Temperature control system and method based on electronic cigarette

Abstract

The invention discloses a temperature control system and method based on an electronic cigarette, and belongs to the technical field of electronic cigarette temperature control, and the method comprises the steps: building a temperature difference mapping matrix through obtaining a heating body infrared radiation signal and a cigarette oil liquid level heat conductivity change parameter; establishing a liquid level-heating surface coupling dynamic model by taking the matrix as input, and extracting nonlinear temperature control characteristic factors; constructing a multi-dimensional state recognition vector based on the factors, and performing fuzzy reasoning in combination with a real-time flow signal of a user suction sensor to generate a dynamic target temperature control curve; a heating body voltage is driven through pulse width modulation, a thermal coupling state identification value is updated according to infrared change and an evaporation rate in a sucking process, a temperature control curve is dynamically corrected, and closed-loop self-adaptive temperature control is realized; according to the method, the temperature control precision and responsiveness can be improved, the smoking experience is improved, and the intelligent control level of the electronic cigarette is enhanced.

Description

Technical Field

[0001] This invention relates to the field of electronic cigarette temperature control technology, specifically to a temperature control system and method based on electronic cigarettes. Background Technology

[0002] With the widespread use of e-cigarettes as a replacement for traditional cigarettes, heating control technology has become a key factor in ensuring user experience and safety. Current e-cigarettes mostly rely on simple resistance-voltage feedback mechanisms for temperature control. However, when faced with different e-liquid types, changes in ambient temperature, airflow disturbances, and differences in user inhalation strength, traditional control strategies struggle to accurately adjust the heating temperature in real time, easily leading to problems such as e-liquid overheating and decomposition, generation of harmful gases, or insufficient vapor production. Furthermore, existing systems often cannot detect changes in the internal thermal conduction characteristics of the cartridge, resulting in a lack of microscopic state perception in the control strategy, severe temperature lag, and impacting the consistency of vapor quality.

[0003] More seriously, current e-cigarettes generally neglect the dynamic interaction between the heating element surface and the e-liquid surface. Especially after prolonged use, the drop in liquid level causes changes in the heat conduction path. Existing systems cannot recognize this microscopic thermal coupling behavior, resulting in heat concentration, localized overheating, or dry burning, increasing potential safety hazards. At the same time, in multi-flavored e-liquids (such as fruit and mint), the thermal response curves of each component differ significantly. Traditional heating curves cannot adaptively match these differences and lack a flexible temperature control mechanism, easily leading to distorted flavor release. Summary of the Invention

[0004] The purpose of this invention is to provide a temperature control system and method based on electronic cigarettes to address the shortcomings in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a temperature control method based on electronic cigarettes, comprising: The infrared radiation signal on the surface of the electronic cigarette heating element and the thermal conductivity change parameter of the e-liquid surface are obtained, and a temperature difference mapping matrix reflecting the degree of coupling of microscopic thermal behavior is generated based on the infrared radiation signal. Using the temperature difference mapping matrix as input, a coupled dynamic model of liquid surface-heating surface is established, and nonlinear temperature control characteristic factors in the heat conduction path are extracted; A multi-dimensional state recognition vector is constructed based on temperature control feature factors, and fuzzy reasoning is performed in conjunction with the real-time flow signal from the user's suction sensor to generate a dynamic target temperature control curve. The dynamic target temperature control curve is used to drive the pulse width modulation signal to adjust the voltage applied to the heating element in real time. During inhalation, the changes in the infrared radiation signal of the heating element and the evaporation rate of the e-liquid surface are continuously sensed. The thermal coupling status indicator value is updated based on the changes in the evaporation rate, and the dynamic target temperature control curve is iteratively adjusted using the indicator value.

[0006] Preferably, the generation of the temperature difference mapping matrix reflecting the degree of coupling of microscopic thermal behavior based on the infrared radiation signal includes: The infrared radiation intensity values ​​of the heating body surface at different inhalation time periods were obtained, and the intensity values ​​were spatially interpolated to construct a thermal distribution map of the heating area. By spatially projecting and registering the heat distribution map with the e-liquid surface position, the heat conduction response data of the corresponding liquid surface area is extracted; Based on the temporal variation characteristics in the thermal conduction response data, the gradient of the change in the thermal conductivity of the liquid surface per unit time is calculated, and the gradient is used as the coupling strength weighting factor. A two-dimensional temperature difference mapping matrix is ​​constructed by using the coupling weight of infrared radiation intensity and thermal conductivity gradient to reflect the microscopic thermal behavior interaction between the heating element and the e-liquid surface.

[0007] Preferably, the establishment of the liquid surface-heating surface coupled dynamic model includes: A multi-frame heat distribution sequence is constructed from the temperature difference mapping matrix in chronological order, and the temperature difference gradient change map between each frame is extracted by difference operation; By using the thermal gradient variation spectrum, the spatial propagation direction of temperature change and its response delay to changes in liquid level are calculated, and a dynamic state vector describing the coupling relationship between the directionality and time of heat conduction is constructed. The dynamic state vector is input into a nonlinear fitting algorithm based on support vector regression to establish a coupling prediction function between the liquid surface and the heating surface. The high-order nonlinear offset factor and the variable-order response curvature in the function are extracted as temperature control feature factors. The extracted temperature control characteristic factors are used to characterize the trend of thermal conductivity change in the heating path.

