Dielectric heating aerosol-generating device with indirect parameter estimation

Abstract

Provided is an aerosol-generating device (120) for dielectrically heating an aerosol-forming substrate (110. 1). The aerosol-generating device (120) comprises a dielectric heater assembly (130, 135) having a heating region (140) for receiving an aerosol-forming article (105). The aerosol-generating device (120) further comprises an radio frequency source (150) for supplying power to the dielectric heater assembly (130, 135) to generate the aerosol over a heating period when the aerosol-forming article (105) is received in the heating region (140). The aerosol-generating device (120) also comprises one or more sensors (185, 187, 188) configured to measure one or more input parameters associated with operating conditions of the aerosol-generating device (120) during the heating period of the aerosol-generating device (120). A control unit (180) comprises a model comprising data indicative of a relationship between the one or more input parameters and an output variable over a range of input parameter values. The control unit (180) is configured to: receive the one or more measured input parameters; and indirectly estimate, using the model, an output variable value based on the received one or more measured input parameters.

Description

[0001] P / 90486.W001

[0002] Dielectric Heating Aerosol-generating Device With Indirect Parameter Estimation

[0003] The present disclosure relates to aerosol-generating devices, and specifically to aerosolgenerating devices configured to heat an aerosol-forming substrate by dielectric heating. Specifically, the disclosure relates to dielectric heating aerosol-generating devices and systems with parameter detection and methods for aerosol-generation using dielectric heating systems.

[0004] Known electrically operated aerosol-generating systems typically heat an aerosol-forming substrate by one or more of: conduction of heat from a heating element to an aerosol-forming substrate, radiation of heat from a heating element to an aerosol-forming substrate or drawing heated air through an aerosol-forming substrate. Most commonly, heating is achieved by passing an electrical current through an electrically resistive heating element, giving rise to Joule heating of the heating element. Inductive heating systems have also been proposed, in which Joule heating occurs as a result of eddy currents induced in a susceptor heating element.

[0005] A problem with these heating mechanisms is that they may give rise to non-uniform heating of the aerosol-forming substrate. The portion of the aerosol-forming substrate closest to the heating element is heated more quickly or to a higher temperature than portions of the aerosolforming substrate more remote from the heating element.

[0006] Systems that dielectrically heat an aerosol-forming substrate have been proposed, which advantageously provide uniform heating of the aerosol-forming substrate. However, known dielectric heating systems are less efficient than inductive heating systems and require complex electrical circuitry to achieve the necessary voltages and frequencies for dielectric heating of an aerosol-forming substrate.

[0007] A further problem with dielectric heating systems is the difficulty in monitoring parameters of the heating operation to control heating power without greatly adding to the cost and complexity of the system.

[0008] It would be desirable to provide an effective solution for monitoring heating operation parameters in dielectric heating systems to facilitate improved heating power control.

[0009] According to the present disclosure, there is provided an aerosol-generating device for dielectrically heating an aerosol-forming article to generate an aerosol including any one or more of the features described below.

[0010] The aerosol-generating device may comprise a dielectric heater assembly having a heating region for receiving an aerosol-forming article. The aerosol-generating device may comprise radiofrequency source, for example an oscillation circuit or an electromagnetic field generator, for example a solid-state transistor based electromagnetic field generator generating an P / 90486.WG01 electromagnetic filed having a frequency of 2.4GHz, 4.8GHz, or other, for supplying power to the dielectric heater assembly to generate the aerosol over a heating period when the aerosol-forming article is received in the heating region. The aerosol-generating device may further comprise one or more sensors configured to measure one or more input parameters associated with operating conditions of the aerosol-generating device during the heating period of the aerosol-generating device. The aerosol-generating device may further comprise a control unit. The control unit may comprise a model comprising data indicative of a relationship between the one or more input parameters and an output variable over a range of input parameter values. The control unit may be configured to receive the one or more measured input parameters. The control unit may be configured to indirectly determine, using the model, an output variable value based on the received one or more measured input parameters.

[0011] As used herein, the term “aerosol-generating device” relates to a device that interacts with an article comprising an aerosol-forming substrate to generate an aerosol.

[0012] As used herein, the term “aerosol-forming substrate” relates to a substrate capable of releasing volatile compounds that can form an aerosol. Such volatile compounds can be released by heating the aerosol-forming substrate.

[0013] As used herein, the term “indirectly determine an output variable” relates to determining an output variable using input parameters associated with variables which are different from the output variable. For example, determining an internal aerosol-forming substrate temperature without directly measuring the internal aerosol-forming substrate temperature.

[0014] According to an example of the disclosure, provided is an aerosol-generating device for dielectrically heating an aerosol-forming substrate. The aerosol-generating device comprises a dielectric heater assembly having a heating region for receiving an aerosol-forming article. The aerosol-generating further device comprises a radio frequency source, for example an oscillation circuit or an electromagnetic field generator for supplying power to the dielectric heater assembly to generate the aerosol over a heating period when the aerosol-forming article is received in the heating region. The aerosol-generating device further comprises one or more sensors configured to measure one or more input parameters associated with operating conditions of the aerosolgenerating device during the heating period of the aerosol-generating device. The aerosolgenerating device further comprises a control unit. The control unit comprises a model comprising data indicative of a relationship between the one or more input parameters and an output variable over a range of input parameter values. The control unit is configured to receive the one or more measured input parameters. The control unit is further configured to indirectly determine, using the model, an output variable value based on the received one or more measured input parameters. P / 90486.W001

[0015] In dielectric heating aerosol-generating systems variables associated with heating operations can be difficult or costly to measure directly, particularly with respect to properties of an aerosol-forming article received in a heating region of an aerosol-generating device. Aerosolgenerating devices according to the present disclosure are configured to address this challenge by indirectly determining output variables based on input parameters which can be easier or less costly to measure during heating operations. This is possible with the use of a control unit comprising a model for mapping the measurable input parameters to the output variable. The indirectly determined output variables may then be used to implement effective control of heating operations in the aerosol-generating device.

[0016] In some examples, the model comprises a database comprising a plurality of known output variable values associated with a range of input parameter values during a heating period of an aerosol-generating device. In some examples, the database comprises a one-to-one correspondence between a plurality of values for the one or more input parameters and a plurality of output variable values. In some examples, the model is configured interpolate or extrapolate an output variable value based on the measured one or more input parameters and the input parameter-output variable correspondence in the database.

[0017] In some examples, the model comprises an artificial neural network trained using a plurality of known output variable values associated with a range of input parameter values for the one or more input parameters during the heating period of an aerosol-generating device.

[0018] In some examples, the model may comprise a mapping function configured to estimate the output variable value based on the measured one or more input parameters. In some examples, the mapping function may be derived from an artificial neural network trained using a plurality of known output variable values associated with a range of input parameter values for the one or more input parameters during the heating period of an aerosol-generating device. Any known artificial neural network training technique known in the art may be used to derive the mapping function and apply the appropriate parament weightings to obtain an accurate output variable value estimation.

[0019] In some examples, the input parameters are associated with a power consumption of the radio-frequency source, for example an oscillation circuit or an electromagnetic field generator. In particular, the input parameters are associated with dielectric heating losses in the dielectric heater assembly. In some examples, the input parameters are associated with switching losses cause by the oscillation circuit, or the solid state transistor of an electromagnetic field generator. In some examples, input parameters are associated with both dielectric heating losses in the dielectric heater assembly and switching losses in the oscillation circuit are used to determine the output P / 90486.W001 variable value. In some examples, a first input parameter may be associated with switching losses in the oscillation circuit and a second input parameter may be associated with dielectric heating losses in the dielectric heating assembly. In some examples, a first input parameter may be associated with switching losses in the oscillation circuit and a second input parameter may be associated with a power consumption of the oscillation circuit.

[0020] In some examples, the output variable is not directly measurable using the one or more sensors.

[0021] In some examples, the one or more sensors comprises one or more temperature sensors. In an example, the one or more temperature sensors are configured to measure a temperature at or inside the heating region of the dielectric heating assembly. In an example, the one or more temperature sensors are configured to measure a temperature at a switching unit of the oscillation circuit. In some examples, the one or more temperature sensors includes temperature sensor positioned at both the switching unit of the oscillation circuit and the heating region of the dielectric heating assembly. In some examples, the one or more temperature sensors are configured to measure an ambient temperature at, inside, or around the aerosol-generating device. In such examples, the one or more input parameters comprise one or more of a temperature at or inside the heating region of the dielectric heating assembly, and an ambient temperature at, inside, or around the aerosol-generating device.

[0022] In some examples, the one or more sensors comprises one or more voltage or current sensors. The one or more voltage and current sensor may be configured to measure a power consumption of one of or both the oscillation circuit and the dielectric heater assembly. In some examples, the one or more voltage or current sensors are configured to measure electrical characteristics of the oscillation circuit. In some examples, the electrical characteristics include one or more of oscillation frequency, peak voltage amplitude and peak current amplitude. In such examples, the one or more input parameters comprise on or more of a power consumption of one of or both the oscillation circuit and the dielectric heater assembly, an oscillation frequency, a peak voltage amplitude and a peak current amplitude of the oscillation circuit.

[0023] In some examples, the one or more sensors comprises a detection device for determining a type of aerosol-forming article. In some examples, the detection device may comprise one or more sensors configured to measure load-dependent parameters of the electric response of the electrically resonant oscillation system including the oscillation circuit and the dielectric heater assembly. In some examples, the one or more sensors are configured to measure one or more of a quality factor and bandwidth of the electrically resonant oscillation system. P / 90486.W001

[0024] In some examples, the one or more sensors comprises one or more airflow or pressure sensors configured to determine an airflow or pressure related to an airflow in an airflow channel of the aerosol-generating device. In such examples, the one or more input parameters comprises an airflow or pressure related to an airflow in the airflow channel of the aerosol-generating device.

[0025] In some examples, the one or more sensors comprises one or more electric field sensors configured to measure a value indicative of an electric field strength in the vicinity of the dielectric heater assembly.

[0026] In some examples, the one or more sensors comprises an inertial measurement unit configured to determine an orientation of the aerosol-generating device. In such examples, the one or more input parameters comprises one or more of an orientation and acceleration of the aerosolgenerating device.

[0027] In some examples, the one or more sensors comprises a humidity sensor configured to provide a humidity value at, inside, or around the aerosol-generating device.

[0028] In some examples, the one or more sensors comprises a sensor for generating geographic context data. In such examples, the one or more input parameters comprises geographic context data.

[0029] It will be appreciated that the one or more sensors may comprise any combination of sensors described in the above examples based on the output variable to be estimated.

[0030] In some examples, the control unit is also configured to receive additional input data and estimate the output variable based on the additional input data in combination with the measured input parameters. In some examples, the additional input data comprises time data associated with the dielectric heating cycle. Such data may be used to associated the measured input parameters with a specific time moment in the dielectric heating cycle. In some examples, the additional input data comprises one or more of a puff count history, a puff volume history, and an aerosol-forming substrate depletion status. In some examples, the additional input data comprises geographic and / or weather data In some examples, the one or more input parameters comprises one or more of geographic data, and weather data. In some examples, the additional input data comprises reference data associated with the one or more input parameters. The reference data may be compared with the measured input parameters for determining the output variable value.

[0031] Based on the input parameters measured, one or more output variables associated with the dielectric heating operation can be measured.

[0032] In some embodiments, the output variable to be estimated is a temperature inside the aerosol-forming article. P / 90486.W001

[0033] In some embodiments, the output variable to be estimated is a presence or absence of an aerosol-generating article in the heating region of the aerosol-generating device.

[0034] In some embodiments, the output variable to be estimated is a puff occurrence.

[0035] In some embodiments, the output variable to be estimated is a puff volume.

[0036] In some embodiments, the output variable to be estimated is a cumulative puff volume during the inhalation session period.

[0037] In some embodiments, the output variable to be estimated is a depletion level of the aerosol-forming substrate.

[0038] In some embodiments, the output variable to be estimated is a an identification of the aerosol-forming article. For examples, the output variable may be an aerosol-forming article type.

[0039] In some embodiments, the output variable to be estimated is a humidity level inside the aerosol-forming article.

[0040] In some embodiments, the output variable to be estimated is an orientation detection of the aerosol-forming article (right or wrong insertion orientation, full or partial insertion).

[0041] In some embodiments, the output variable to be estimated is a detection whether the aerosol-forming article has been previously used and / or depleted.

[0042] In some embodiments, the output variable to be estimated is a temperature profile or gradient inside the aerosol-forming substrate.

[0043] It will be appreciated that numerous combinations of the described input parameters and output variables are possible, however preferred examples are set out below.

[0044] In some examples, the one or more input parameters comprises an average electric field strength between an electrode arrangement of the dielectric heater assembly, wherein the estimated output variable is a temperature inside the aerosol-forming article. In some examples, the one or more input parameters further comprises one or more of a temperature within the heating region of the dielectric heater assembly, a time since the start of the dielectric heating operation, and a temperature at a switching unit in the oscillation circuit.

[0045] In some examples, the one or more input parameters comprises a temperature measurement from outside the aerosol-forming substrate, wherein the output variable is a temperature inside the aerosol-forming article. In some examples, the temperature measurement can include one or more of a temperature at one or more locations within the heating region of the dielectric heater assembly, and a temperature at a switching unit in the oscillation circuit.

[0046] In some examples, the one or more input parameters comprises a temperature of a switching unit in the oscillation circuit, and wherein the output variable is a temperature inside the aerosol-forming article. In some examples, the temperature measurement can include one or more P / 90486.W001 of a temperature at one or more locations within the heating region of the dielectric heater assembly.

[0047] In some examples, the one or more input parameters comprises an electric field strength at one or more locations within the heating region, wherein the output variable is a temperature inside the aerosol-forming article. In some examples, the one or more input parameters further comprises a power supplied to the oscillation circuit during the heating operation. In some examples, the one or more input parameters further comprises a temperature within the heating region of the dielectric heater assembly.

