An intelligent identification method for abnormal state of a humidifier

Abstract

The application provides a humidifier abnormal state intelligent identification method, comprising: acquiring a water tank inversion signal through a hardware interface, collecting an initial driving power of an atomizing piece in a starting stage and a starting transient oscillation signal according to the water tank inversion signal; performing frequency domain conversion on the starting transient oscillation signal to obtain a frequency domain response distribution of an atomizing cavity, and determining the sound radiation pressure intensity in the atomizing cavity according to the initial driving power and the frequency domain response distribution of the atomizing cavity; according to the loss degree of the atomizing amount and the compression degree of the bubble, the serious grade of the air cushion layer retention is divided, and the serious grade covers three levels of mild bubble local aggregation, moderate air cushion sheet covering and severe air cushion overall covering; and according to the serious grade, a graded early warning signal for the starting stage is output, and the graded early warning signal corresponds to the indication of the atomizing starting state of the humidifier and the driving power compensation demand.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to an intelligent method for identifying abnormal states of a humidifier. Background Technology

[0002] Humidifiers are core devices for improving indoor microclimates and are widely used in living rooms, bedrooms, offices, baby rooms, and medical care environments during dry seasons. The stability and reliability of their atomization startup directly affect the user experience and determine the long-term lifespan of the device, making them a crucial aspect of smart home appliances. An existing technology, CN106642518A, discloses a method and device for detecting water shortage in a humidifier. This method uses a water level sensor to monitor the water level parameters in the tank, controls a solenoid valve to add water to the tank based on the water level status, and confirms water shortage when the solenoid valve is energized more than a preset number of times and the tank remains water-deficient. However, this disclosed technology relies solely on water level parameters and water replenishment results to determine water shortage anomalies. It fails to detect issues such as bubble retention caused by rapid tank inversion or immediate startup after movement, sound energy transmission obstructed by the air cushion layer, and resulting startup oscillations, buzzing, and startup failures. In everyday use, user behavior is far more complex than the ideal model suggests.

[0003] For example, users often quickly invert the water tank when adding water, or turn on the humidifier before it has completely settled, or even press the switch directly after moving the humidifier. These seemingly ordinary actions instantly draw a large amount of air into the atomizing chamber, forming a cluster of bubbles. These bubbles are not evenly dispersed; rather, due to the limitations of the atomizing chamber structure and the characteristics of water turbulence patterns, they are more likely to form a dense accumulation area directly above the atomizing plate. When an ultrasonic humidifier's atomizing plate is working, it releases acoustic energy into the water through high-frequency vibration, breaking water molecules into tiny droplets. When the surface of the atomizing plate is covered by a cluster of bubbles, the difference in acoustic impedance between the gas and the liquid causes the energy transfer path to become disordered. Following the long-standing hardware design inertia in the industry, in order to shorten the startup waiting time and get the device into normal atomization mode as quickly as possible, there is often a tendency to directly increase the driving power of the atomizing plate during the startup phase, hoping to overcome the surface tension of the water with stronger vibrations.

[0004] When faced with bubble retention, this treatment method significantly increases the driving power, generating a stronger acoustic radiation pressure near the atomizing plate. This pressure, which should push the water to form a mist column, instead compresses and fixes the accumulated bubbles to the surface of the atomizing plate due to the presence of bubbles, making it difficult for them to detach and float. This forms an extremely stable air cushion layer. This air cushion acts like an invisible barrier between the atomizing plate and the water, completely blocking the path of sound energy transmission to the upper liquid layer. Ultimately, this causes repeated, violent vibrations during startup, accompanied by abnormal humming, and can even lead to complete startup failure, forcing the user to shut down and restart, resulting in a very poor user experience.

[0005] Therefore, in special working conditions where rapid inversion of the water tank or similar actions cause air bubble retention, how to identify the physical contradiction between increased driving power and the formation of the air cushion layer, and provide reliable early warning of potential anomalies during the startup phase, has become a key issue in overcoming the bottleneck of intelligent recognition in humidifiers, and is also a specific challenge that the current smart home appliance industry urgently needs to address. Summary of the Invention

[0006] This invention provides a method for intelligent identification of abnormal states in humidifiers, the method comprising: The water tank inversion signal is obtained through the hardware interface, and the initial driving power and transient oscillation signal of the atomizing plate during the start-up phase are collected based on the water tank inversion signal. The transient oscillation signal at startup is converted into the frequency domain to obtain the frequency domain response distribution of the atomizing cavity. The acoustic radiation pressure intensity in the atomizing cavity is determined based on the initial driving power and the frequency domain response distribution of the atomizing cavity. Threshold evaluation of acoustic radiation pressure intensity is performed to screen out the frequency range where the acoustic radiation pressure intensity exceeds the critical value for bubble compression initiation. The high-frequency loss component dissipated by bubble compression within the frequency range is determined based on the frequency domain response distribution of the atomizing cavity. The degree of bubble compression on the surface of the atomizing plate is identified based on the high-frequency loss component. The coverage pattern of the air cushion on the surface of the atomizing plate is identified based on the degree of bubble compression. Based on the coverage pattern of the air cushion on the surface of the atomizing plate, the thickness of the surface air cushion is extracted. Based on the thickness of the surface air cushion, the sound energy blocking rate for sound energy transmission to the upper liquid layer is calculated. Identify the portion of the sound energy blocking rate that exceeds the preset blocking threshold, extract the target driving power when the atomizing sheet generates over-threshold blocking, and determine the degree of atomization loss caused by the air cushion layer based on the target driving power and the surface air cushion thickness. Based on the degree of atomization loss and bubble compression, the severity level of air cushion layer retention is classified, including mild local bubble aggregation, moderate air cushion sheet coverage, and severe air cushion complete coverage. Based on the severity level, a graded warning signal is output for the startup phase. The graded warning signal corresponds to the humidifier's atomization startup status and drive power compensation requirements.