[0008] Preferably, the generation of the dynamic target temperature control curve includes: A multidimensional state recognition vector is constructed based on the extracted nonlinear temperature control feature factors. The state recognition vector includes four components: thermal diffusion response curvature, thermal conduction offset amplitude, liquid surface dynamic gradient, and heating delay time. By synchronizing the state recognition vector with the flow signal collected in real time by the user's suction sensor, a state-behavior joint input set is constructed, wherein the flow signal includes three sub-items: suction strength, inhalation duration, and inhalation fluctuation frequency. Perform fuzzy logic reasoning on the joint input set, and map different thermal states and user behaviors to temperature regulation response levels based on the preset fuzzy rule base, and output fuzzy control quantities; The fuzzy control quantity is defuzzified and converted into a specific temperature control target value, thereby generating a dynamic target temperature control curve for controlling the heating element voltage or duty cycle.

[0009] Preferably, the dynamic target temperature control curve is used to drive a pulse width modulation signal to adjust the voltage applied to the heating element in real time, including: By setting the difference between the target temperature range and the current actual temperature, the required temperature rise rate is calculated, and the temperature rise rate is mapped to the duty cycle adjustment range to generate the corresponding duty cycle adjustment command. The dynamic target temperature control curve is discretized over time, and the target temperature value in each time period is extracted as the adjustment node. For each adjustment node, a pulse width modulation signal is generated according to the duty cycle command. The pulse width modulation signal drives the heating circuit at a fixed frequency to adjust the voltage of the heating element in real time. During the heating process, the real-time temperature signal of the heating element is continuously monitored, and the subsequent duty cycle is dynamically adjusted based on the deviation between the actual temperature and the target temperature.

[0010] Preferably, updating the thermal coupling state flag value based on the change in the evaporation rate, and using the flag value to iteratively adjust the dynamic target temperature control curve, includes: Infrared radiation intensity signals from the surface of the heating element are continuously collected, and the derivative of its time-varying behavior is extracted using curve fitting to construct an infrared variation vector that reflects the thermal dynamic response rate of the heating region. Based on liquid level height sensing or evaporation mass estimation methods, the evaporation rate of e-liquid liquid surface per unit time is measured in real time, and its change slope is calculated to characterize the acceleration or deceleration trend of evaporation behavior. The infrared variation vector is fused with the evaporation rate variation slope to generate a thermal coupling state identifier value, which reflects the dynamic change in heat exchange efficiency between the heating element and the e-liquid.

[0011] Preferably, the thermal coupling state identifier value is input as a control factor into the dynamic temperature control curve generation process, and the slope or amplitude of the target temperature curve in the next unit time is dynamically adjusted according to the trend of the identifier value change, so as to realize temperature control correction during the inhalation process.

[0012] The present invention also provides a temperature control system based on electronic cigarettes, comprising: Infrared sensing and thermal behavior acquisition module: acquires the infrared radiation signal on the surface of the electronic cigarette heating element and the thermal conductivity change parameter of the e-liquid surface, and generates a temperature difference mapping matrix that reflects the degree of coupling of microscopic thermal behavior based on the infrared radiation signal. Thermal coupling modeling and feature extraction module: Using the temperature difference mapping matrix as input, a dynamic model of liquid surface-heating surface coupling is established, and nonlinear temperature control feature factors in the heat conduction path are extracted; Temperature control strategy generation module: Constructs a multi-dimensional state recognition vector based on temperature control feature factors, and combines it with the real-time flow signal from the user's suction sensor to perform fuzzy inference and generate a dynamic target temperature control curve; Temperature control execution module: uses the dynamic target temperature control curve to drive the pulse width modulation signal and adjusts the voltage applied to the heating element in real time; Adaptive temperature control correction module: During the smoking process, it continuously senses the changes in the infrared radiation signal of the heating element and the evaporation rate of the e-liquid surface, updates the thermal coupling status flag value according to the changes in the evaporation rate, and uses the flag value to iteratively adjust the dynamic target temperature control curve.

[0013] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention constructs a temperature difference mapping matrix based on infrared radiation signals and changes in thermal conductivity to establish a dynamic model of thermal coupling between the liquid surface and the heating surface, and extracts nonlinear temperature control characteristic factors, significantly improving the perceptibility and controllability of the microscopic heating process of electronic cigarettes. Compared with traditional methods that rely on a single temperature sensor or a fixed heating curve, this invention can achieve dynamic modeling of the thermal response state during inhalation, possessing higher control precision and stronger adaptive capabilities, effectively avoiding problems such as overheating and dry burning or insufficient smoke, and ensuring the consistency of smoke quality and flavor.