[0048] In some examples, the dielectric heater assembly comprises an asymmetrical electrode arrangement, configured to produce a non-uniform electric field across the heating region. In some examples, the dielectric heater assembly comprises an electrode arrangement having two opposing electrodes, wherein the two electrodes are non-parallel with each other.

[0049] In some examples, the one or more input parameters comprises a temperature in a plurality of different locations within the heating region of the dielectric heater assembly in order to estimate a temperature distribution profile along an aerosol-forming article. In such examples, the control unit may be further configured to determine, based on the temperature distribution profile, a location within the aerosol-forming article which is at a target temperature.

[0050] In some examples, the control unit is configured to collected data on the one or more input parameters over a time period to generate a corresponding one or more input parameter profiles for the dielectric heating operation over the time period. In some examples, the control unit is configured to perform data processing on the one or more input parameter profiles. In some examples, the control unit is configured to determine a first order derivative of the one or more input parameter profiles. In some examples, the control unit is configured to determine a second order derivative of the one or more input parameter profiles. In some examples, the control unit is configured to determine an average value of the one or more input profiles. In some examples, the control unit is configured to determine a median value of the one or more input parameter profiles. In some examples, the control unit is configured to filter the one or more input parameter profiles. In some examples, the control unit is configured to compare the one or more input parameter profiles with one or more input parameter reference profiles stored in memory.

[0051] In some examples, the control unit comprises a counter, wherein a counting value from the counter during the heating period is used as an additional input for the control unit to estimate the output variable value.

[0052] In some examples, the one or more input parameters comprises an oscillation frequency of the oscillation circuit. P / 90486.WG01

[0053] In some examples, the one or more input parameters comprises an oscillation frequency of the oscillation circuit, wherein the estimated output variable is a temperature inside the aerosolforming article. In some examples, the one or more input parameters further comprises a temperature within the heating region of the dielectric heater assembly, a time since the start of the dielectric heating operation, and a temperature at a switching unit in the oscillation circuit.

[0054] In some examples, the one or more input parameters comprises an oscillation frequency of the oscillation circuit, wherein the output variable is aerosol-forming article identity. In such examples, additional input data may comprise In some examples, the one or more input parameters further comprises one or more of a temperature within the heating region of the dielectric heater assembly, a time since the start of the dielectric heating operation, and a temperature at a switching unit in the oscillation circuit.

[0055] In some examples, the one or more input parameters comprises an oscillation frequency of the oscillation circuit, and wherein the output variable is aerosol-forming article substrate depletion level. In some examples, the one or more input parameters further comprises one or more of a temperature within the heating region of the dielectric heater assembly, a time since the start of the dielectric heating operation, and a temperature at a switching unit in the oscillation circuit.

[0056] Aerosol formers such as glycerol, polypropene glycol (PG) and vegetable glycerin (VG) exhibit similar dielectric properties that vary with temperature and frequency. Generally, for heat- not-burn (HnB) products, the large portion of the relative permittivity of the aerosol-forming substrate is given by the infusion or presence of the aerosol former, as generally the tobacco or non-tobacco-based substrate base has a substantially lower relative permittivity (dry tobacco about 2.5 to 3). Since such aerosol formers are commonly used to infuse tobacco substrates or other non-tobacco based substrates for HnB products and as a main liquid substrate component, the dielectric response of an aerosol-forming substrate during a heating operation can be indicative of parameters associated with the heating operation.

[0057] In a dielectric heating aerosol-generating device that operates with an inverting switching unit and an LC feedback loop having the load capacitor within the feedback loop, the oscillation frequency, typically in a range between 100MHz and 1 GHz, will vary with a change of the capacitance value of the load capacitor. Given the presence of aerosol formers and some other highly dielectric liquids such as water inside aerosol-forming substrates, the oscillation frequency of the oscillator will vary upon heating the aerosol-forming substrate.

[0058] In some examples, the control unit is configured to estimate the output variable value based on a comparison of the oscillation frequency profile for a dielectric heating operation with a P / 90486.W001 reference profile stored in the control unit. In some examples, the reference profile corresponds to a frequency profile from a previous dielectric heating operation.

[0059] In some examples, the control unit is configured to estimate the output variable value based on a comparison of one or more properties associated with the frequency profile with one or more reference frequency profile properties.

[0060] In some examples, the control unit is communicatively coupled with a database comprising stored reference data. The reference data may comprise one or more of reference profiles, reference properties and reference values.

[0061] In some examples, the one or more properties comprises a frequency change falling below a threshold value. During an initial start-up heating phase, the oscillation frequency falls as the aerosol former is heated, causing its dielectric constant to increase and consequently leading to a high load capacitance. The oscillation frequency will continue to fall until a predeterminable temperature is reached at which the aerosol former experiences no further change in relative permittivity. As such, the point at which the oscillation frequency ceases to fall can serve as a temperature and timing marker point. Such information can also be used to determine a type of aerosol-forming substrate based on the dielectric response of the aerosol former. After this point, there is a gradual increase of the oscillation frequency resulting from a decrease in relative permittivity of the aerosol-forming substrate, due to some temperature-related evacuation of liquids from the aerosol-forming substrate. With the further advancing of time, the oscillation frequency increase accelerates, as higher temperatures are reached and some vaporization begins to happen, notably the vaporization of water at 100°C. The temperature-related depletion of liquids from aerosol-forming substrate thereby counteracts the temperature dependent increase of the relative permittivity. The extent of the increase in oscillation frequency during this period can be indicative of a humidity content of the aerosol-forming substrate.

[0062] In some examples, the one or more properties comprises a frequency change exceeding a threshold value. Following the initial start-up heating phase when the device reaches a steadystate temperature, observable increases in oscillation frequency occur during user puffs. This is due to a sudden evacuation of aerosol-former from the aerosol-forming substrate, which leads to a sudden decrease of the relative permittivity of aerosol-forming substrate and causes a sharp increase in oscillation frequency. As such, rapid increases in oscillation frequency can be used to detect not only user puff, but also user puff volumes based on the extent of the oscillation frequency increase. Such information may be collated in order to update a puff count, and a depletion status for the aerosol-forming substrate. P / 90486.W001

[0063] During the time between puffs, the oscillation frequency slightly decreases again due to the cooling effect to the aerosol-forming substrate from the colder incoming air drawn by the puff. Therefore, in some examples, this reduction in oscillation frequency may be used to indicate the end of a puff. Shortly after this cooling effect, the oscillation frequency starts increasing due to constant evaporation of aerosol-former at the steady-state temperature. Over subsequent heating operations, the oscillation frequency increases with increased depletion of the aerosol-forming substrate, until the difference between the oscillation frequency after and before a puff is no longer observable. Therefore, a post-puff oscillation frequency change falling below a threshold value can be used to indicate that the aerosol-forming substrate is fully depleted of aerosol-former.

[0064] The use of the first order derivative of the frequency profile (and also the second order derivative to further enhance detection and estimation results) can be used to estimate different parameters around the aerosolization process of the aerosol-forming substrate.

[0065] In some examples, the control unit is further configured to determine a first order derivative of the frequency profile, and determine the one or more parameters related to the dielectric heating operation based on one or more properties associated with the first order derivative of the frequency profile. In some examples, the one or more properties comprises a zero-crossing of the first order derivative of the frequency profile. Specifically, the zero-crossing of the first order derivative of the frequency profile can be used as a temperature and timing marker point during the initial start-up heating phase that the aerosol-forming substrate has reached an initial preheating temperature.

[0066] In some examples, the one or more properties comprises an amplitude of the first order derivative of the frequency profile. Specifically, the amplitude of the first order derivative of the frequency profile can be used to detect one or more of a puff, a puff volume, puff number and a depletion status for the aerosol-forming substrate.

[0067] In some examples, the control unit is configured to instruct an action based on the estimated output variable value. In some examples, the control unit is configured to instruct the action when the output variable value crosses a predetermined threshold value. In some examples, the predetermined threshold value is based on an average output variable value estimated over a plurality of preceding heating periods.

[0068] In some examples, the device further comprises a user interface, wherein the control unit is configured to notify a user on the estimated output variable using the user interface. In some examples, the user interface comprises one or more of LEDs, haptic feedback elements, a display screen, audio feedback devices or a combination of these. In some examples, the action comprises the use of a user interface to notify the user of device status based on the estimated P / 90486.W001 output variable value. Such indicators may indicate one or more of pre-heating phase completion, correct or incorrect aerosol-forming article insertion in the device, puff count, aerosol-forming substrate depletion status, fully depleted aerosol-forming substrate.

[0069] In some examples, the action comprises termination of a heating period. In some examples, the heating period includes a pre-heating period and an inhalation session period, wherein the action comprises termination of the pre-heating period, or termination of the inhalation session period. In some examples, the heating period includes a pre-heating period and an inhalation session period, wherein the action comprises a change from the pre-heating period to the inhalation session period. In some examples, the pre-heating period has a duration of one of 5, 10 or 15 seconds.

[0070] In some examples, the action comprises a control of a power supplied to the dielectric heater assembly. In some examples, the control unit may control the power supplied to the dielectric heater assembly by changing one or more of the oscillation frequency, or the supply voltage of the oscillation circuit. In an example, the action may comprise transitioning from a preheating phase power to a steady-state phase power. In some examples, wherein the output variable relates to a temperature within the aerosol-forming substrate, the aerosol-generating device may be configured to control the power delivered to the aerosol-forming substrate based on the estimated output variable. The temperature generated via dielectric heating by the oscillation circuit, inside or at the substrate that is located in the heating zone of the load capacitor, for example after a pre-heating phase, can be in a range between 80°C to 365°C, more preferably between 80°C to 320°C, and more preferably between 150°C to 250°C, and more preferably between 180°C to 240°C, and more preferably between 200°C to 220°C. Preferably, the temperature of the substrate is kept lower than a temperature where pyrolysis could occur in the material composition of the aerosol forming substrate. Preferably, the temperature is such that one or more active ingredients of the substrate are vaporized and thereafter aerosolized, for example simultaneously or sequentially during a heating session, for example a heating session for a duration of one puff or inhalation, for example a duration in a range between 0.5 seconds to 10 seconds, or a heating session for a duration that covers multiple puffs or inhalation, for example a duration of more than 10 seconds and less than 10 minutes, or preferably, more than 30 seconds and less than 6 minutes. In some examples, a control unit of the aerosol-generating device or the settings of the oscillation circuit can be configured such that they control or set the temperature to such a range between 80°C to 365°C, more preferably between 80°C to 320°C, or more preferably 180°C to 250°C by open-loop control, or closed-loop control. P / 90486.W001

[0071] In some examples, the action comprises a selection of a heating profile for the aerosolforming substrate based on the estimated output variable value. In some examples, the control unit may determine based on the estimated output variable value that a particular type of aerosolforming substrate is present, and select a heating profile from a database based on the type of aerosol-forming substrate.

[0072] In some examples, the configuration and dimensioning of the oscillation circuit is selected to passively limit the temperature of the aerosol-forming substrate to a range between 80°C to 365°C, more preferably between 80°C to 320°C, or more preferably 180°C to 220°C, without the use of active temperature control.

[0073] In some examples, the device further comprises a puff detection system for detecting user puffs. The control unit may be configured to use user puff information from the puff detection system in conjunction with the measured one or more input parameters to determine the output variable value associated the dielectric heating operation. In some examples, the user puff detection may be used as a timing marker in order to associate a specific portion of the measured one or more input parameters with a specific event in the dielectric heating operation.

[0074] The disclosure further provides an aerosol-generating system comprising an aerosolgenerating device according to any of the above examples and an article comprising an aerosolforming substrate. In an example, the aerosol-forming substrate comprises an aerosol former. The article is arranged in the heating region of the device to enable dielectric heating of the aerosolforming substrate.

[0075] In some examples, the aerosol former comprises at least one of glycerol, polypropene glycol and vegetable glycerin.

[0076] The disclosure further provides a method for generating an aerosol from an aerosol-forming article using a dielectric heating aerosol-generating device according to any of the described examples. The method comprises measuring, using the one or more sensors, one or more input parameters associated with operating conditions of the aerosol-generating device during the heating period of the aerosol-generating device. The method further comprises estimating, using a model comprising data indicative of a relationship between the one or more input parameters and an output variable over a range of input parameter values, an output variable value based on the measured one or more measured input parameters.

[0077] It is to be appreciated that the features of the above examples of the disclosure are complementary with one another, except where stated otherwise, and so features of different P / 90486.W001 examples may be readily implemented in the device or method of other examples. Further optional examples of the disclosure are set out below.

[0078] Ex1 . An aerosol-generating device for dielectrically heating an aerosol-forming article to generate an aerosol, the aerosol-generating device comprising: a dielectric heater assembly having a heating region for receiving an aerosol-forming article; a radio frequency source for supplying power to the dielectric heater assembly to generate the aerosol over a heating period when the aerosol-forming article is received in the heating region; one or more sensors configured to measure one or more input parameters associated with operating conditions of the aerosol-generating device during the heating period of the aerosolgenerating device; and a control unit comprising a model comprising data indicative of a relationship between the one or more input parameters and an output variable over a range of input parameter values, the control unit being configured to: receive the one or more measured input parameters; and indirectly estimate, using the model, an output variable value based on the received one or more measured input parameters.

[0079] Ex 2. An aerosol-generating device according to Ex 1 , wherein the model comprises a database comprising a plurality of known output variable values associated with a range of input parameter values during a heating period of an aerosol-generating device.

[0080] Ex 3. An aerosol-generating device according to Ex 2, wherein the database comprises a one-to-one correspondence between a plurality of values for the one or more input parameters and a plurality of output variable values.

[0081] Ex 4. An aerosol-generating device according to Ex 1 , wherein the model comprises an artificial neural network trained using a plurality of known output variable values associated with a range of input parameter values for the one or more input parameters during the heating period of an aerosol-generating device.

[0082] Ex 5. An aerosol-generating device according to any preceding Ex, wherein the one or more input parameters are associated with a power consumption in the radio frequency source.