[0007] Preferably, the step of acquiring the water tank inversion signal through a hardware interface and collecting the initial driving power and transient oscillation signal of the atomizing plate during the startup phase based on the water tank inversion signal includes: The inverted signal output by the Hall element at the bottom of the water tank is received through the hardware interface. When the output level of the Hall element changes from low to high, the acquisition timing of the atomizing plate is activated to obtain the start-up phase trigger mark. The current and voltage sampling values ​​of the atomizing plate driving circuit are retrieved according to the trigger flag of the startup phase, and multiplied point by point in the initial power-on interval to obtain the initial driving power and power input curve. The transient oscillation signal is picked up by a piezoelectric vibration sensor. After filtering out power frequency interference, it is aligned with the power input curve according to the acquisition timing to obtain the start-up transient oscillation signal.

[0008] Preferably, the step of performing frequency domain conversion on the start-up transient oscillation signal to obtain the frequency domain response distribution of the atomizing cavity, and determining the acoustic radiation pressure intensity within the atomizing cavity based on the initial driving power and the frequency domain response distribution of the atomizing cavity, includes: The fast Fourier transform is used to perform frequency domain transformation on the start-up transient oscillation signal segment by segment along the acquisition time sequence, extract the amplitude and phase of each frequency point, and arrange the frequency points from low to high to obtain the frequency domain response distribution of the atomizing cavity. The resonant frequency point of the atomizing cavity is determined for the highest amplitude frequency point. The peak values ​​of the spectrum in the neighborhood of the resonant frequency point are extracted, and the amplitude and the initial driving power are aligned along the same time stamp. The square of the amplitude is accumulated point by point to obtain the energy density of the water body. Based on the difference in acoustic impedance between the two sides of the energy density and impedance interface, the radiation intensity pointing upwards to the liquid layer is converted according to the transmission ratio, thus obtaining the acoustic radiation pressure intensity in the region directly above the atomizing plate.

[0009] Preferably, the step of threshold evaluation of acoustic radiation pressure intensity, screening out frequency ranges where the acoustic radiation pressure intensity exceeds the critical value for bubble compression initiation, determining the high-frequency waste component dissipated by bubble compression within the frequency range based on the frequency domain response distribution of the atomizing cavity, and identifying the degree of bubble compression on the surface of the atomizing plate based on the high-frequency waste component includes: The acoustic radiation pressure intensity is compared with the bubble compression start-up critical value at each frequency point, and the frequency points that exceed the critical value are retained and the adjacent frequency points that exceed the critical value are merged into the frequency range. The amplitude values ​​of each frequency point falling within the frequency range are extracted from the frequency domain response distribution of the atomizing cavity. The reference amplitude value in the same range during normal atomization is subtracted from the current amplitude point by point, and the amplitude difference is accumulated according to the frequency point to obtain the high-frequency waste component. Based on the proportion of the high-frequency waste component in the frequency range, and combined with the spectral line broadening amplitude of the resonant frequency neighborhood amplitude spreading to both sides, the degree of bubble compression on the surface of the atomizing plate is obtained.

[0010] Preferably, the step of identifying the coverage pattern of the air cushion on the surface of the atomizing plate based on the degree of bubble compression, extracting the thickness of the surface air cushion based on the coverage pattern of the air cushion on the surface of the atomizing plate, and calculating the sound energy blocking rate for sound energy transmission to the upper liquid layer based on the surface air cushion thickness includes: The edge contour of the air cushion is delineated based on the adhesion distribution of the bubble compression degree on the surface of the atomizing plate. The boundary of the transition from the dense compression area to the sparse adhesion area of ​​the air cushion is determined by the edge contour. The coverage shape of the air cushion is determined based on the proportion of the area enclosed by the edge contour to the sound-receiving surface of the atomizing plate. The coverage shape is determined to show whether the air cushion is locally patchy or continuously covered on the surface of the atomizing plate. The sound energy transmission path through the air cushion is defined according to the coverage shape. The sound energy path length along the transmission path is extracted and converted into the longitudinal distance of gas filling along the normal of the air cushion to obtain the surface air cushion thickness. The transmission margin of sound energy after passing through the air cushion along the transmission path is determined based on the thickness of the surface air cushion. The proportion of sound energy intercepted by the air cushion is obtained by subtracting the transmission margin from the sound energy pointing upwards to the liquid layer before passing through the air cushion. The sound energy blocking rate of sound energy transmission to the liquid layer upwards is obtained by dividing the sound energy before passing through the air cushion.