[0014] 2. This invention introduces joint modeling of state recognition vectors and user suction behavior signals, and combines this with a fuzzy inference mechanism to generate a dynamic target temperature control curve. Then, it utilizes real-time infrared and evaporation feedback information for closed-loop correction. This invention achieves intelligent, adaptive, and real-time feedback adjustment of electronic cigarette temperature control. This solution not only improves the adaptability of the temperature control system to individual smoking differences and environmental changes, but also enhances the device's response speed and safety. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a flowchart of the method of the present invention.

[0017] Figure 2 This is a flowchart of the system modules of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1, please refer to Figure 1 As shown in this embodiment, a temperature control method based on electronic cigarettes includes: The infrared radiation signal on the surface of the electronic cigarette heating element and the thermal conductivity change parameter of the e-liquid surface are obtained, and a temperature difference mapping matrix reflecting the degree of coupling of microscopic thermal behavior is generated based on the infrared radiation signal. Using the temperature difference mapping matrix as input, a coupled dynamic model of liquid surface-heating surface is established, and nonlinear temperature control characteristic factors in the heat conduction path are extracted; A multi-dimensional state recognition vector is constructed based on temperature control feature factors, and fuzzy reasoning is performed in conjunction with the real-time flow signal from the user's suction sensor to generate a dynamic target temperature control curve. The dynamic target temperature control curve is used to drive the pulse width modulation signal to adjust the voltage applied to the heating element in real time. During inhalation, the changes in the infrared radiation signal of the heating element and the evaporation rate of the e-liquid surface are continuously sensed. The thermal coupling status indicator value is updated based on the changes in the evaporation rate, and the dynamic target temperature control curve is iteratively adjusted using the indicator value.

[0020] Using a short-wave infrared sensor array installed near the heating element, the infrared radiation intensity values ​​of the heating element surface at different stages before, during, and after absorption were continuously recorded at sampling frequencies ranging from 1 to 10 Hz. To improve the spatial resolution of the measurement, the collected discrete infrared signals were spatially reconstructed using a bilinear interpolation method, forming a two-dimensional thermal map containing pixel-level thermal intensity distribution. This thermal map can be regarded as a quantitative description of the temperature radiation intensity at each point per unit area of ​​the heated region.

[0021] A liquid surface positioning model was established, and spatial projection registration was performed on each coordinate point in the aforementioned thermal spectrum to identify the liquid surface region corresponding to the infrared signal. The liquid surface positioning model was constructed based on the heat capacity inversion method: the height of the liquid surface was deduced by using the ratio between the initial liquid temperature rise rate and the heat input per unit volume. In this embodiment, the initial liquid surface position was set as the zero-point reference, and the slope change of the thermal response curve after each heating was used as the dynamic indicator parameter for the liquid surface descent.

[0022] The registered thermal response data is processed over time to extract the temperature change rate sequence within the inhalation cycle. The thermal conductivity gradient is defined as the ratio of the derivative of temperature change with respect to infrared signal intensity per unit time; this ratio reflects the thermal energy conversion efficiency of the e-liquid during its transition from liquid to gaseous state. In actual calculations, the first derivative of the temperature change sequence is calculated using a difference algorithm, and then normalized based on the actual thermal input values ​​to obtain the thermal conductivity gradient array.

[0023] To further establish the thermal coupling relationship between the heating element and the liquid surface, a coupling strength weighting factor is introduced. This factor is defined as the product of the rate of change of infrared radiation intensity and the gradient of thermal conductivity, i.e.: The coupling strength factor is calculated as: Infrared radiation change rate × Thermal conductivity change gradient; where the infrared radiation change rate is the difference in radiation intensity between the same pixel in two adjacent infrared images divided by the time interval, expressed in watts per square meter per second. This product operation yields a weighting factor matrix containing both spatial and temporal dimensions.

[0024] After obtaining the coupling strength factor, it is used as a weight to perform a weighted transformation on the infrared thermogram, generating a two-dimensional temperature difference mapping matrix. Each element in this matrix represents the actual thermal interaction intensity between the heating element and the e-liquid surface at a specific spatial location and during a specific inhalation stage. The unit of the mapping matrix is ​​degrees Celsius, indicating the instantaneous temperature difference at that point during the heat absorption or release process.

[0025] To eliminate interference from ambient temperature fluctuations, a reference temperature calibration mechanism is introduced before the temperature difference mapping matrix is ​​generated. This mechanism calculates the static average radiation value based on the thermal image data within the first 5 seconds of inhalation and uses it as a zero-temperature difference reference line to perform offset correction on each subsequent frame of infrared image.

[0026] Furthermore, to achieve continuous identification of changes in the thermal coupling state, the generated temperature difference mapping matrix is ​​dynamically updated. Specifically, a new set of infrared signals and liquid surface thermal conductivity parameters are acquired every 0.5 seconds, and the above processing procedure is repeated. The latest 10 sets of mapping matrices are retained using a sliding window mechanism, forming a matrix sequence. Principal component analysis is applied to this sequence to extract the dominant thermal change direction and intensity trend, thereby establishing the temporal evolution trajectory of the thermal coupling state.