[0083] Ex 6. An aerosol-generating device according to any preceding Ex, wherein the one or more sensors comprises one or more of: P / 90486.W001

[0084] • a temperature sensor configured to measure a temperature at the heating region;

[0085] • a voltage or current sensor configured to measure a power consumption of the radio frequency source;

[0086] • a detection device configured to determine a type of aerosol-forming article;

[0087] • an airflow or pressure sensor configured to determine an airflow or pressure related to an airflow in an airflow channel of the aerosol-generating device;

[0088] • an electric field sensor configured to measure a value indicative of an electric field strength formed within the dielectric heater assembly;

[0089] • an inertial measurement unit configured to measure an orientation of the aerosolgenerating device;

[0090] • a frequency measurement unit configured to measure an oscillation frequency of the radio frequency source;

[0091] • a temperature sensor configured to measure an ambient temperature at, inside, or around the aerosol-generating device; and

[0092] • a humidity sensor configured to provide a humidity value at, inside, or around the aerosolgenerating device.

[0093] Ex 7. An aerosol-generating device according to any preceding Ex, wherein the one or more input parameters comprises one or more of:

[0094] • a temperature at or inside the heating region;

[0095] • a power consumption of the radio frequency source;

[0096] • a type of aerosol-forming article;

[0097] • an airflow or pressure related to an airflow in an airflow channel of the aerosol-generating device;

[0098] • an orientation / acceleration of the aerosol-generating device;

[0099] • a voltage, current or oscillation frequency in the radio frequency source;

[0100] • an ambient temperature at, inside, or around the aerosol-generating device; and

[0101] • a humidity at, inside, or around the aerosol-generating device.

[0102] Ex 8. An aerosol-generating device according to any preceding Ex, wherein the output variable comprises one or more of:

[0103] • a temperature inside the aerosol-forming article;

[0104] • a presence or absence of an aerosol-generating article in the heating region;

[0105] • a puff occurrence; P / 90486.W001

[0106] • a puff volume;

[0107] • a cumulative puff volume during an inhalation session period;

[0108] • a depletion level of aerosol-forming substrate in the aerosol-forming article;

[0109] • an identification of the aerosol-forming article;

[0110] • an orientation detection of the aerosol-forming article;

[0111] • a detection whether the aerosol-forming article has been previously used; and

[0112] • a temperature profile inside the aerosol-forming article.

[0113] Ex 9. An aerosol-generating device according to any preceding Ex, wherein the one or more input parameters comprises an oscillation frequency of the radio frequency source, and wherein the output variable is a temperature inside the aerosol-forming article.

[0114] Ex 10. An aerosol-generating device according to Ex 9, wherein the one or more input parameters further comprises a temperature in or adjacent to the heating region of the aerosolgenerating device.

[0115] Ex 11 . An aerosol-generating device according to any one of Ex 1 to 8, wherein the one or more input parameters comprises a value indicative of a strength of an alternating electric field between an electrode arrangement in the dielectric heater assembly, and / or a temperature in the heating region of the aerosol-generating device, and wherein the output variable is a temperature inside the aerosol-forming article.

[0116] Ex 12. An aerosol-generating device according to any one of Ex 1 to 8, wherein the one or more input parameters comprises a value indicative of a frequency of an alternating electric field in the heating region, and wherein the output variable is aerosol-forming article identity.

[0117] Ex 13. An aerosol-generating device according to any one of Ex 1 to 8, wherein the one or more input parameters comprises a value indicative of a frequency of an alternating electric field in the heating region, and wherein the output variable is aerosol-forming article substrate depletion level.

[0118] Ex 14. An aerosol-generating device according to any one of Ex 1 to 8, wherein the one or more input parameters comprises a temperature at a plurality of different locations along the P / 90486.W001 heating region, and wherein the output variable is an aerosol-forming substrate temperature distribution profile.

[0119] Ex 15. An aerosol-generating device according to any one of Ex 1 to 8, wherein the one or more input parameters comprises a value indicative of a strength of an alternating electric field at a plurality of locations within an electrode arrangement of the dielectric heater assembly, and wherein the output variable is an aerosol-forming substrate temperature distribution profile.

[0120] Ex 16. An aerosol-generating device according to any one of Ex 14 to 15, wherein the control unit is further configured to determine, based on the temperature distribution profile a location in the aerosol-forming article which is at a target temperature.

[0121] Ex 17. An aerosol-generating device according to any one of Ex 1 to 8, wherein the one or more input parameters comprises a temperature of a switching unit in the radio frequency source and a temperature in the heating region of the aerosol-generating device, and wherein the output variable is a temperature inside the aerosol-forming article.

[0122] Ex 18. An aerosol-generating device according to any one of Ex 9 to 17, wherein the one or more input parameters further comprises a power supplied to the dielectric heater assembly.

[0123] Ex 19. An aerosol-generating device according to any preceding Ex, wherein the control unit is configured to measure the one or more input parameters over a period of time to determine one or more input parameter profiles for the dielectric heating operation.

[0124] Ex 20. An aerosol-generating device according to Ex 19, wherein the control unit is configured to determine one or more properties associated with the one or more input parameter profiles, and compare the one or more properties with one or more reference properties.

[0125] Ex 21 . An aerosol-generating device according to Ex 19, wherein the control unit is configured to determine one or more properties associated with a first order derivative or a second order derivative of the one or more input parameter profiles. P / 90486.W001

[0126] Ex 22. An aerosol-generating device according to any preceding Ex, wherein the control unit comprises a counter, wherein a counting value from the counter during the heating period is used as an additional input for the control unit to estimate the output variable value.

[0127] Ex 23. An aerosol-generating device according to any preceding Ex, wherein the control unit is configured to instruct an action based on the estimated output variable value.

[0128] Ex 24. An aerosol-generating device according to Ex 23, wherein the control unit is configured to instruct the action when the output variable value crosses a threshold value.

[0129] Ex 25. An aerosol-generating device according to any one of Ex 23 to 24, wherein the action comprises adjusting a power supplied to the dielectric heater assembly.

[0130] Ex 26. An aerosol-generating device according to any one of Ex 25, wherein the action comprises termination of a heating period.

[0131] Ex 27. An aerosol-generating device according to any one of Ex 23 to 24, wherein the action comprises using a user interface to indicate a status of the aerosol-generating device to a user.

[0132] Ex 28. An aerosol-generating device according to any one of Ex 23 to 24, wherein the heating period includes a pre-heating period and an inhalation session period, wherein the action comprises a change from the pre-heating period to the inhalation session period.

[0133] Ex 29. An aerosol-generating device according to any one of Ex 23 to 24, wherein the heating period includes a pre-heating period and an inhalation session period, wherein the action comprises selecting a heating profile for the aerosol-forming article for the inhalation session period.

[0134] Ex 30. An aerosol-generating system comprising: an aerosol-generating device according to any preceding Ex; and an aerosol-forming article received in the heating region of the aerosol-generating device. P / 90486.WQ01

[0135] 31 . A method for generating an aerosol from an aerosol-forming article using a dielectric heating aerosol-generating device, the aerosol-generating device comprising: a dielectric heater assembly having a heating region for receiving an aerosol-forming article; an radio frequency source for supplying power to the dielectric heater assembly to generate the aerosol over a heating period when the aerosol-forming article is received in the heating region; the method comprising: measuring, using one or more sensors, one or more input parameters associated with operating conditions of the aerosol-generating device during the heating period of the aerosolgenerating device; and indirectly estimating, using a model comprising data indicative of a relationship between the one or more input parameters and an output variable over a range of input parameter values, an output variable value based on the measured one or more measured input parameters.

[0136] Any of all the features of the above described oscillation circuit may be provided on a single low dielectric constant carrier material.

[0137] Brief Description of Drawings

[0138] The invention will be further described, by way of example only, with reference to the accompanying drawings in which:

[0139] Figure 1 is a schematic illustration of a dielectric heating aerosol-generating system according to embodiments of the disclosure;

[0140] Figure 2 is a schematic illustration of an oscillation circuit for use in the dielectric heating aerosol-generating system of Figure 1 , according to embodiments of the disclosure;

[0141] Figure 3a-b are a schematic illustrations of an oscillation circuit showing two different phase-shifting elements, one exemplarily implemented as a resonance circuit, one exemplarily implemented as a capacitive element, to achieve a 180 degrees phase shift;

[0142] Figure 3c is a schematic illustration of an oscillation circuit showing two different phaseshifting elements, one exemplarily implemented as a resonant heating regionhaving parallel resonance properties, one exemplarily implemented as a capacitive element, to achieve a 180 degrees phase shift;

[0143] Figure 4 illustrates an oscillation circuit diagram according to embodiments of the disclosure;

[0144] Figure 5A-F illustrate how a Quartz-mimicking or Quartz equivalent circuit may be derived, as a non-limiting example of a parallel-resonant circuit, according to embodiments of the disclosure; P / 90486.W001

[0145] Figure 6 illustrates a frequency analyzer plot of a parallel resonant circuit showing the effect of the switching frequency on the phase shift and impedance of a parallel resonant circuit;

[0146] Figure 7 is a schematic illustration of a control system for the dielectric heating aerosolgenerating system shown in figure 1 , according to embodiments of the disclosure;

[0147] Figure 8 is a schematic illustration of a control system for the dielectric heating aerosolgenerating system, according to embodiments of the disclosure;

[0148] Figure 9 is a flow diagram illustrating a method for parameter estimation in a dielectric heating aerosol-generating device, according to embodiments of the disclosure;

[0149] Figure 10 is a schematic illustration of the processing steps carried out by a control unit to estimate an output variable value, according to embodiments of the disclosure;

[0150] Figure 11 shows a method for generating a software that can be uploaded as a firmware to characterize puff volume;

[0151] Figure 12a illustrates a method for producing a machine learning model for an aerosolgenerating device for generating an aerosol from an aerosol-forming article;

[0152] Figure 12b shows a method for characterizing a puff volume using a machine learning model on an aerosol-generating device;

[0153] Figure 13 illustrates a method of characterizing a usage session using a machine learning model;

[0154] Figure 14 shows a system and computing apparatus that may be used to implement methods according to an example embodiment;

[0155] Figure 15 illustrates a data collection scheme for training data, according to embodiments of the disclosure;

[0156] Figure 16 shows a system for creating training data using human puffs, according to embodiments of the disclosure;

[0157] Figure 17 illustrates a change in power during a puff, according to embodiments of the disclosure;

[0158] Figure 18 is a graph illustrating the relationship between dielectric heating power and the oscillation frequency profile for a dielectric heating operation, according to embodiments of the disclosure;

[0159] Figure 19 is a graph illustrating the temperature of an aerosol-forming article measured as a function of time and heating power in an aerosol-generating system during a heating operation, according to embodiments of the disclosure;

[0160] Figures 20a, 20b and 20c are schematic illustrations of frequency responses of an electrically resonant oscillation system according to an embodiment of the disclosure; P / 90486.W001

[0161] Figure 21 is a graph illustrating how the frequency dependency of the relative permittivity of Glycerol varies with temperature.

[0162] Figure 22a is a graph illustrating an oscillation frequency profile measured during a dielectric heating operation, according to an embodiment of the disclosure;

[0163] Figure 22b is a graph illustrating a first derivative of the graph of figure 16a;

[0164] Figure 23 is a flow diagram illustrating a control process for a pre-heating phase of a dielectric heating operation, according to an embodiment of the disclosure; and

[0165] Figure 24 is a flow diagram illustrating a control process for a steady-state phase of a dielectric heating operation, according to an embodiment of the disclosure.

[0166] Specific Description

[0167] Figure 1 is a schematic illustration of a dielectric heating aerosol-generating system 100 according to an embodiment of the disclosure. The system 100 comprises an article 105 comprising an aerosol-forming substrate 110 and an aerosol-generating device 120 for heating the aerosol-forming substrate 110. The aerosol-generating device 120 comprises a first electrode 130 and a second electrode 135 separated by a region 140 for receiving the article 105. The region 140 and the article 105 are sized such that the aerosol-forming substrate 110 is in contact or in close proximity to both the first electrode 130 and the second electrode 135 when received within the region 140. Moreover, the first electrode 130 and the second electrode 135 form part of a feedback loop of an oscillation circuit 150 via a first and second electrical contact 160, 165.

[0168] In other examples, the first electrode 130 and the second electrode 135 may form part of the article 105 comprising the aerosol-forming substrate 110. In such embodiments, a heating regionbetween the first and second electrical contacts 160, 165 is sized such that, when the aerosol-forming article 105 is housed within the region 140, an electrical connection is made between the first electrode 130 and the first electrical contact 160, and the second electrode 135 and the second electrical contact 165.

[0169] In some embodiments, the width of the article 105 comprising the aerosol-forming substrate 110 is slightly greater than the spacing between the first electrode 130 and the second electrode 135, such that the distal end of the aerosol-generating substrate 110 is slightly compressed between the first electrode 130 and the second electrode 135. In some embodiments, the article 105, in an initial, uncompressed form has a width between 5-30% larger than the distance between the first electrode 130 and the second electrode 135. This may reduce or prevent the build-up of air between the first electrode 130 and the second electrode 135 when the aerosol-forming article 105 is received in the region 140, and decrease a distance between first and second electrodes 130, 135 for dielectric heating, thereby improving dielectric properties of the load capacitor CL and the P / 90486.W001 accuracy of any measurements or determinations of the dielectric properties of the aerosol-forming substrate 110 performed by the aerosol-generating device 120.

