,and j )log[p(x i ,and j ) / p(x i )p(and j )]; Among them, MI(X i ,Y j ) represents the mutual information value between the i-th feature and the j-th type of fault, X i represents the i-th characteristic variable extracted from the operating state parameters, Y j represents the jth fault type, x i Represents feature X i The specific value of y j Indicates fault type Y j The specific category value of p(x i ,y j ) is X i and Y j The joint probability distribution of p(x i ) and p(y j ) is the marginal probability distribution; sort the features from high to low according to the mutual information value, select the top b features as the model input, and form the final feature subset; where b is a set value; divide the feature subset F and the corresponding fault label Y into training set and test set in a ratio of 7:3; then use the random forest algorithm for training, set the number of decision trees and the maximum depth, and use the fault type as the output label; finally, optimize the model parameters through 10-fold cross-validation, and stop training when the model accuracy and recall rate reach the preset threshold.

6. The intelligent management method for a radio frequency energy output device according to claim 1, characterized in that: The step S300 further includes: Step S303: During the operation of the equipment, set a timing cycle to collect operating parameters in real time, extract feature vectors and input them into the fault classification model; when the model outputs an abnormal classification result, start the multi-parameter association analysis process: based on the top 8 features ranked by mutual information value in step S302, establish a feature fault mapping table, and map the feature anomalies to specific components in combination with the equipment hardware architecture; trace the historical change trajectory of the abnormal features, and locate the specific faulty component in combination with the feature fault mapping table; after the fault is located, the system automatically generates a visual diagnostic report: the parameter change curve within 2 minutes before and after the fault occurs is displayed in the form of a timeline, and the abnormal data points are marked in red; the fault tracing process is presented through a flow chart, and the judgment basis of each layer node is marked.

7. The intelligent management method for a radio frequency energy output device according to claim 1, characterized in that: The step S400 includes: Step S401: extract f cases of treatment data from a historical database, including tissue impedance characteristics, temperature distribution characteristics, and operation feedback information; bin the temperature and impedance data, set different thresholds for the tissue impedance change rate and temperature rise rate, and divide the temperature and impedance into three intervals: low, medium, and high; use the Apriori algorithm to analyze the association rules between the device operating status and the power control strategy: the input items are the tissue impedance interval, temperature distribution characteristics, and treatment stage; the output items are the power adjustment amplitude, adjustment frequency, and duration; calculate the efficacy weight of each association rule, count the number of cases marked as having significant efficacy in the operation feedback when the rule is applied in historical treatments, divide the number by the total number of applications of the rule, and obtain the proportion of cases with significant efficacy as the quantitative value of the efficacy weight; calculate the safety weight, count the number of adverse events that occur during the application of the rule, and subtract the adverse event incidence rate from 1 to obtain the safety weight value, wherein the adverse event incidence rate = the number of adverse event occurrences / the total number of applications; finally, obtain the final weight of each rule through weighted summation; Step S402: define quantitative treatment effect and safety evaluation indicators, the treatment effect indicators include the target tissue temperature compliance rate, that is, the proportion of time during the treatment process that the target tissue temperature is maintained in the preset effective treatment temperature range; safety indicators include the high temperature area ratio and the number of times the temperature rise rate exceeds the standard; based on these indicators, compare the actual parameters of the historical treatment plan with the preset threshold value: if the target tissue temperature compliance rate is greater than or equal to the threshold, it is judged that the treatment effect is effective; if the high temperature area ratio is less than or equal to the threshold and the number of times the temperature rise rate exceeds the standard is ≤2 times, it is judged that the safety standard is met; finally, the plan level is divided based on the efficacy and safety results; the equipment operation state is defined as the state space, and the power regulation strategy is used as the action space. A reward function is constructed based on the treatment effect and safety indicators, and a corresponding reward value is assigned to each action. A positive reward is given if the standard is met, and a negative penalty is given if the standard is not met; during the optimization process, the plan library is updated with a preset number of iterations, and the high-quality plans ranked in the top a in reward value are retained, where a is a set value.

8. The intelligent management method for a radio frequency energy output device according to claim 1, characterized in that: The step S400 further includes: Step S403: Generate an energy control plan based on the patient's individualized parameters. Before treatment, obtain the patient's tissue characteristic parameters and the treatment goals set by the doctor. Based on these individualized inputs, match the most suitable energy control strategy from the optimized plan library, and output an individualized plan that includes a full-process power regulation curve, safety warning threshold adjustment, and real-time feedback strategy.

9. An intelligent management system for a radio frequency energy output device, characterized in that: The system includes a parameter acquisition and thermal field modeling module, a double-layer power control module, a fault detection module, and a personalized solution optimization module; The parameter acquisition and thermal field modeling module deploys an integrated sensor array on the RF host and handpiece electrodes to synchronously collect four core parameters: load voltage, output current, electrode temperature, and tissue impedance. After aligning the multi-source parameters using a time synchronization algorithm, a dynamic thermal impedance collaborative analysis model is constructed. Based on the model output, a three-dimensional thermal field simulation map of the RF energy application area is generated. Combined with the real-time monitoring of tissue impedance change characteristics, the heat accumulation trend in the RF energy application area is dynamically predicted. The dual-layer power control module constructs a dual-layer control architecture, including a first control layer and a second control layer; the first control layer takes preset treatment target parameters and safety threshold conditions as input, generates a dynamic power regulation rule base and outputs control instructions; The second control layer dynamically adjusts the radio frequency energy output parameters according to the control instruction; In the control execution phase, the power regulation strategy is dynamically implemented based on the prediction results of the dynamic thermal impedance collaborative analysis model; During the power regulation process, the fault detection module continuously collects operating status parameters of the RF energy output device to form a multi-operating condition data set and build a fault classification model. It detects abnormal operating conditions through real-time feature extraction and the fault classification model, locates the root cause of the fault through multi-parameter correlation analysis, and finally generates a visual diagnostic report. The personalized solution optimization module constructs a mapping relationship between the device operating status and the power control strategy based on the tissue impedance changes, temperature distribution and operation feedback information in the historical treatment data; Through an iterative optimization mechanism, the treatment effect and safety indicators are evaluated to generate energy control solutions that meet individual needs.

Images