[0027] In the temperature control method based on electronic cigarettes proposed in this invention, in order to achieve in-depth modeling and temperature regulation optimization of the heat conduction behavior between the heating element and the e-liquid surface, it is necessary to further establish a coupled dynamic model of the liquid surface and heating surface based on the obtained temperature difference mapping matrix. This model can extract temperature control characteristic factors reflecting the dynamic nonlinear changes in the heat conduction path, providing a key basis for subsequent temperature regulation strategies.

[0028] The construction of the coupled dynamic model is based on the spatiotemporal trend analysis of the temperature difference mapping matrix sequence. The temperature difference mapping matrix is ​​a two-dimensional data structure, where each element represents the temperature difference at a specific spatial location at a certain moment, in degrees Celsius. The temperature difference mapping matrices obtained at different time points are arranged in the order of sampling time to form a heat distribution time series. To capture the heat conduction characteristics evolving over time, a frame-by-frame differencing algorithm is used to calculate the rate of temperature difference change at each corresponding point between two adjacent heat maps, thereby generating a temperature gradient change map.

[0029] The differential processing method involves subtracting the temperature difference between the nth frame and the (n-1th)th frame, dividing by the sampling time interval, and obtaining the temperature change rate of each pixel per unit time, thus constructing a temperature gradient map. This map reflects the speed and directionality of heat diffusion during the heating process.

[0030] To reflect the impact of dynamic changes in the liquid level on the heat conduction path, a dynamic state vector is established by measuring the correlation between the center offset of the heat diffusion region and the liquid level height. This vector consists of three elements: first, the temporal change of the coordinates of the heat diffusion centroid; second, the rate of heat expansion in the vertical direction; and third, the time difference between the liquid surface evaporation rate and its thermal response delay.

[0031] The coordinates of the centroid of heat diffusion are obtained by calculating the mass center of the high-temperature region in the temperature gradient diagram, indicating the location where heat is mainly concentrated. The vertical heat expansion rate is estimated by statistically analyzing the rate of temperature change along the normal direction of the liquid surface. The thermal response delay time difference refers to the delay time for the heat distribution to respond after the liquid surface drops, used to assess the time lag of the heat transfer path.

[0032] The dynamic state vector is used as the input variable, and a nonlinear prediction function is constructed using the Support Vector Regression (SVR) method. SVR is a supervised learning method commonly used to handle high-dimensional, nonlinear relationship problems, and it has good generalization ability. The SVR model uses time series data as training samples as input, and its goal is to predict the nonlinear response trend of future thermal diffusion behavior under the background of liquid level drop.

[0033] The training process for SVR includes: Using the historical dynamic state vector as input, the feature values ​​of the temperature difference gradient map at the next time node (such as the location of the maximum value, the angle of the expansion direction, etc.) are used as the target output.

[0034] Radial basis function (RBF) is used as the kernel function to improve the model's ability to fit nonlinear relationships.

[0035] The penalty coefficient C and kernel function width were selected using cross-validation. This is to ensure a balance in regression performance.

[0036] The SVR model is fitted using the training data, and nonlinear predictions are made for the state vector of new inputs.

[0037] The trained SVR model yields a nonlinear prediction function relating liquid level change to thermal diffusion behavior. To further quantify the significant variations in this function, two types of temperature control characteristic factors are extracted: Higher-order nonlinear offset factor: refers to the degree of abrupt change in the derivative at the inflection point of the nonlinear change in the prediction function, reflecting the sensitivity of thermal diffusion behavior to changes in the liquid level.

[0038] Variable-order response curvature factor: refers to the rate of change of curvature of the thermal response curve within a specific time period, representing the trend of heat conduction efficiency over time.

[0039] The two feature factors mentioned above were obtained by performing high-order derivative analysis on the output curve of the SVR model. The specific calculation method is as follows: first, the output curve is fitted with a second derivative to identify the time period with the largest curvature change; then, the ratio of the derivative before and after the local extreme point is calculated in these time periods to quantify the degree of nonlinear shift and response curvature.

[0040] Ultimately, the extracted nonlinear temperature control characteristic factors will serve as dynamic parameters for adjusting the temperature control strategy, participating in the construction of the target temperature curve and the feedback judgment of heating control during subsequent inhalation cycles.

[0041] To achieve intelligent temperature control during e-cigarette use and improve the consistency of user inhalation and flavor reproduction, this invention extracts nonlinear temperature control characteristic factors, further constructs a multi-dimensional state recognition vector, and combines it with real-time flow signals collected by the user's suction sensor to generate a dynamic target temperature control curve using fuzzy inference, thereby achieving precise control of the heating element.

[0042] The core technology of this process lies in establishing a state-behavior collaborative recognition mechanism, which models the thermal state parameters during the inhalation process together with the user's behavior parameters, and uses fuzzy logic to process complex nonlinear and multivariate inputs to generate a dynamic adjustment target for heating intensity, ultimately outputting a dynamic control curve of temperature change over time.