[0170] The aerosol-forming substrate 110 may comprise tobacco-based or non-tobacco based materials having an aerosol forming material therein and one or more active agents or ingredients, such as nicotine, pharmaceutical, botanicals, flavorants, liquid substrates with one or more active agents or ingredients, or a combination thereof. The aerosol-forming substrate 110 can also be a liquid aerosol-forming substrate and thereby the aerosol-forming article 105 can be in the form of a cartridge, capsule, or liquid container, and the electrodes 130, 135 can be configured as a wicking element or capillary element for liquid transfer. For example, it is possible that first and second electrodes 130, 135 form a capillary structure that is part of the aerosol-forming article 105 or reaches into an inner volume of the aerosol-forming article 105 that can heat and vaporize a liquid aerosol-forming substrate 110 located in the inner volume. For example, the first and second electrodes 130, 135 can be embodied as parallelly arranged plates separated by a distance that forms a capillary channel, for example in a range between 0.1 mm to 2mm, depending on the desired capillary strength or rise. The first and second electrodes can be arranged as two matrices or arrays of pin-like, rod-like, or tab-like electrodes with opposite polarity, the two matrices or arrays interposed between each other, forming a capillary structure therebetween, for example with an average distance between neighboring pin-like electrodes being in a range between 0.1 mm to 2mm, depending on the desired capillary strength or rise. In another variant, the wicking element can be a separate element that is interposed between two electrodes 130, 135, for example flat or slightly curved electrodes 130, 135.

[0171] The aerosol-generating device 120 further comprises a power supply 170 and a controller 180 electrically coupled to the oscillation circuit 150. The controller 180 is communicatively coupled to a frequency measurement unit 185 configured to measure frequency data indicative of an oscillation frequency of the alternating electric field across the first and second electrodes 130, 135 during a dielectric heating operation. The controller 180 is configured to receive the measured frequency data from the frequency measurement unit 185 and determine one or more parameters related to the dielectric heating operation based on the received frequency data. Controller 180 is also communicatively coupled two a first and second temperature sensor 187, 188, configured to measure a temperature at a heating region 140 and at a switching unit in the oscillation circuit respectively. The controller 180 may be configured to use the temperature data received from the first and second temperature sensors 187, 188, to determine a dielectric heating loss in the aerosol-generating system and estimate a temperature of the aerosol-forming substrate 110. P / 90486.W001

[0172] In some embodiments, the controller 180 may be configured to estimate various parameters associated with a dielectric heating operation before the aerosol-forming article 105 has been fully heated to an aerosolization temperature during an initial pre-heating phase. For example, during a pre-heating phase of an aerosol-forming article, an RF electric field can be applied to the electrode assembly 130, 135 to start heating up the aerosol-forming substrate 110 in the aerosol-forming article 105. During this stage, as the aerosol-forming substrate 110 is gradually heated from an environmental temperature (e.g. 20°C) to reach an aerosolization temperature (e.g. 200°C), which allows to estimate or detect a variety of parameters of the aerosol-forming article.

[0173] For example, it could be detected whether the aerosol-forming substrate is inserted or not. All whilst performing the regular heat up-phase. Upon insertion of an aerosol-forming article, the controller 180 may be configured to initiate the pre-heating phase for a given time duration (e.g. 5, 10, 15 seconds) that is short enough that the operational temperature is not yet reached (e.g. below 200°C), or until a given temperature is reached (e.g. a temperature above boiling temperature of water, e.g.120°C), for example one that is below the operational temperature (e.g. measured by one or more temperature sensors). During the pre-heating phase, the controller 180 may be configured to measure and store data one or more input parameters (for example power consumption of the oscillation circuit 150 or the temperature at the heating region 140). The controller 180 may then be configured to analyse the measured input parameter data by use of the machine learning / mapping model to estimate one or more output variable values.

[0174] In this embodiment, the power supply 170 can be a rechargeable lithium ion battery, for example with one or more lithium ion battery cells, and the aerosol-generating device 120 comprises a power connector that enables the aerosol-generating device 120 to be connected to a mains power supply for recharging the power supply. Providing the aerosol-generating device 120 with a power supply, such as a battery, enables the aerosol-generating device 120 to be portable and used outdoors or in locations in which a mains power supply is not available.

[0175] In use, power is supplied to the oscillation circuit 150 from the power supply 170 when a user activates the aerosol-generating device 120. In this embodiment, the aerosol-generating device 120 is activated by a user pressing an activation button (not shown) that can be provided on an external surface of the aerosol-generating device 120. It will be appreciated that in other embodiments, the aerosol-generating device 120 may be activated in another manner, such as on detection of a user drawing on a mouthpiece (not shown) by a puff sensor provided on the mouthpiece, or a user holding the aerosol-generating device 120. When power is supplied to the oscillation circuit 150, the oscillation circuit 150 generates an alternating electric field across the P / 90486.W001 first and second electrodes 130, 135 to dielectrically heat the aerosol-forming substrate 110 in the region 140, releasing volatile compounds.

[0176] The aerosol-generating system 100 is also configured for measuring a dielectric property of the aerosol-forming article 105 or the aerosol-forming substrate 110 using the electrodes 130, 135 that are employed for dielectric heating of the aerosol-forming substrate 110. In some examples, the first and second electrodes 130, 135 can be used for dielectric measurements, either during the heating process or separately to the heating process. In this embodiment, the aerosolgenerating system 100 can be configured to determine the presence of the aerosol-forming article 105 between the first electrode 130 and the second electrode 135. The aerosol-generating system can be configured to measure a dielectric property, for example an instant value, timely-evolution, or change of a dielectric property, for example to determine whether the aerosol-forming article 105 meets specific criteria or is an authentic substrate. The aerosol-generating system in this example is also configured to control the heating of the aerosol-forming substrate 110 based on the measured dielectric property of the aerosol-forming article 105.

[0177] Figure 2 is a schematic illustration of an oscillation circuit 250 for use in the aerosolgenerating system 100 of Figure 1 , according to an embodiment of the disclosure. Oscillation circuit 250 comprises a switching unit 260 interconnected with a resonator feedback loop 270 to provide for a self-oscillating signal to the switching unit 260. The switching unit 260 comprises a single transistor, such as a bipolar junction transistor (BJT) or a field effect transistor (FET).

[0178] The oscillation circuit 250 can further comprise a choke 280 that acts on an input to the feedback loop 270 to provide for a stimulation signal, for example a stimulation voltage. The oscillation circuit also comprise a biasing unit 290 acting on the feedback loop 270 for providing a variable or controllable biasing signal, for example a biasing voltage for setting the operating conditions. In the variant shown, the feedback signal can be described as a voltage. The output voltage UOUT of the switching unit 260 is coupled to the feedback loop 270 providing a feedback switching signal in the form of a voltage UIN to the switching unit 260. The configuration of the feedback loop 270 is such that the output signal, e.g. the voltage UOUT of the switching unit 260 can undergo a phase change and arrives inverted at the input UIN of the switching unit 260 for resonant oscillation. In other configurations, a current could be used as the feedback signal with a switching unit 260 comprising a BJT.

[0179] The feedback loop 270 is configured to be self-oscillating and will oscillate at or close to a given resonance frequency determined by the values of the passive components of the feedback loop 270. Feedback loop 270 is configured to provide a 180° phase shift from the output UOUT to P / 90486.W001 input UIN of switching unit 260 for oscillation, and in addition, the transistor T is configured for inverting operation.

[0180] As shown in Figures 3a and 3b, feedback loop 270 includes a resonant circuit 272 comprising the load capacitor CL providing for a first 90 degrees phase shift or quarter wave shift to the feedback signal. Feedback loop 270 further includes a capacitive element 274 providing for a second 90 degrees phase shift or quarter wave shift to the feedback signal, such that the feedback signal reaching the input of the switching unit 260 is inverted and phase-shifted by 180 degrees. Switching unit 260 is itself configured for inverted switching operation to provide a 180 degree phase shift between the input UIN and the output UOUT of the switching unit 260.

[0181] Resonant circuit 272 comprises the first and second electrodes 130, 135, together forming a load capacitor CL. When an aerosol-forming substrate 110 is situated between the first and second electrodes 130, 135, it forms part of the load capacitor CL. Importantly, the load capacitor CL is formed in the feedback loop 270, and not at a separate output or part of a separate circuitry that is connected to the switching unit 260. This enables a high-frequency oscillating voltage to be created across the electrodes of load capacitor CL , which is needed for sufficient and efficient dielectric heating of the aerosol-forming substrate 110, without having an additional output or circuit to the already resonating feedback loop 270, which would create unnecessary losses and circuit complexity. The resonant circuit 272 may comprise a series resonator circuit or a parallel resonator circuit, examples of which are described in greater detail below.

[0182] In an alternative embodiment, resonant circuit 272 may include a resonant heating region272, as illustrated in Figure 3c, a transmission line, or a combination of a resonant heating regionand a transmission line. The resonant circuit 272 including resonant cavity, transmission line, or a combination of a resonant heating regionand a transmission line may have an interior volume configured to receive an aerosol-forming substrate 110, for example having an opening for inserting an aerosol-forming substrate 110. In an example, the resonant circuit 272 including resonant cavity, transmission line, or a combination of a resonant heating regionand a transmission line can be configured as a A / 4 resonator. The resonant circuit 272 including resonant cavity, transmission line, or a combination of a resonant heating regionand a transmission line may be configured to behave like an RLC circuit. The inductor L and the capacitor C are arranged in parallel to each other, to have a parallel resonance or a frequency close to the parallel resonance that can be stimulated by the switching unit 260. Resonant circuit 272 including the resonant cavity, transmission line, or a combination of a resonant heating regionand a transmission line may be coupled to the feedback loop 270 using one or more of a capacitive coupling, an inductive P / 90486.W001 antenna coupling (magnetic coupling), a direct electric coupling, or a window coupling (e.g. coupling with a loop).

[0183] The resonant circuit 272 including resonant cavity, transmission line, or a combination of a resonant heating regionand a transmission line can have any shape, but preferably has a cylindrical shape or a rectangular parallelepiped shape. In one embodiment, the resonant heating regioncan be configured as a split-ring resonator.

[0184] The resonant circuit 272 including resonant cavity, transmission line, or combination of resonant heating regionand transmission line, can also be supplied with an electromagnetic field from an electromagnetic field generator that is configured to generate radio-frequency (RF) electromagnetic radiation having a predetermined frequency composition, for example as shown in International Patent Publication Nos. WO2022 / 128290, WO2020 / 244996, and WO2021 / 013477, these references herewith incorporated by reference in their entirety.

[0185] Figure 4 illustrates an oscillation circuit 350 according to a non-limiting, exemplary embodiment of the disclosure. Oscillation circuit 350 comprises a switching unit 260 in the form of a transistor T having an intrinsic capacitance Ci. Moreover, transistor T is configured for inverting operation, for example as an inverting common source FET, MOSFET, or a common emitter BJT. The source terminal of transistor T can be coupled to a DC power supply via a choke 280. Between the gate and source terminals of transistor T extends a feedback loop 270. The feedback loop 270 comprises a resonant circuit 272 including a load capacitor CL having a first and second electrode 130, 135 separated by an aerosol-forming substrate 110. In the variant shown, the resonator circuit 272 is also connected to ground via a delay line D and a capacitor C2 connected in series to the delay line D . The circuit 350 further comprises a biasing unit 290 coupled to the gate terminal of the transistor T via the delay line DL. AS shown in Figure 4, the biasing unit 290 is electrically connected between the delay line DL and the capacitor C2, so that the biasing unit 290 is somewhat isolated from the high oscillation frequency of the feedback loop 270.

[0186] The delay line D is a time delay element, for example an element that has inductive behavior, for slowing down the arriving voltage wave from the feedback loop 270 during a period of the oscillation. This allows to tune the resonant circuit 272 to a desired switching and oscillation frequency, to move the oscillation frequency away from the natural resonant frequency given by the resonant circuit 272. This ensures that oscillation circuit 350 remains in a predefined frequency operating range to provide for the requisite inverted or 90° phase shifted feedback and also to make sure that the feedback loop has a low impedance to provide for a high gain, as described in greater detail below. P / 90486.W001

[0187] The oscillation circuit 350 is shown with electrical contacts 160, 165 that are arranged on each side of the load capacitor CL. In some embodiments, the first and second electrodes 130, 135 may be removable from the oscillation circuit 350, or may form part of the aerosol-forming article 110. In such embodiments, electrical contacts 160, 165 provide an electrical connection between the first and second electrode 130, 135 and the feedback loop 270. In embodiments where the load capacitor CL is fixed within the feedback loop 270, for example, such that an aerosol-forming substrate 110 can be inserted and removed to and from a region 140 formed in between the first and second electrodes 130, 135, electrical contacts 160, 165 provide electrical connections from the first and second electrodes to the next components in the feedback loop 270, e.g. inductors Li and L2.

[0188] With respect to the power supply voltage, a DC power supply voltage is provided, that is preferably in a range that is suitable for battery operation with one or more standard battery cells. Preferably, the DC power supply voltage is below 14V. For example, it is possible to operate the oscillation circuit 350 on a single battery cell, for example an 18650 battery cell (Li-Ion), or a similar battery cell, that provides for 3.2V to 3.9V. However, more preferably, a voltage of one battery cell of an exemplary 3.5V to 7V for power supply can be boosted, for example by a DC-DC converter (e.g. a boost circuit), or a voltage doubler. Alternatively or in addition, two or more battery cells can be used in series, or other configurations or arrangements that allows to increase a voltage from one or more battery cell can be used. It is also possible to have a controllable output voltage (e.g. DC-DC converter, voltage regulator), to control the temperature of heating by a change to the DC supply voltage, or to boost the voltage (for example to 10-12V) for maximum power at the preheating stage, to speed up the preheating stage with the goal to reach the aerosolization temperature quickly. Control of the DC supply voltage is one way that makes it possible to rapidly change heating power despite the oscillation circuit 350 freely oscillating.

[0189] A capacitor Ci is arranged in parallel to the transistor T and therefore in parallel with the intrinsic capacitor of the transistor T (e.g. a field effect transistor). This facilitates a less voltagedependent oscillation and frequency, stabilizes the oscillation, and also improves the overall dielectric heating efficiency. Capacitance of capacitor Ci is chosen to be larger than the maximal intrinsic capacitor Ci of transistor T at the operating conditions, so that the variation of the intrinsic transistor based on frequency, temperature, etc. has much less or negligible influence on the feedback loop 270. For example, in a non-limiting embodiment, the value can be in a range between 2pF to 10OpF, more preferably in a range between 5pF and 50pF.