[0043] The multidimensional state recognition vector is based on temperature control feature factors, integrating multiple key dimensions such as thermal diffusion, thermal response, liquid surface characteristics, and time delay. In this embodiment, the state vector is set as a four-dimensional structure, containing the following four components: Thermal diffusion response curvature: refers to the degree of change in the curve shape of heat distribution during spatial propagation. This value is obtained by fitting a thermal boundary curve after edge detection of the temperature difference mapping matrix and statistically analyzing its second derivative. Thermal conduction offset amplitude: This represents the maximum spatial offset distance of the heat center position relative to the initial homogenization zone, in millimeters, and is used to reflect the changing trend of the heat conduction path. Dynamic gradient of liquid surface: refers to the rate of change of the rate of descent of the e-liquid surface per unit time, which indicates the intensity of the interference of the evaporation process on the heat conduction path; Heating delay time: This refers to the time delay from the start of liquid level change to the appearance of temperature difference response in the heated area, measured in milliseconds, and is used to assess thermal response hysteresis.

[0044] The process of generating the state vector is as follows: every 0.5 seconds, the current temperature difference mapping matrix is ​​automatically extracted, and the above four indicators are calculated by combining the liquid level height data and the thermal response curve to form the state vector V=[K1,K2,K3,K4], where K1 to K4 represent the standardized values ​​of the above four dimensions respectively.

[0045] This invention utilizes a micro differential pressure sensor integrated into the cigarette holder to collect the user's suction behavior in real time. The sampling frequency is set to 10 times per second, and the collected signal includes the following three dimensions: Suction strength: Based on the output voltage change of the differential pressure sensor and the standard flow conversion curve, it is converted into a flow rate value in milliliters per second; Inhalation duration: The time interval from when the suction signal rises to when it falls, in milliseconds; Inhalation oscillation frequency: refers to the number of flow rate oscillations that occur during a single inhalation, used to reflect the stability of a user's inhalation behavior.

[0046] The above three data points form a user behavior vector B=[F,T,W], where F is the suction intensity, T is the duration, and W is the inhalation fluctuation frequency.

[0047] To achieve collaborative recognition of state and behavior, the state vector V and behavior vector B are aligned in time and combined to form the input dataset S=[K1,K2,K3,K4,F,T,W], which serves as the input basis for fuzzy inference.

[0048] This invention employs fuzzy control theory to process the aforementioned multidimensional input dataset, achieving a nonlinear mapping between state-behavior and target heating temperature. Fuzzy control mainly includes the following steps: Each item in the input variable S is transformed into a linguistic variable, and multiple fuzzy membership functions are defined. For example: The thermal diffusion response curvature is divided into three levels: "low", "medium" and "high". Suction strength is categorized as "weak", "moderate", and "strong"; The dynamic gradient of the liquid surface is divided into "stable", "fluctuating", and "violent".

[0049] Each level corresponds to a triangular or trapezoidal membership function, with the membership degree ranging from 0 to 1.

[0050] Based on experience with drug use scenarios and extensive data observation, a fuzzy rule base was constructed. Each rule takes the following form: If the thermal diffusion response is "high" and the liquid level gradient is "intense" and the suction strength is "strong", then the temperature control response is "increased heating intensity".

[0051] This invention designs no fewer than 25 multi-input fuzzy rules to cover common combinations of hot states and user behaviors.

[0052] The Mamdani inference mechanism is used to activate fuzzy rules that meet the conditions, and weights are evaluated based on the input membership degree to finally obtain the superposition result of multiple output fuzzy sets.

[0053] The fuzzy output is converted into a specific temperature control target value. The centroid method is used to calculate the expected value of the output result, and the instantaneous temperature control response value is obtained in degrees Celsius.

[0054] The aforementioned fuzzy inference generates a temperature control target value every 0.5 seconds, and multiple consecutive target values ​​are combined to form a complete dynamic temperature control curve. This curve represents the target temperature change trajectory that the heating element should achieve within the current inhalation cycle, and is used to guide the subsequent heating control unit to adjust the voltage or pulse width ratio to achieve fine-tuned temperature regulation.

[0055] The dynamic target temperature control curve is a continuous function with time on the horizontal axis and temperature on the vertical axis, representing the target temperature to be achieved at each moment within the e-cigarette inhalation cycle. To convert this curve into a PWM control signal, the target temperature rise rate must first be calculated, i.e., the rate at which the desired temperature rises relative to the previous control cycle. Let the current target temperature be T(n), and the target temperature of the previous cycle be... If the control period is Δt, then the temperature rise rate R can be expressed as: Temperature rise rate This rate represents the desired rate of temperature change per unit time, measured in degrees Celsius per second. Based on the relationship between thermal power and heating efficiency, the temperature rise rate R is mapped to the required heating intensity level. Then, according to a preset duty cycle correspondence table, the target duty cycle value D is generated. The mapping relationship is obtained through experimental calibration. For example, under standard conditions, a temperature rise of 5 degrees Celsius per second corresponds to a duty cycle of 60%, and a temperature rise of 10 degrees Celsius per second corresponds to a duty cycle of 85%. Based on this, a temperature rise rate-duty cycle mapping curve is constructed, forming a lookup table or regression function.

[0056] The dynamic temperature control curve is a continuously changing function, which is inconvenient to use directly for drive control. Therefore, the curve needs to be discretized over time. Specifically, the curve is sampled at fixed time intervals Δt (e.g., 0.5 seconds), and the target temperature value T(n) for each time interval is extracted, generating N discrete control nodes to form a temperature control sequence: T(1), T(2), ..., T(N). Each target temperature value serves as the temperature target for a temperature control command, and together with the target temperature from the previous moment, it is used to calculate the temperature rise rate, thereby deriving the PWM duty cycle value. This discretization process provides an operable numerical basis for subsequent control.

[0057] Based on the duty cycle D, a pulse width modulation (PWM) signal is generated to control the power supply voltage of the heating element. The PWM signal is a rectangular wave with a fixed frequency and a variable duty cycle, defined as the ratio of the high-level duration to the entire cycle time. The voltage energy per unit cycle is adjusted by changing the high-level duration, thereby controlling the actual heating power.

[0058] In this embodiment, the PWM frequency is set to 1 kHz, and the duty cycle ranges from 0% to 100%. Each discrete temperature control node corresponds to a duty cycle value. The PWM signal is generated in real time by a microcontroller or control chip and output to the heating element circuit to control its conduction time, thereby changing the average voltage and thus achieving temperature regulation.

[0059] For example, when the target duty cycle is 80%, it means that in each cycle, the high level lasts for 0.8 milliseconds and the low level lasts for 0.2 milliseconds; if the target is 30%, then the high level lasts for 0.3 milliseconds and the low level lasts for 0.7 milliseconds. The higher the duty cycle, the stronger the heating power and the faster the temperature rises.

[0060] To ensure that the actual heating temperature remains consistent with the dynamic target temperature control curve, this invention designs a closed-loop control mechanism based on real-time temperature feedback. During each temperature control cycle, the temperature sensor collects the current temperature of the heating element (Tactual) and compares it with the current target temperature (T(n)) to obtain the temperature deviation value E: temperature deviation. If the deviation exceeds the set allowable error threshold (e.g., ±2 degrees Celsius), the deviation value is used as an adjustment factor in the next duty cycle calculation to dynamically correct the target duty cycle. For example, if the current deviation is positive and exceeds the threshold, it indicates insufficient heating, and the duty cycle should be increased; if it is negative, it indicates overheating, and the duty cycle should be decreased. The duty cycle adjustment range can be set through the proportional gain coefficient k: duty cycle adjustment amount. The value of k can be between 0.5% and 2%, depending on the type of e-liquid and the atomizer parameters.

[0061] This feedback process updates once per cycle, effectively preventing abnormal states such as overheating and underheating, and achieving stable temperature tracking.

[0062] During inhalation, a near-infrared or short-wave infrared sensor array continuously collects radiation intensity signals from key areas on the surface of the heating element. The signal sampling frequency can be set to 10 times per second or higher to ensure data continuity and response speed.

[0063] To extract the dynamic characteristics of the radiation signal, the infrared intensity data at each time point is differentially analyzed with the data at the previous time point to obtain the intensity change value per unit time. A time series curve is formed by plotting time on the horizontal axis and infrared intensity on the vertical axis. By applying the first derivative to this curve, the radiation intensity change rate sequence is obtained, denoted as the infrared change vector. The mean of this vector reflects the current thermal response intensity of the heating element, while the slope indicates the stability and dynamism of the heating behavior.

[0064] To eliminate the influence of ambient temperature disturbances, a moving average filter can be used to smooth the raw radiation data. It is recommended to set the filter window width to 3 to 5 sampling points. Simultaneously, an interference threshold should be set; if the rate of change of a sampling point exceeds three times the average value, it is considered an abnormal fluctuation and should be discarded.

[0065] The evaporation rate of e-liquid can be estimated in the following two ways: Indirect heat capacity method: By monitoring the temperature diffusion trend of the area above the liquid surface using an infrared sensor, and combining this with the known specific heat capacity of the e-liquid and the change in liquid level, the amount of liquid sinking per unit time can be deduced.

[0066] Sensor estimation method: Obtain liquid level change data through capacitive or ultrasonic liquid level sensors and calculate the liquid level drop rate.

[0067] Let the current liquid level height be H(n), and the previous time step be... If the sampling period is Δt, then the evaporation rate Ve is: Evaporation rate ; Furthermore, by applying the derivative to the evaporation rate, we obtain the slope Re of the evaporation rate change, which reflects the accelerating or decelerating trend of the evaporation rate. A positive value indicates accelerated evaporation, a negative value indicates decelerated evaporation, and a zero value indicates a stable evaporation rate.

[0068] The thermal coupling state identifier value is used to comprehensively reflect the coupling relationship between the dynamic characteristics of infrared radiation and the evaporation rate behavior. In this invention, this identifier value is defined in a fusion manner as: Identifier Value ;in: is the mean or slope of the infrared variation vector, representing the intensity of the surface temperature response of the heating element; Re is the slope of the evaporation rate change; α and β are weighting factors, representing the relative influence of the infrared signal and evaporation rate on the thermal coupling state, and can be set to the default value. , Fine-tuning can be made based on the actual type of e-liquid or the characteristics of the atomizer.