[0190] Capacitive element 274 comprises a capacitor C2arranged at the output or end of the resonant circuit 272. In one embodiment, capacitive element 274 comprises more than one P / 90486.WQ01 capacitor. As described above, capacitive element 274 has the function of providing a 90° phase shift to the feedback voltage of feedback loop 270 with minimized losses or other undesired effects, and it therefore needs to have a high-quality factor or Q factor, preferably above 1000 at 100MHz. The capacitance value for C2 of the capacitive element 274 should be relatively high as compared to Ci, for example in a range between 500pF to 100nF, more preferably between 1 nF and 50nF, which leads to a low impedance of capacitive element 274. In one embodiment, the capacitive element 274 can be implemented as a RC network to provide for the 90° phase shift, for example using two single-resistor-capacitor networks, having two capacitors in the feedback loop, each capacitor connected to ground via a resistor.

[0191] Resonant circuit 272, together with capacitive element 274, provides for a 180° phase shift and a voltage gain from the output UOUT to the input UIN, and transistor T (for example a FET) is configured for inverting operation, thereby also providing for another 180° phase shift. This results in a resonant or close-to resonant oscillation and an amplified voltage UL across the electrodes 130 and 135 of the load capacitor CL, as compared to the DC supply voltage. When operating close to resonance, the resonant circuit 272 circuit behaves inductively, having a high Q factor. Furthermore, the feedback loop 270 is impedance-matched with the transistor T, to provide for a high gain, leading to an increased voltage across the load capacitor CL. Also, preferably, this gain is achieved without the use of an additional voltage or current amplifying passive element, such as a tapped inductor or a transformer located in the circuit that forms the feedback loop 270, as such passive elements are difficult and lossy to operate and design at frequencies greater than 50 MHz.

[0192] The combination of capacitor Ci, the feedback loop with resonant circuit 272 and capacitive element 274 can also be described as a bandpass filter or Pi or TT network that generates a 180° phase shift. In the illustrated embodiment, the resonant circuit 272 of the feedback loop 270 is not connected to ground, but is suspended with ends at each capacitor Ci and C2, thereby not having a direct ground connection at either end of resonant circuit 272, reducing stray elements and ground influences for more predictable operation.

[0193] At the operating frequency, the resonant circuit 272 including load capacitor CL acts as an inductive load providing a first 90° phase shift, also referred to as a quarter-wave phase shift, and capacitive element 274 exemplarily including a high quality factor capacitor C2 connected to ground, provides for the second 90° phase shift or quarter-wave phase shift.

[0194] Oscillation circuit 350 can be described or characterized as a Pierce oscillation circuit with a modified feedback loop 270, where the physical Quartz element is replaced by a Quartz-mimicking or Quartz electric equivalent circuit to provide for an inverting feedback to the switching unit 260 that also operates in an inverted mode. In some embodiments, oscillation circuit 272 can be based P / 90486.WQ01 on other resonant feedback loop oscillation circuit configurations, for example, but not limited to, the use of a Colpitts or Hartley type oscillator, using an inverting transistor T.

[0195] Figures 5A to 5F illustrate how a Quartz-mimicking or Quartz equivalent circuit may be derived using a variant of a parallel resonator circuit PRC, that can serve as an exemplary and non-limiting embodiment as resonant circuit 272. A Quartz-mimicking or Quartz equivalent circuit can have a parallel resonance at a given frequency. This can be seen as a circuit with two branches, one representing the mechanical oscillation and one representing the electric behavior, as illustrated in figure 5A and 5B. The mechanical oscillation is represented by a first branch having load capacitor CL and an inductor LTOT. The electric oscillation is represented by a second branch arranged in parallel to the first branch, having a Capacitor CE. This configuration leads to a series connection of two capacitors CL and CE (seen around the loop formed by the two branches) so that CE will decrease the overall capacitive value of the equivalent circuit. Also, this circuit provides an inductive phase shift of about 90° within certain defined frequency range.

[0196] Referring now to Figure 5C, one branch of the parallel resonator circuit PRC comprises an inductor LTOT and the load capacitor CL connected in series of the first branch. This branch can be improved by splitting LTOT into two inductors Li and L2on each side of the load capacitor CL, as shown in Figure 5D, to provide a split inductor or split-coil design and a more symmetric application of the voltage UL across the load capacitor CL, thereby improving dielectric heating efficiency. To provide for parallel resonance, the capacitor CE of the second branch can be replaced with an inductor (as shown in Figure 5F) due to capacitor CE‘S minimal capacitive effect on CL.

[0197] In a non-limiting example starting from the split inductor Li and L2of the resonant circuit of Figure 5D, inductors Li and L2can be mutually magnetically coupled to form a mutual inductance M, thereby forming the parallel circuit branch or second branch of the resonant circuit 272, as shown in Figure 5E. The mutual magnetic coupling can be achieved by the close proximity of the two inductors L1 and L2 with alignment of winding axis of the coils, or by use of a mutual magnetic core, or both. This has the advantage of providing a parallel-resonator circuit PRC without the use of additional wires for the second branch, and without additional windings or separate magnetic cores for a second parallelly-arranged inductor. This also allows for a symmetric arrangement that favors and facilitates the inductive coupling of the two inductors Li, L2and the balancing of the voltage UL over load capacitor CL. The symmetry of the two branches in either direction with first branch Li - CL - L2and the second branch with inductor LE, representing the two mutual inductance values, facilitates the symmetrical balancing of the voltage over the electrodes of the load capacitor CL, which consequently reduces losses created at the load capacitor CL. This split inductor principle can also be referred to as a split coil resonator. P / 90486.WG01

[0198] The resonant circuit 272 could also be implemented as shown in the Figure 5F, where the mutual inductance M (seen two times due to the mutuality) is replaced by a separate inductive element, for example inductor LE.

[0199] The values of components in this resonant circuit 272 are preferably chosen to be in the following exemplary and non-limiting ranges. LTOT can be a range between 10nH to 50nH, more preferably between 15nH and 40nH, which is the equivalent of Li plus L2, LEcould be in a range between 7nH and 30nH, more preferably between 10nH and 20nH, and the value of the load capacitor can be in a range between 0.5pF to 5pF, more preferably between 1 pF to 3pF.

[0200] The resonant circuit 272 can be configured as another type of tank circuit providing for the 90° phase shift in a given frequency range. In one embodiment, the resonant circuit 272 can be implemented as a series resonant circuit, having the load capacitor CL connected in series with one or more inductive elements, configured to provide for an inductive response or 90 degrees phase shift in a given frequency range that is suitable for dielectric heating.

[0201] In order to produce the required dielectric losses in the load capacitor CL, a high frequency is needed. The dielectric losses will increase somewhat proportionally to the frequency (in this case the oscillation frequency fsof the oscillation circuit) of the voltage applied to the load capacitor CL. However, higher switching frequencies fslead to greater switching losses in the transistor. Where very high switching frequencies are used, e.g. 1 GHz or more, expensive circuit designs are necessary to make the circuit operable (e.g. GaAn transistors, GHz-type circuit design e.g.).

[0202] To compensate for a lower switching frequency, a higher oscillation AC voltage UL (for example measured by the RMS or peak voltage) can be provided across the load capacitor CL to deliver the necessary power to the load capacitor CL. When a balance has to be struck between increasing the switching frequency fsand increasing the voltage over the load capacitor CL in order to increase the heating power (and therefore dielectric losses PL in the load capacitor CL which can include the aerosol-forming substrate 110), the voltage increase has a stronger effect on the heating power than frequency increase, as voltage has a second order relationship with power. The dielectric losses in load capacitor CL are also dependent on the distance separating the two electrodes 130 and 135, which influences the electric field strength EF across the aerosol-forming substrate 110.

[0203] For mobile portable designs of dielectric heater that are handled by a human, safety considerations with respect to voltage insulation and potential dielectric breakdown due to contamination and general inhomogeneous material of the substrate material S of the aerosolforming substrate 110 are a serious concern. Therefore, there are working ranges for switching frequency fs and electric field strength EF, for a hand-held and human-operated portable device P / 90486.W001

[0204] The electric field strength is dependent on voltage UL applied to capacitor CL and a maximal distance d between the first and second electrodes 130, 135.

[0205] It is preferable that the switching frequency is limited to a certain range, firstly to ensure sufficient power losses PL across the load capacitor CL, and secondly to avoid excessive switching losses. Limiting the value of fsenables the use of straightforward circuit design, for example but not limited to the use of a LDMOS or other standard transistor used for RF circuits. Preferably, the switching frequency should be in a range between 100MHz and 1 .2 GHz, more preferably between 150MHZ to 1GHz, more preferably between 200MHz and 900MHz. At the same time, the maximal-value of electric field strength between the electrodes of the load capacitor CL should be limited to a reasonable value that allows for simple electric insulation materials and designs, taking into account impurities and inhomogeneous substrate designs, the use of thin electrodes that may be located in close proximity to each other (for example but not limited to 1 mm-10mm), potential exposure or close proximity to human body, and potential improper human manipulation. Preferably the average electric field strength across the electrodes of the load capacitor is maximally 120V / mm, more preferably maximally 100V / mm, even more preferably maximally 80V / mm.

[0206] Preferably, during nominal heating operation (not during a start-up or warm-up phase that lasts less than 30 seconds where the efficiency can be lower), the heating efficiency should be at least 60% (desired power losses PL versus all the other losses, which can include switching losses of transistor T, losses caused by biasing circuit and choke, inductive-resistive losses of the inductors Li and L2, capacitive-resistive losses from capacitors Ci and C2, and resistive losses of electrodes Ei and E2and the wiring). As an example, the overall power could be 10W, while the effective power or heating losses PL should be 6W or more, with 4W or less of non-heating losses or other heating losses not caused in the substrate 110 between the first and second electrodes 130, 135. More than 6-7W of power losses is not desirable for a handheld device. Preferably the aerosol-generating device is configured such that there are less than 5W of losses during nominal heating operation or any other operation lasting for more than 30 seconds. The average dielectric heating power density provided by the aerosol-generating device may be in a range between 1 W / cm3 to 25W / cm3 per volume of aerosol-forming substrate material over a time period of less than 15 minutes, preferably between 1.5W / cm3 and 15W / cm3.

[0207] During an initial start-up phase, the average dielectric heating power density may be controlled or set to operate in a range between 7W / cm3 to 25W / cm3. During a nominal (maintenance) heating phase, the average dielectric heating power density may be controlled or set to operate in a range between 1 W / cm3 to 7W / cm3. Preferably, during the start-up phase, the P / 90486.WG01 average dielectric heating power density is controlled or set to operate in a range between 8W / cm3 to 20W / cm3. Preferably, during the nominal heating phase the average dielectric heating power density is controlled or set to operate in a range between 1 W / cm3 to 5W / cm3.

[0208] To have a proper inverting effect and a 180° phase shift on the feedback loop 270 between UIN and UOUT, the oscillation circuit 350 must remain in a frequency operating range where the behavior of the feedback loop 270 is highly inductive. In the example comprising a parallel resonator circuit (PRC), the series resonance frequency fsER (resonant frequency) is relatively close to the parallel resonance frequency fpAR (antiresonant frequency). If the oscillation frequency fs of the PRC exceeds the parallel resonance frequency fPAR, the feedback loop 270 will act capacitively and not provide the necessary phase inversion to the feedback loop 270. Furthermore, the equivalent impedance of the circuit will increase to an extent that is too high for efficient dielectric heating as the oscillation circuit 350 will not be able to provide a high signal gain.

[0209] Oscillations in the feedback loop 270 will be naturally drawn towards the parallel resonance (antiresonant) frequency of the resonant circuit 272. However, the addition of a delay line D can introduce a slight time delay limiting the oscillation frequency below the parallel resonance frequency. Figure 6 shows a frequency analyzer plot of an exemplary resonant circuit 272, specifically a plot of a parallel resonator circuit PRC showing the relationship between the oscillation frequency (with a series resonance at 855 MHz and a parallel resonance at 1 .246 GHz), the phase shift across the PRC (with a relatively flat inductive 90° degrees frequency response between the two resonant frequencies) and the effective impedance of the PRC. More specifically, it can be seen from Figure 6 that the 90° phase shift starts dropping before the parallel resonance frequency fPAR is reached. After the parallel resonance frequency f PAR, the phase shift response drops below 0° to capacitive behavior and the impedance is very high, e.g. 2.4kQ. The ideal operating frequency range is closer to the series resonance frequency fsER where the phase shift is still 90° and the impedance response is low. The impedance of the resonant circuits at the operating frequency of the oscillation circuit may be between 0.5Q and 10O, preferably between 10 and 50, and more preferably, between 1 .50 and 30. The parallel resonance frequency fPAR can be above 1 GHz, e.g. 1 GHz to 1 .5GHz, while the actual switching frequency fs can be below 1 GHz, and this lower switching frequency is caused by the delay line DL.

[0210] Ideally, the oscillation frequency fs should be set to be below the parallel resonance frequency fPAR but above the series resonance frequency fsER, to make sure that two conditions are fulfilled, firstly (i) that the resonant circuit behaves inductively to provide a 90° phase shift, and secondly (ii) to make sure that the impedance of the resonant circuit (an therefore the feedback loop 270) is low, as illustrated in the graphs of FIG. 6. For example, a resulting impedance of the P / 90486.WG01 feedback loop at the oscillation frequency fs of the oscillation circuit can be in a range of approximately 100mQ to 2Q. Preferably, the delay line DL is configured such that the oscillation frequency fs is closer to the series resonance frequency fsER than to the parallel resonance frequency fPAR, thereby maintaining a low resonant circuit impedance while operating at a frequency range where the resonant circuit provides the 90° phase shift. The time delay caused by delay line DL needs to be relatively short, as the series resonance and the parallel resonance of a parallel oscillating circuit PRC are close to each other, relative to the overall frequency range. Preferably, the delay caused by delay line DL that acts of feedback loop 270 should be in a range between 5% to 35% of the period of the parallel resonance frequency fPAR, providing that the above two conditions (i) and (ii) are fulfilled. In an embodiment, the delay caused by the delay line DL that acts on feedback loop 270 is in a range between 35% and 90% of a difference between the period of the parallel resonance frequency fPAR and the period of the series resonance frequency fsER, again providing that the above two conditions (i) and (ii) are fulfilled, more preferably a range between 50% and 85%. For example, taking the illustration of FIG. 6 and as a non-limiting numerical example, assuming that a parallel resonance frequency is at 1 .25 GHz, therefore having a period of 800ps (picoseconds), and a series resonance is at 855MHZ, and therefore a period of 1169ps, there is a difference of 369ps between the period of fPAR and fsER- The time delay caused by delay line DL can be in the above range, for example at 70% of the period difference between fPAR and fsER, thereby being 258ps, thereby making sure that the feedback loop 270 has the desired inductive behavior and low impedance that is necessary to provide inverting feedback at high gain.

[0211] Preferably, the delay line is implemented as a meandering conductive element having dominantly inductive behavior, for example a meandering element having from two (2) to twelve (12) meandering branches, more preferably from three (3) to eight (8) meandering branches. Such implementations exhibit minimal stray inductive and capacitive behavior. Various delay line structures can be used to provide the desired function, for example an Omega-shaped coil, single planar coil, flat inductor, wavy line, zig-zag line, or a sawtooth line. It is also possible to provide the required delay line functionality by a specific transmission line design. For example, it is possible that the physical element of the delay line DL is implemented as a conductor in a printed circuit board, for example implemented as a microstrip patch antenna. In some embodiments, a low-pass filter may be used as the delay line DL, however this will have an impact on the shape of the oscillating voltage, whereas a delay line DL that provides for a short time delay by inductive effect will not impact the wave shape. In the embodiment illustrated in Figure 4, delay line DL is placed P / 90486.W001 between the feedback loop output of the resonant circuit 272 and capacitive element 274, but other arrangements are also possible.

[0212] Figure 7 is a schematic illustration of a control system for the dielectric heating aerosolgenerating system shown in figure 1 , according to embodiments of the disclosure.

[0213] An oscillation unit OSC having an inverting switching unit and an LC feedback loop is provided, the feedback loop including two inductors L1 , L2 and the load capacitor CL. The load capacitor CL comprises a removable aerosol-forming substrate S arranged within an electrode arrangement E1 , E2 for exposure to an alternating electric field, for dielectric heating. An electric field frequency probe ANT, for example an antenna, is arranged to be exposed to the alternating electric field to generate a signal that is indicative of the frequency of the alternating electric field, which is provided to a frequency counter FC. The frequency counter FC delivers its signal on the frequency to a controller, for example a microprocessor.

[0214] In some embodiments, the control system may utilize one or more of resonant antennas, microstrips, waveguides, for high-frequency sensing, or a device that can measure the frequency of the magnetic field departed by any one of the two inductors L1 , L2.

[0215] The controller can be operatively connected to a user interface Ul to allow for communication of statuses and results to a user, for example but not limited to one or more LEDs, haptic feedback, display screen, audio feedback, or a combination of those. The controller can also control or otherwise act on the electric power flow that is supplied to the oscillator unit OSC, with a power supply unit PSU. For example, the power supply unit PSU can be instructed increase or decrease a supply voltage by acting on a DC-DC converter (e.g. boost converter, buck-boost converter, buck converter, full bridge or half bridge DC-DC converter) to provide for a variable voltage to oscillator unit OSC, or by performing an on / off control to control the electric power delivery to the DC-DC converter with a fixed voltage supply, for example by bang-bang chopping control or pulse width modulation (PWM).

[0216] Figure 8 is a schematic illustration of a control system for the dielectric heating aerosolgenerating system, according to embodiments of the disclosure. In the system illustrated in figure 8, an oscillation circuit is coupled to dielectric heater assembly comprising an electrode arrangement having a pair of electrodes E1 , E2 configured to dielectrically heat an aerosol-forming substrate S. The pair of electrodes E1 , E2 are aligned such that they are not parallel to each other or by having a variable distance between each other along a given direction, in order to produce a varying electric field strength along the aerosol-forming substrate S for a given RF voltage supplied to the electrodes. The exemplary system further comprises two temperature sensors P / 90486.W001 communicatively coupled with a control unit, the two temperature sensors being positioned at opposite ends of the aerosol-forming substrate S.

[0217] In an embodiment, the control unit is configured to estimate a temperature distribution profile inside a substrate S, based on temperature measurements from one or from both the two temperature sensors outside of the substrate S, for example a temperature distribution profile along an insertion or longitudinal direction of the substrate S. This allows to determine a spatial location of a desired target temperature within the substrate, so that the heating can be controlled to “move” the desired target temperature Ttarget through the substrate in a controlled manner. With reference to figure 8, a spatial temperature distribution can be created inside the substrate S, where a higher temperature is established inside the substrate S at the left side where the electrodes are closer to each other, whilst the temperature on the right side is lower, as the electrodes are more spaced apart. This spatial temperature distribution profile can be timely moved to the right, so that the target temperature Ttarget is “moved” through the substrate, for example a specific vaporization or aerosolization temperature (e.g. 220°). For example, this can be done by gradually increasing the supply voltage to the oscillation circuit.

[0218] Accordingly, the control unit is configured to use the external temperature values to estimate, within a short time interval, the temperature values of the spatially-variable temperature profile along to X-axis (as referenced in the figure), as the input values. For example, using a model (e.g., a thermal dynamic modelling, or an artificial intelligence model) that can be trained with data from a test temperature sensor system that crosses the substrate S along the X-axis, the testing temperature sensor being able to pick up different temperature values along the X-axis. The trained model, upon being fed with the real temperature measurement, can calculate or estimate, within a given time period, an estimated or inferred spatially variable temperature profile, using the control unit. As further data processing step, the control unit can determine at with location along the X-axis the desired target temperature is located.

[0219] This parameter estimation allows the control unit to move the target temperature Ttarget through the substrate S in a controlled matter, as the control unit can increase the power supply voltage to the oscillation circuit (or alternatively take other action, e.g. increase the oscillation frequency), to increase the heating power. For example, the control unit can have a data on a desired location of a given target temperature at a given time in the substrate, and the estimated / calculated location of the target temperature can be subtracted therefrom, to a have a location difference that can be used in a control feedback loop control.

[0220] Such controlled moving of the target temperature can be used for a controlled and zoned release of the aerosol throughout a substrate that has one dominant direction, e.g. a direction P / 90486.W001 along the cylindrical axis of a stick-like consumable, or a longitudinal direction along the insertion direction of a flat consumable.

[0221] In an alternative embodiment, instead of estimating the temperature, or in combination with an estimation of the temperature, it would be possible to measure an electric field strength at one or more locations inside the heating region as an input value, for example at locations between the electrode plates E1 , E2, and thereby estimate an electric field strength profile along the X-axis. The calculation or estimation of the electric field distribution profile could be based on a knowledge table or formula, based on the geometry of the electrode plates E1 , E2. This would allow to precisely control an electric field distribution profile. Given that the electric field distribution is given by the geometry and arrangement of the electrodes, one electric field measurement could be sufficient to calculate an entire electric field profile along a given direction.

[0222] Figure 9 is a flow diagram illustrating a method for parameter estimation in a dielectric heating aerosol-generating device, according to embodiments of the disclosure. Figure 9 illustrates how the output variable estimation and response may be applied to both a pre-heating stage of a heating period and an inhalation session stage of the heating period. In some embodiments, different input parameters may be measured during the different stages of the heating period in order to estimate different output variables. In some embodiments, a plurality of different input parameters may be measured during each stage of the heating period in order to estimate a plurality of different output variables.

[0223] Figure 10 is a schematic illustration of the processing steps carried out by a control unit to estimate an output variable value, according to embodiments of the disclosure. As an initial step S10, measured input parameters are received by the control unit. The control unit is configured to analyse S20 the input parameters to determine one or more properties associated with the measured input parameters. The control unit is then configured to process S25 the one or more properties associated with the measured input parameters using a model to map S30 the measured input parameter to an output variable value. As described throughout the disclosure, the model may comprise one or more of: a database or look-up table of input parameters and correlated output variable values; an artificial neural network trained using a plurality of known output variable values associated with a range of input parameter values for the one or more input parameters during the heating period of an aerosol-generating device; or a mathematical mapping function.

[0224] Figure 11 shows a method for generating a data structure as a software code or firmware to characterize heating period output variable values or firmware generation for heating period output variable value prediction. In a first phase, a plurality of data sets is stored for a number of different P / 90486.W001 volume buckets (the higher the number of volume buckets, the higher will be the accuracy of the predictions). Data is collected for each volume bucket 410. Each data set is collected for different operating conditions and scenarios. For example, a machine learning model may be trained using consumables with different Al, consumables with different cast leaf agglomeration, consumables with different humidity level, or devices with different battery status. Collected input parameters may include all possible parameters which can be measured during the heating session, for example, the parameters may include the voltage, the current, the power, temperature of the thermistor (used as puff sensor), electric field strength and oscillation frequency. To create training data, a precise output variable value is known for all of the different environmental conditions. The machine learning model is trained based on the fact that data varies (because of the different scenarios used) but they all return output variable values characterized by the same generating device 420.

[0225] Using the training data, the machine learning algorithm is capable of establishing correlations between the input parameters which are used during real device operations to make output variable value predictions. Once the training data is collected, it is then fed to the machine learning algorithm to be trained and to ultimately return a “trained” firmware 430 capable of, during operation of the device, receive input parameter samples during the heating session (such as current, voltage, temperature, electric field frequency, electric field amplitude and the like) and return in output a prediction of the output variable value (such as one or more temperature values within the aerosol-forming substrate) for the heating session. The “trained” firmware 430 is loaded into the aerosol-generating device.

[0226] Figure 12A illustrates a method 500 for producing a machine learning model for an aerosolgenerating device for generating an aerosol from an aerosol-forming article. The method 500 includes training 510 the machine learning model using a plurality of known output variable values over a range of operating conditions associated with one or more parameters of the aerosolgenerating device. The trained machine learning model is then transferred to the aerosolgenerating device 520. For example, the machine learning model is trained 510 using data collected from aerosol-generating devices with a plurality of different properties and a plurality of different operating conditions. For example, the plurality of known output variable values can be known temperatures or temperature profiles of an aerosol-forming substrate and the current output variable can be a current temperature or temperature profile of the aerosol-forming substrate that is heated by the aerosol-generating device. The plurality of different properties includes one or more of different aerosol-forming articles, different operating power, different humidity levels, different environmental pressure levels, different temperature levels, or different aerosol-forming P / 90486.W001 article states, different types of aerosol-forming articles, oscillation frequency, dielectric losses, switching losses, electric field strength, heating region temperature for example.

[0227] Once the machine learning model is trained by step 510, data of the machine learning model can be uploaded to the aerosol-generating device for operation. This can be done via a communication interface, for example but not-limited via a USB-C data interface, where data of the machine learning module is uploaded as a firmware, or as a data structure that can be accessed by the existing firmware, to a memory device of the aerosol-generating device. The machine learning model can thereafter be used by the controller of aerosol-generating device for operation. Uploaded to the memory device of the aerosol-generating device, the machine learning model can be configured to receive as input one or more measured parameter values of the aerosol-generating device associated with current operating conditions and output a current output variable during the usage session. The current output variable may be a current puff volume, for example. The machine learning model can be based on different machine learning networks or structures, and can comprises one or more of a neural network, for example but not limited to an Artificial Neural Network (ANN), Convolutional Neural Network (CNN), Convolutional Recurrent Neural Network (CRNN) and Recurrent Neural Network (RNN), k-nearest neighbors algorithm (k- NN), support vector machines (SVM), a linear regression model, a decision tree, a decision forest, a random forest (RF), multi-layer perceptrons (MLP) fuzzy logic, correlation model, a classification model, and a clustering model.

[0228] During training 510 of the machine learning model, it is possible that different types and architectures of machine learning models are trained, and the resulting trained machine learning models are benchmarked, classified, or ranked with a test data set, so that a ranking for the different trained machine learning models is generated for comparison purposes. In this respect, the method of training 510 can include a sub-step where results of the different machine learning models are displayed as a ranking, for example based on different ranking criteria. This sub-step can be such that a list is displayed in a window or area of a graphical user interface (GUI) of a display device of a computer. This allows the operator or user of the training method to select one of the different machine learning models for implementation with an aerosol-generating device. In this respect, the different machine learning models can be tested by inference of the trained machine learning model with benchmark or test data, to be ranked with a ranking value, for example a percentage value of errors or deviation from an actual value of the output variable, for example, a cumulative puff volume during a usage session. Thereby, it is possible to choose the best suited trained machine learning model for upload and operation with the aerosol-generating device for operationalization. For example, a purely result-oriented comparison can be done P / 90486.W001 related to the estimated or calculated one or more output variables, where the performance of the test or benchmark data for the specific output variable is used. For example, a trained machine learning model could be chosen based on the best accuracy of the estimation or calculation of the temperature or temperature profile of the aerosol-forming substrate. This could be a statistical variation for an individual heating period, or cumulative over multiple heating periods or usage sessions. However, it is also possible to take other results into account, for example use of memory space used by the uploaded and trained machine learning model, power consumption of the trained machine learning model, latency time duration of the estimation or calculation by the trained machine learning model, potential false result generation of the trained machine learning model, for example caused by local maxima / minima.

[0229] Figure 12B shows a method 550 for characterizing an output variable value using a machine learning model on an aerosol-generating device. One or more parameters associated with current operating conditions of the aerosol-generating device are monitored 560 during operation of the aerosol-generating device. The one or more parameters associated with the aerosol device include one or more of a voltage, a current, a power, and a temperature associated with the aerosol-generating device, an oscillation frequency of the oscillation circuit, a temperature of a switching unit, a temperature of the heating region, a frequency of the alternating electric field at the load capacitor.