[0069] The Sc value is essentially a quantitative indicator with dimensionless units, used to describe the changes in heat conduction efficiency and heat response rate under the current inhalation state.

[0070] The identifier value range can be set to [-1, 1], where: A value close to 1 indicates a rapid thermal response and intense evaporation, which may pose a risk of overheating. A value close to 0 indicates that the thermal response and evaporation are in equilibrium, which is an ideal state. A value close to -1 indicates a lag in thermal response or slow evaporation, which may result in insufficient smoke.

[0071] Based on the changes in the aforementioned identifier values, the dynamic target temperature control curve for the current and subsequent stages is iteratively corrected. The correction strategy is as follows: Curve slope adjustment method: If Sc > positive threshold (e.g., +0.6), it indicates increased evaporation and potentially excessive heating intensity. In this case, reduce the slope of the temperature control curve to slow the rate of temperature rise; if Sc < negative threshold... This indicates a heating lag, and the slope of the temperature control curve should be increased. Amplitude fine-tuning method: When Sc deviates from zero for a long time (e.g., for 3 consecutive sampling periods), the overall temperature amplitude of the curve can be adjusted, such as decreasing by 3 degrees Celsius or increasing by 5 degrees Celsius. Cycle adaptive method: In the later part of the inhalation cycle, if the Sc fluctuation tends to stabilize, the temperature control adjustment cycle can be appropriately shortened to increase the response frequency and improve the temperature control sensitivity.

[0072] The iterative update process can be performed every 0.5 seconds. The temperature control processor reads the current flag value and updates the target value of the next time period of the temperature control target curve according to the preset rules, thereby dynamically generating new control commands and driving the PWM duty cycle adjustment.

[0073] Example 2, please refer to Figure 2 As shown in the figure, the temperature control system based on electronic cigarettes described in this embodiment includes: Infrared sensing and thermal behavior acquisition module: acquires the infrared radiation signal on the surface of the electronic cigarette heating element and the thermal conductivity change parameter of the e-liquid surface, and generates a temperature difference mapping matrix that reflects the degree of coupling of microscopic thermal behavior based on the infrared radiation signal. Thermal coupling modeling and feature extraction module: Using the temperature difference mapping matrix as input, a dynamic model of liquid surface-heating surface coupling is established, and nonlinear temperature control feature factors in the heat conduction path are extracted; Temperature control strategy generation module: Constructs a multi-dimensional state recognition vector based on temperature control feature factors, and combines it with the real-time flow signal from the user's suction sensor to perform fuzzy inference and generate a dynamic target temperature control curve; Temperature control execution module: uses the dynamic target temperature control curve to drive the pulse width modulation signal and adjusts the voltage applied to the heating element in real time; Adaptive temperature control correction module: During the smoking process, it continuously senses the changes in the infrared radiation signal of the heating element and the evaporation rate of the e-liquid surface, updates the thermal coupling status flag value according to the changes in the evaporation rate, and uses the flag value to iteratively adjust the dynamic target temperature control curve.

[0074] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A temperature control method based on electronic cigarettes, characterized in that: include: The infrared radiation signal on the surface of the electronic cigarette heating element and the thermal conductivity change parameter of the e-liquid surface are obtained, and a temperature difference mapping matrix reflecting the degree of coupling of microscopic thermal behavior is generated based on the infrared radiation signal. Using the temperature difference mapping matrix as input, a coupled dynamic model of liquid surface-heating surface is established, and nonlinear temperature control characteristic factors in the heat conduction path are extracted; A multi-dimensional state recognition vector is constructed based on temperature control feature factors, and fuzzy reasoning is performed in conjunction with the real-time flow signal from the user's suction sensor to generate a dynamic target temperature control curve. The dynamic target temperature control curve is used to drive the pulse width modulation signal to adjust the voltage applied to the heating element in real time. During inhalation, the changes in the infrared radiation signal of the heating element and the evaporation rate of the e-liquid surface are continuously sensed. The thermal coupling status indicator value is updated based on the changes in the evaporation rate, and the dynamic target temperature control curve is iteratively adjusted using the indicator value.

2. The temperature control method based on electronic cigarettes according to claim 1, characterized in that: The generation of the temperature difference mapping matrix reflecting the degree of coupling of microscopic thermal behavior based on infrared radiation signals includes: The infrared radiation intensity values ​​of the heating body surface at different inhalation time periods were obtained, and the intensity values ​​were spatially interpolated to construct a thermal distribution map of the heating area. By spatially projecting and registering the heat distribution map with the e-liquid surface position, the heat conduction response data of the corresponding liquid surface area is extracted; Based on the temporal variation characteristics in the thermal conduction response data, the gradient of the change in the thermal conductivity of the liquid surface per unit time is calculated, and the gradient is used as the coupling strength weighting factor. A two-dimensional temperature difference mapping matrix is ​​constructed by using the coupling weight of infrared radiation intensity and thermal conductivity gradient to reflect the microscopic thermal behavior interaction between the heating element and the e-liquid surface.