[0230] A machine learning model is used 570 to determine a current output variable based on the one or more parameters. The current output variable may control operation of the aerosolgenerating device. For example, the current output variable may be used to end a usage session of the aerosol-generating device, for example, stopping or ramping down the heating. While examples herein describe the current output variable as being a temperature or temperature profile of the aerosol-forming substrate, it is to be understood that the current output variables may include one or more other output variables associated with the aerosol-generating device, for example based on other parameters associated with the dielectric heating operation.

[0231] Figure 13 illustrates a method of characterizing a usage session. A user inserts an aerosolforming article into the aerosol-generating device and initiates a usage session by actuating a user button. This indicates the start of the usage session 601 . Power is supplied from a battery in the aerosol-generating device to the heater until the temperature of the aerosol-forming substrate reaches a predetermined operating temperature. This temperature may be, for example, about 330 degrees Centigrade.

[0232] Parameters associated with the aerosol-generating device are monitored 602. In an example, the power signal is monitored 602. When the user puffs 603, the heater is cooled P / 90486.W001 because of the cooling effect of the airflow, as environmental and relative cold air is drawn into the heated substrate by the user. Thus, the power that needs to be supplied to the heater to maintain the operating temperature increases. The power supplied increases and the correct temperature is maintained. The power signal may be determined by using the machine learning model and the one or more sensors described herein.

[0233] The presence of a user puff is detected 604. For example, the user puff may be detected by analysing the power signal or using the machine learning model. A puff start point and a puff end point are determined by means of this analysis. The volume of aerosol generated during the puff is calculated 606 using the machine learning model. The puff volume is added to a cumulative total of volume generated during the usage session.

[0234] It is determined 607 whether a cumulative puff volume equals or exceeds a predefined maximum allowed (threshold) volume. If the cumulative total volume equals or exceeds the predetermined maximum permissible aerosol volume for the usage session (for example 660 ml), the usage session is ended 608. If the cumulative total volume does not equal or exceeds the predetermined maximum permissible aerosol volume for the usage session then the session remains active and the user may take another puff 602. The usage session remains active until the user has generated the maximum permissible aerosol volume or until a maximum time threshold is reached, for example.

[0235] The methods and processes described above can be implemented on computer hardware. In figure 14, a block diagram shows a system and computing apparatus 700 that may be used to implement methods according to an example embodiment (for example, as a computer, a mobile device, a smart sensor, a control system, and the like.). The components may be implemented as integrated circuits (les), portions thereof, discrete electronic devices, or other modules, instruction sets, programmable logic or algorithms, hardware, hardware accelerators, software, firmware, or a combination thereof.

[0236] One or more sensors 712, 713, 714 are disposed on the aerosol-generating apparatus. The sensors may include one or more of a power sensor, a voltage sensor, a current sensor, one or more temperature sensors (for example, a heating region temperature sensor and a PCB temperature sensor), and an accelerometer or a position sensor, or any other sensor described in this disclosure. In this respect the aerosol-generating device can be equipped with an accelerometer or a position sensor, for example an inertial measurement unit (IMU), that can measure movements, orientation, and accelerations to the aerosol-generating device while the user is holding the device during a usage session, or while taking one or more puffs. With data from the accelerometer, it is possible that false or ghost puffs are detected and eliminated from the P / 90486.WG01 puff detection and puff volume. In this respect, some types of motions that are departed to the aerosol-generating device could cause a ghost puff or an air motion that is not intended or done by the user, and data patterns from the accelerometer could provide for valid data to exclude these puffs. Therefore, data of such sensor can be provided as input data to the trained machine learning network for the estimation or calculation of the one or more output variables. It is also that certain types of unreliable or expensive sensors are avoided, or its data is not used. For example, the one or more sensors or plurality of sensors may not include any sensors such as humidity sensors that capture a humidity level from the external environment, as these types of sensors may be easily covered or otherwise obstructed for reliable measurements, or other sensors that can be clogged or otherwise be rendered inoperative due to particle accumulation and other environmental issues, for example flow sensors. The sensor data may be fed into controller 720.

[0237] The control unit 720 may include a user interface 725 conventional computing hardware such as a central processor 721 , memory 722, input / output (I / O) interfaces 723, and a non-volatile data storage unit 724 (for example, hard disk drives, solid state drives). The processor 721 may include any one or more of a microprocessor, a controller, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or equivalent discrete or integrated logic circuitry. In some embodiments, the processor 721 may include multiple components, such as any combination of one or more microprocessors, one or more controllers, one or more DSPs, one or more ASICs, or one or more FPGAs, as well as other discrete or integrated logic circuitry. The functions attributed to the control unit 720 or processor 721 herein may be embodied as software, firmware, hardware, or any combination of these. Certain functionality of the control unit 720 may also be performed in the cloud or other distributed computing systems operably connected to the processor 721 .

[0238] The memory 722 may include any volatile, non-volatile, magnetic, optical, or electrical media, such as a random-access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), electrically erasable programmable ROM (EEPROM), flash memory, or any other digital media. While shown as both being incorporated into the controller 720, the memory 722 and the processor 721 could be contained in separate modules.

[0239] The control unit 720 includes an external data interface 726 that receives data from the sensors 712, 713, 714. The data storage unit 724 stores a machine learning model 728 configured to predict current output variables based on input parameters sensed via sensors 712, 713, 714. Usage session management 732 may be used to determine a usage session and output variable value characterization 730 based on an output from the machine learning model. For example, P / 90486.W001 usage session management may determine that a usage session should be ended based on an output variable value reaching a predefined threshold.

[0240] Figure 15 illustrates a data collection scheme for training data for puff volume estimation. In this example, three different volume buckets are shown, but it is to be understood that more or fewer known volume buckets may be used for the training data. Each of the volume buckets determines various parameters for four different scenarios. It is to be understood, the more or less than four scenarios may be used.

[0241] Table 1 shows predictions versus measurements for puff volume using the machine learning model.

[0242] Table 1

[0243] In this example, the accuracy for the 35ML and the 70ML volume buckets was 84% and the accuracy rating of the 55ML puff volume was 86%. In this example, more than 20k candidates were used to train the model. The candidates are split 80 % for training and 20 % for testing. The prediction result is the result of correct identification of classes on the 20% test candidates. The data was generated by using a classifier model trained for three (3) classes. First, the data was split into two (2) buckets, with one bucket for training, and the other one for testing. This can be done automatically by a machine learning tool, for example a commercially available machine learning tool that can operate on a commercially available computer, such as a PC or a Macintosh computer. The machine learning tool then generates a first model. This first model is then run on the computer and can be inputted with test data set or a benchmark data set for so that the performance or accuracy of the first model can be evaluated. The first model running on the computer can output the score / accuracy of test data or benchmark data set. The accuracy numbers, for example, in percentage of error, are represented in Table 1 . Next, one or more additional models can be tested with the test or benchmark data set, to see the score / accuracy data for comparison and selection purposes.

[0244] Figure 16 illustrates a data collection method for generating a data structure as a software code or firmware to characterize puff volume using human puffs by a single user or a plurality of users of an aerosol-generating device 910. The method can be performed instead of using the P / 90486.W001 smoking or inhalation machine in a controlled manner to generate the airflow, or can be performed by using both the smoking or inhalation machine. The human puffs are taken during a high number of usage sessions with the aerosol-generating device 910. The puffs taken by human inhalation are measured and characterized accurately by using a reference flow meter 920 that provides flow measurement data 930. The reference flow meter is fluidically arranged between a mouth of the user and the aerosol-generating device 910. In an example, the reference flow meter 920 is disposed on a mouthpiece side of the aerosol-generating device 910. This flow measurement data 930 may be used to provide a reference value 940 for the puff volume for each puff, instead of using the airflow generated by the smoking or inhalation machine, or reference values can be generated by both human inhalation and smoking or inhalation machine. For each puff various other internal sensor data 950 is obtained from the aerosol-generating device 910. For example, one or more of a power sensor, a voltage sensor, a current sensor, one or more temperature sensors (for example, a heating regiontemperature sensor and a PCB temperature sensor), and an accelerometer. The sensor data 950 and the reference values 940 may be used to generate training data for use with one or more different machine learning algorithms to generate one or more trained machine learning models 960, 970, 980. The trained models 960, 970, 980 may include one or more of an aerosol-generating device simulator, a model generated with a classifier and a model generated with a classifier with a low memory footprint.

[0245] In a variant, instead of performing a data collection method for generating a data structure to characterize puff volume, it is possible that other types of known output variables mentioned in this disclosure are used to train the machine learning model. For example, it is possible to train a machine learning model for detecting whether the aerosol-forming substrate is hot (exceeds a certain threshold temperature) or not while a puff is taken. With such trained machine learning model, it is possible to provide for information to the aerosol-generating device 10 that will allow to mitigate against the hot aerosol effect. In such system that performs the data collection method, a reference temperature sensor can be arranged between the mouthpiece of the aerosol-generating article and the smoking or inhalation machine (or human user), that can be used to generate a reference temperature value during a plurality of puffs, serving as an example of a known output variable value. This data can be used for training the machine learning model, such that the trained and operationalized machine learning model can infer a temperature of the aerosol that exits from the mouthpiece. As another example, the machine learning model can be trained by a human user, where the identity of the user serves as a known output variable. P / 90486.W001

[0246] Figure 17 shows that the energy strictly associated to a puff may be calculated as the integral calculus of the power signal during the puff, minus the energy that would be spent anyway even without a puff, as indicated in [1] below.

[0247] Similarly, it is also possible to correlate the power to the air flow, which equals the volume per time unit. One or more of the sensors described herein may be used to determine the data indicative of a change in heating power as a result of air cooling.

[0248] The usage session has a maximum permissible volume of aerosol to be delivered. Every puff contributes to the maximum permissible volume. Once the threshold has been reached, the experience ends. Therefore, the experience is not tied to a predetermined number of puffs, but to the way the user actually puffs on the device. According to another embodiment, it is possible to use data gathered from the aerosol-generating device during the one or more usage sessions by a specific user can be used to further improve the machine learning model. For example, according to an embodiment, a method for improving a machine learning model can be provided, where first a user uses his aerosol-generating device, so that data on the usage sessions can be recorded. For example, data from one or more sensors can be recorded to a memory device during one or more usage session from a user, for example power data on the power consumption of the heater, and temperature data from the one or more temperature sensors. In another variant, data from an acceleration sensor, for example an IMU can be recorded. In addition, the corresponding generated or estimated output data of the machine learning model can be recorded to a memory device, which includes a current output variables that will be based on the recorded one or more parameters, to thereby create a user-specific data set of the usage and results of the currently uploaded and operating machine learning model. Next, in another step, a data communication can be established with an external data processing device, to download the user-specific data set that has been generated by the trained machine learning model and provided to a computing environment to further improve the trained machine learning model, by using the user-specific data set. The downloaded data set can be used to correct an error in the calculation or estimation.

[0249] Figure 18 is a graph illustrating the relationship between dielectric heating power and the oscillation frequency profile for a dielectric heating operation with an aerosol-forming device or system that has a dielectric heating assembly, according to embodiments of the disclosure. The P / 90486.WG01 long line on the left of the graph represents a sweep of frequency to locate the resonant frequency, and then the more horizontal lines represents the oscillation frequency. In the graph, the red plot starts at a frequency of approximately 2.67 GHz and then overtime shifts towards lower values, ending at around 2.61 GHz. The frequency decreases as the heating operation progresses due to temperature change. As the aerosol-forming substrate is dielectrically heated, the substrate is brought above vaporization temperature of one or more substances within the aerosol-forming substrate with the consequence of evaporating polar molecules which are the main ones causing the heating. The system can detect that the load has changed. So, the primary effect of the heating is the change of chemistry of the substrate, for example caused by depletion of liquid aerosolformer from the substrate by vaporization.

[0250] Figure 19 is a graph illustrating the temperature of an aerosol-forming substrate measured as a function of time and heating power in an aerosol-generating device or system during a dielectric heating operation. Accordingly, within the first few seconds of a heating operation, a higher heating power leads to higher substrate temperatures.

[0251] The oscillation frequency-heating power relationship for the aerosol-generating device or system illustrated in figure 18 may be used in conjunction with the substrate temperature-heating power relationship for the aerosol-generating system illustrated in figure 19, in order to create a correlation between each substrate temperature value associated with a respective oscillation frequency value during a heating operation. The correlated values may be stored in memory in a control unit of the aerosol-generating device. Alternatively, a mathematic mapping function may be derived to transform an oscillation frequency value into a corresponding substrate temperature value, wherein the mapping function is stored in memory in a control unit of the aerosol-generating device. It will be appreciated that a similar approach may be used to determine models for estimating other output variables associated with dielectric heating operations.

[0252] Figures 20a, 20b, and 20c show different load-dependent parameters of the electric response of an electrically resonant oscillation system or electromagnetic field generator of an aerosol-generating device, according to embodiments of the disclosure. Each of Figures 20a, 20b, and 20c show frequency responses of the electrically resonant oscillation system of an aerosolgenerating device (i.e. the dependence of the signal strength on the frequency of the applied electric field) in different configurations.

[0253] Figure 20a shows the frequency response of an electrically resonant oscillation system or electromagnetic field generator when an aerosol-generating article is received in the article region and when an aerosol-generating article is not received in the article region. As shown in Figure 20a, the frequency response of the electrically resonant oscillation system when an aerosol- P / 90486.W001 generating article is received in the article region is different to the frequency response of the electrically resonant oscillation system when an aerosol-generating article is received in the article region. The frequency response of the electrically resonant oscillation system when an aerosolgenerating article is received in the article region is shifted, such that the resonant frequency, cor, (i.e. the peak of the response) when an aerosol-generating article is received in the article region occurs at a different frequency to the resonant frequency, coO, when an aerosol-generating article is not received in the article region. This frequency shift can be measured and used to determine when an aerosol-generating article is received in the article region. The magnitude of the frequency shift can also vary depending on the electrical properties of the aerosol-generating article received in the article region. As such, the magnitude of the frequency shift can also be used to determine the type of aerosol-generating article received in the article region.