3. The temperature control method based on electronic cigarettes according to claim 2, characterized in that: The establishment of the coupled dynamic model of the liquid surface and the heating surface includes: A multi-frame heat distribution sequence is constructed from the temperature difference mapping matrix in chronological order, and the temperature difference gradient change map between each frame is extracted by difference operation; By using the thermal gradient variation spectrum, the spatial propagation direction of temperature change and its response delay to changes in liquid level are calculated, and a dynamic state vector describing the coupling relationship between the directionality and time of heat conduction is constructed. The dynamic state vector is input into a nonlinear fitting algorithm based on support vector regression to establish a coupling prediction function between the liquid surface and the heating surface. The high-order nonlinear offset factor and the variable-order response curvature in the function are extracted as temperature control feature factors. The extracted temperature control characteristic factors are used to characterize the trend of thermal conductivity change in the heating path.

4. The temperature control method based on electronic cigarettes according to claim 3, characterized in that: The generated dynamic target temperature control curve includes: A multidimensional state recognition vector is constructed based on the extracted nonlinear temperature control feature factors. The state recognition vector includes four components: thermal diffusion response curvature, thermal conduction offset amplitude, liquid surface dynamic gradient, and heating delay time. By synchronizing the state recognition vector with the flow signal collected in real time by the user's suction sensor, a state-behavior joint input set is constructed, wherein the flow signal includes three sub-items: suction strength, inhalation duration, and inhalation fluctuation frequency. Perform fuzzy logic reasoning on the joint input set, and map different thermal states and user behaviors to temperature regulation response levels based on the preset fuzzy rule base, and output fuzzy control quantities; The fuzzy control quantity is defuzzified and converted into a specific temperature control target value, thereby generating a dynamic target temperature control curve for controlling the heating element voltage or duty cycle.

5. The temperature control method based on electronic cigarettes according to claim 1, characterized in that: Using the dynamic target temperature control curve to drive the pulse width modulation signal and adjust the voltage applied to the heating element in real time includes: By setting the difference between the target temperature range and the current actual temperature, the required temperature rise rate is calculated, and the temperature rise rate is mapped to the duty cycle adjustment range to generate the corresponding duty cycle adjustment command. The dynamic target temperature control curve is discretized over time, and the target temperature value in each time period is extracted as the adjustment node. For each adjustment node, a pulse width modulation signal is generated according to the duty cycle command. The pulse width modulation signal drives the heating circuit at a fixed frequency to adjust the voltage of the heating element in real time. During the heating process, the real-time temperature signal of the heating element is continuously monitored, and the subsequent duty cycle is dynamically adjusted based on the deviation between the actual temperature and the target temperature.

6. The temperature control method based on electronic cigarettes according to claim 1, characterized in that: The thermal coupling state flag value is updated based on the change in the evaporation rate, and the dynamic target temperature control curve is iteratively adjusted using the flag value, including: Infrared radiation intensity signals from the surface of the heating element are continuously collected, and the derivative of its time-varying behavior is extracted using curve fitting to construct an infrared variation vector that reflects the thermal dynamic response rate of the heating region. Based on liquid level height sensing or evaporation mass estimation methods, the evaporation rate of e-liquid liquid surface per unit time is measured in real time, and its change slope is calculated to characterize the acceleration or deceleration trend of evaporation behavior. The infrared variation vector is fused with the evaporation rate variation slope to generate a thermal coupling state identifier value, which reflects the dynamic change in heat exchange efficiency between the heating element and the e-liquid.

7. A temperature control method based on electronic cigarettes according to claim 6, characterized in that: The thermal coupling state identifier value is input as a control factor into the dynamic temperature control curve generation process, and the slope or amplitude of the target temperature curve in the next unit time is dynamically adjusted according to the trend of the identifier value change, so as to realize the temperature control correction during the inhalation process.

8. A temperature control system based on an electronic cigarette, used to implement the temperature control method based on an electronic cigarette as described in any one of claims 1-7, characterized in that: include: Infrared sensing and thermal behavior acquisition module: acquires the infrared radiation signal on the surface of the electronic cigarette heating element and the thermal conductivity change parameter of the e-liquid surface, and generates a temperature difference mapping matrix that reflects the degree of coupling of microscopic thermal behavior based on the infrared radiation signal. Thermal coupling modeling and feature extraction module: Using the temperature difference mapping matrix as input, a dynamic model of liquid surface-heating surface coupling is established, and nonlinear temperature control feature factors in the heat conduction path are extracted; Temperature control strategy generation module: Constructs a multi-dimensional state recognition vector based on temperature control feature factors, and combines it with the real-time flow signal from the user's suction sensor to perform fuzzy inference and generate a dynamic target temperature control curve; Temperature control execution module: uses the dynamic target temperature control curve to drive the pulse width modulation signal and adjusts the voltage applied to the heating element in real time; Adaptive temperature control correction module: During the smoking process, it continuously senses the changes in the infrared radiation signal of the heating element and the evaporation rate of the e-liquid surface, updates the thermal coupling status flag value according to the changes in the evaporation rate, and uses the flag value to iteratively adjust the dynamic target temperature control curve.

Images