[0254] Figure 20b shows a frequency response of an electrically resonant oscillation system when an aerosol-generating article is received in the heating region of an aerosol-generating device. The frequency response has a bandwidth, Aw. The bandwidth, Aw, of the frequency response can also be indicative of the type of aerosol-generating article received in the article region. As such, the bandwidth of the frequency response may be used to determine the type of aerosol-generating article received in the article region.

[0255] Figure 20c shows three different frequency responses of an electrically resonant oscillation system, when a first aerosol-generating article is received in the article region, when a second aerosol-generating article is received in the article region, and when no aerosol-generating article is received in the article region. As shown in Figure 20c, each of the frequency responses is different. For each frequency response, the frequency shift and the bandwidth is unique. Accordingly, a combination of the frequency shift and the bandwidth of the frequency response may be used to determine whether an aerosol-generating article is received in the article region, and the type of aerosol-generating article received in the article region. Using more than one loaddependent property of the frequency response of the electrically resonant oscillation system may provide a more accurate determination of whether an aerosol-generating article is received in the article region, and the type of aerosol-generating article received in the article region, compared to using a single load-dependent parameter.

[0256] In some embodiments, the control electronics are configured to measure load-dependent parameters by varying the frequency of the alternating electric field generated in the article region. Accordingly, the control electronics are configured to control the aerosol-generating device by controlling the supply power to the dielectric heater to generate an alternating electric field in the article region with a varying frequency, the frequency varying in a sweep over a predetermined P / 90486.WG01 range of frequencies. The control electronics monitor the electric response of the oscillation system across the range of varying frequencies, and generate a frequency plot, which enables loaddependent parameters, such as the resonant frequency, to be identified, and used to identify the type of aerosol-generating article that is received in the article region.

[0257] Figure 21 is a graph illustrating how the frequency dependency of the relative permittivity of Glycerol varies with temperature, an example of an aerosol former commonly used in aerosolforming substrates. As shown in figure 21 , the relative permittivity of Glycerol increases with temperature (see e.g. at 100MHz, at 273K about 15, at 295K about 40), leading to a higher capacitance value of the load capacitor, causing the oscillation frequency of the self-oscillating oscillation circuit to drop. Conversely, the frequency-dependency of the relative permittivity of Glycerol can be seen, showing that with an increase of the oscillation frequency of the alternating electric field, the relative permittivity decreases, leading to a lower capacitance value of the load capacitor. This is due to the so-called relaxation effect of the Glycerol. This is turn will force the oscillation to increase. It can also be seen that the drop of relative permittivity can be substantial for a given frequency range.

[0258] Figure 22a illustrates the evolution of the switching frequency of the oscillation circuit over time, during a heating session including a heat-up phase and a steady state phase or session / experience phase, where the puffs are taken.

[0259] At time instant A the heating session is started with the heat-up phase, and the oscillation frequency starts at an initial value, for example, at about 500MHz. The oscillation frequency then decreases as the glycerol becomes hotter, as the dielectric constant increases (leading to a higher value of the load capacitance), until a time instant B is reached, during which there is no change in oscillation frequency as there is no change in relative permittivity. Time instant B will happen as soon as a predeterminable temperature of the substrate has been reached and can serve as a temperature and timing marker point. As an example, at an oscillation frequency of about 500 MHz, for Glycerol as the aerosol former, the temperature can be about 73°C.

[0260] After the time instant B, there is a gradual increase of the oscillation frequency resulting from a decrease in relative permittivity of the substrate, due to some temperature-related evacuation of liquids from the substrate S. With the further advancing of time, the oscillation frequency increase accelerates, as higher temperatures are reached and some vaporization begins to happen, notably the vaporization of water at 100°C. The temperature-related depletion of liquids from substrate S thereby counteracts the temperature dependent increase of the relative permittivity. During this transition from B to START, the temperature of the substrate will reach the P / 90486.W001 steady-state temperature for aerosolization, for example a temperature between 200°C and 250°C, a time moment where the experience or user session can start, allowing the user to take puffs.

[0261] After time instant START, with puff-taking events starting at C1 , C2, etc. and ending at E1 , E2, etc., respectively, with the transition D.

[0262] At a given time instant C1 , the user takes a first puff. This leads to a sudden evacuation of aerosol-former from substrate S, which leads to a sudden decrease of the relative permittivity of substrate S, leading to a sharp increase of the oscillation frequency. Therefore, upon the end of the puff at time instant E1 , the oscillator will operate at a higher frequency than at instant C1 . During the time between puffs, for example the time duration labelled as D, it shows that the oscillation frequency slightly decreases again due to the cooling effect to the substrate from the colder incoming air drawn by the puff, but shortly after starts increasing due to constant evaporation of aerosol-former at the steady-state temperature. Additional puffs that are taken at time instances C2, C3, C4, etc. show the stepwise increase of the oscillation frequency triggered by the puff, however, with the increased number of puffs (2 and more), the cooling effect by the incoming air that leads to a decrease in oscillation frequency decreases and then disappears, see, e.g. at C3 to E3.

[0263] Also, it can be seen that over the entire heating process (heat-up phase and steady state phase) the oscillation frequency increases with increased depletion of the substrate S, until there is no observable change in the oscillation frequency after and before a puff, indicating that the substrate is fully depleted of aerosol-former.

[0264] Figure 22b is a graph illustrating a first derivative of the graph of figure 22a. The use of the first order derivative (and also the second order derivative to further enhance detection and estimation results) can be used to estimate different parameters around the aerosolization process of substrate S. For example, with the zero crossing at time instant B, which occurs once the substrate S reaches a given temperature (e.g. 73°C). Furthermore, amplitude peaks (AM1 , AM2 AM3 etc...) on the first order derivative of the measured frequency response over time may be indicative of both user puffs and puff volumes.

[0265] Figure 23 is a flow diagram illustrating a control process for a pre-heating phase of a dielectric heating operation using the control unit according to figure 7. The pre-heating phase is a time period where the heater starts heating such that the substrate is brought to a temperature for inhalation, for example 180°C to 220°C. For HnB substrates, this time period is usually about 10 seconds to 40 seconds, while for e-vapor and liquid vaporization, this time period is usually much shorter, 200ms or less. P / 90486.W001

[0266] Upon starting the heater (start pre-heat), for example by powering the oscillator unit OSC, data indicative of the frequency of the oscillator unit OSC is captured and processed, for example to normalize the date, filtered for noise, etc. Next, the data can be analyzed to detect a pre-heating frequency profile, that allows to estimate or detect one or more parameters of the pre-heating phase, for example to estimate parameters around the substrate S before the substrate S is ready for the user session. For example, the pre-heating frequency profile can be processed and compared to reference data. The pre-heating frequency profile can be processed to perform a first or second order derivative, to compare values obtained with reference data, to detect and estimate a time instant and value of the maximal frequency obtained (see time instant B above), serving as an indicator for the temperature. This reference temperature value and its timing can serve to determine a time instant when the start-up phase can be stopped, and the steady state heating phase can be performed.

[0267] Based on the analysed pre-heating frequency profile, by performing a comparison with reference data, one or more parameters related to the article can be detected. As non-limiting examples for the one or more parameters, it can be detected whether the article has been previously used and is fully or at least partially depleted (showing oscillation frequencies that are too high), whether the substrate S or article have been inserted in the correct direction (also showing too high oscillation frequencies), determine the type of substrate or article, detect humidity content of the article, determine atmospheric conditions of the environment. In the context of e- Vapor and liquid substrates, it can be determined whether there is sufficient liquid in a liquid transfer element to safely allow for a puff.

[0268] This analysis may facilitate the performance of different tasks by the controller, for example to instruct different actions, change settings or the profile of the device, or inform the user via Ul based on the one or more parameters. For example, actions can be taken such as aborting the heating if a used article or unauthorized article is being heated, determining a time moment when to switch from the pre-heating phase to the steady-state phase (e.g. predicting when the steadystate temperature of the substrate is reached).

[0269] In some embodiments, the controller may set or modify a temperature profile for the steadystate phase for the specific article, set a temperature profile based on the humidity content of the article to mitigate against a hot aerosol effect or based on different atmospheric conditions.

[0270] Parameters or data associated with the parameters can be communicated to the user with a user interface (Ul). For example, an error signal can be communicated, for example via a display or LED light, or instructions can be provided to the user to remedy an error situation, e.g. turning the article, inserting a new article, article identification information, etc. P / 90486.W001

[0271] Figure 24 is a flow diagram illustrating a control process for a steady-state phase of a dielectric heating operation using the control system according to figure 7.

[0272] Initially, data indicative of the oscillation frequency of the oscillator unit OSC is captured and processed, for example to normalize the data, filtered for noise, etc. Next, the data can be analyzed to detect different types of parameters, for example (1) parameters that are specific to an individual puff that has been taken, or (2) parameters related to the overall heating session, or both.

[0273] With respect to (1), a puff can be detected by a change in frequency to the oscillator unit OSC, for example by using the first order derivative and thresholding, see time instances C1 , C2, C3, etc. of figure 22a. Second order derivative can be used to enhance the detection. Amplitude and width (amplitudes AM1 , AM2, AM3, etc., pulse widths) of the first order derivative can undergo data processing to estimate puff strength and puff volume for each puff, to thereby monitor stick depletion.

[0274] In addition, with respect to (2), the measured frequency profile can be continuously analyzed with data processing to detect different parameters, for example to deduct the current temperature of the substrate S, to deduct the stick depletion level. Temperature and / or stick depletion level can be communicated to the user via the III.

[0275] The embodiments described above are exemplary embodiments only, and various other embodiments according with this disclosure are also envisaged.

[0276] For the purpose of the present description and of the appended claims, except where otherwise indicated, all numbers expressing amounts, quantities, percentages, and so forth, are to be understood as being modified in all instances by the term “about”. Also, all ranges include the maximum and minimum points disclosed and include any intermediate ranges therein, which may or may not be specifically enumerated herein. In this context, therefore, a number A is understood as A ± 5% of A.

Claims

P / 90486.W001CLAIMS1 . An aerosol-generating device for dielectrically heating an aerosol-forming article to generate an aerosol, the aerosol-generating device comprising: a dielectric heater assembly having a heating region for removably receiving an aerosolforming article; a radio frequency source for supplying power to the dielectric heater assembly to generate the aerosol over a heating period when the aerosol-forming article is received in the heating region; one or more sensors configured to measure one or more input parameters associated with operating conditions of the aerosol-generating device during the heating period of the aerosolgenerating device; and a control unit comprising a model comprising data indicative of a relationship between the one or more input parameters and one or more temperature values inside an aerosol-forming substrate of the aerosol-forming article over a range of input parameter values, the control unit being configured to: receive the one or more measured input parameters; and indirectly estimate, using the model, one or more temperature values inside the aerosol-forming substrate based on the received one or more measured input parameters.

2. An aerosol-generating device according to claim 1 , wherein the model comprises an artificial neural network trained using a plurality of known temperature values inside an aerosolforming substrate associated with a range of input parameter values for the one or more input parameters during the heating period of an aerosol-generating device.

3. An aerosol-generating device according to any preceding claim, wherein the one or more input parameters are associated with a power consumption in the radio frequency source.

4. An aerosol-generating device according to any preceding claim, wherein the one or more input parameters comprises an oscillation frequency of the radio frequency source.

5. An aerosol-generating device according to any one of claims 1 to 3, wherein the one or more input parameters comprises a value indicative of a strength of an alternating electric field between an electrode arrangement in the dielectric heater assembly, and / or a temperature in the heating region of the aerosol-generating device.P / 90486.W0016. An aerosol-generating device according to any one of claims 1 to 3, wherein the one or more input parameters comprises a value indicative of a frequency of an alternating electric field in the heating region, and wherein the control unit is further configured to: indirectly estimate, an aerosol-forming article substrate depletion level based on the received one or more measured input parameters.

7. An aerosol-generating device according to any one of claims 1 to 3, wherein the one or more input parameters comprises a temperature at a plurality of different locations along the heating region, and wherein the control unit is further configured to: indirectly estimate, an aerosol-forming substrate temperature distribution profile based on the received one or more measured input parameters.

8. An aerosol-generating device according to any one of claims 1 to 3, wherein the one or more input parameters comprises a value indicative of a strength of an alternating electric field at a plurality of locations within an electrode arrangement of the dielectric heater assembly, and wherein the control unit is further configured to: indirectly estimate, an aerosol-forming substrate temperature distribution profile based on the received one or more measured input parameters.

9. An aerosol-generating device according to any one of claims 1 to 3, wherein the one or more input parameters comprises a temperature of a switching unit in the radio frequency source and a temperature in or adjacent to the heating region of the aerosol-generating device.

10. An aerosol-generating device according to any preceding claim, wherein the control unit is configured to measure the one or more input parameters over a period of time to determine one or more input parameter profiles for the dielectric heating operation.

11. An aerosol-generating device according to claim 10, wherein the control unit is configured to determine one or more properties associated with a first order derivative or a second order derivative of the one or more input parameter profiles.

12. An aerosol-generating device according to any preceding claim, wherein the control unit is configured to instruct an action based on the estimated one or more temperature values inside the aerosol-forming substrate.P / 90486.W00113. An aerosol-generating device according to claim 12, wherein the action comprises adjusting a power supplied to the dielectric heater assembly.

14. An aerosol-generating device according to any one of claims 12 to 13, wherein the heating period includes a pre-heating period and an inhalation session period, wherein the action comprises a change from the pre-heating period to the inhalation session period.

15. An aerosol-generating system comprising: an aerosol-generating device according to any preceding claim; and an aerosol-forming article received in the heating region of the aerosol-generating device.

Images