[0011] Preferably, the step of identifying the portion of the acoustic energy blocking rate that exceeds a preset blocking threshold, extracting the target driving power when the atomizing sheet generates blocking beyond the threshold, and determining the degree of atomization loss caused by the air cushion layer based on the target driving power and the surface air cushion thickness includes: The acoustic energy blocking rate is compared with a preset blocking threshold point by point along the acquisition time sequence, and consecutive points where the acoustic energy blocking rate exceeds the preset blocking threshold are merged into an over-threshold segment; Extract the target driving power applied to the atomizing plate by the atomizing plate driving circuit within the over-threshold range. The target driving power indicates the amount of electrical energy input that is raised when the air cushion is pushed open when the air cushion barrier is increased. Based on the difference between the target driving power and the driving power consumed to maintain an equal amount of mist column during normal atomization, the extra power redundancy consumed by the atomizing plate to open the air cushion is obtained. Based on the power redundancy and the surface air cushion thickness, the proportion of droplets that fall out when the air cushion thickness increases is obtained, and the degree of atomization loss caused by the air cushion layer is determined.

[0012] Preferably, the severity level of air cushion layer retention is determined based on the degree of atomization loss and bubble compression. This severity level covers three grades: mild localized bubble aggregation, moderate sheet-like air cushion coverage, and severe complete air cushion coverage, including: The degree of atomization loss is divided into small, medium and large segments from small to large, and the degree of bubble compression is divided into light, medium and heavy segments from light to heavy. The segments into which the two fall are combined in pairs to obtain the combined values ​​that characterize the retention state of the air cushion layer. Based on the pre-established judgment boundaries between sections and levels, the combination of falling into a small section and a light section is classified as a light bubble local aggregation state, the combination of falling into a medium section and a medium section is classified as a moderate air cushion sheet coverage state, and the combination of falling into a large section and a heavy section is classified as a severe air cushion full coverage state, thus obtaining the severity level of air cushion layer retention.

[0013] Preferably, the step of outputting a graded warning signal for the startup phase based on the severity level, wherein the graded warning signal corresponds to indicating the atomization startup status and drive power compensation requirements of the humidifier, includes: Based on the pre-established mapping between severity levels and warning code values, a warning code value is assigned to each of the following states: mild localized bubble aggregation, moderate air cushion sheet coverage, and severe air cushion complete coverage, thus obtaining a graded warning signal for the initiation phase. Based on the warning code value carried by the graded warning signal, the atomization start-up status indication and drive power compensation requirement are retrieved. The atomization start-up status indication indicates normal start-up, delayed start-up, or start-up obstruction. The drive power compensation requirement indicates the compensation level from low to high. The graded warning signal corresponding to the humidifier atomization start-up status and drive power compensation requirement is output.

[0014] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses an intelligent identification method for abnormal states of humidifiers. Addressing the problem that a cushion layer easily forms on the surface of the atomizing plate after inverted startup, causing a sudden and undetectable drop in atomization efficiency, the method captures the water tank inversion signal via a hardware interface and simultaneously collects the initial drive power and transient oscillation signal during the startup phase. The oscillation signal is then frequency-domain converted to obtain the frequency response distribution of the atomization cavity. Combined with the initial drive power, the acoustic radiation pressure intensity within the atomization cavity is calculated. This allows for the selection of frequency ranges exceeding the critical value for bubble compression startup, from which the high-frequency energy dissipation component dissipated during bubble compression is extracted to identify the abnormal state. The system identifies the degree of bubble compression on the surface of the atomizing plate, then reconstructs the air cushion coverage shape and extracts the air cushion thickness. It calculates the sound energy blocking rate for sound energy transmission to the upper liquid layer, and assesses the degree of atomization loss by combining the target driving power when the blocking exceeds the threshold with the air cushion thickness. Finally, it classifies the severity into three levels: mild local aggregation, moderate sheet-like coverage, and severe complete sealing, based on the degree of bubble compression. It outputs graded warnings and indicates the driving power compensation requirements, thereby achieving refined perception and proactive intervention of abnormal air cushion retention during the startup phase, significantly improving the atomization stability and operational reliability of the humidifier under inverted operating conditions. Attached Figure Description

[0015] Figure 1 This is a flowchart of an intelligent identification method for abnormal states of a humidifier according to the present invention.

[0016] Figure 2 This is a schematic diagram of an intelligent identification method for abnormal states of a humidifier according to the present invention.

[0017] Figure 3 This is another schematic diagram of an intelligent identification method for abnormal states of a humidifier according to the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] like Figures 1-3 This embodiment of a humidifier abnormal state intelligent identification method may specifically include: Step S101: Obtain the water tank inversion signal through the hardware interface, and collect the initial driving power and transient oscillation signal of the atomizing plate during the startup phase based on the water tank inversion signal.

[0020] The hardware interface receives an inverted signal output after the water tank is installed in place. This inverted signal is obtained by a Hall element located at the bottom of the water tank, which senses the rotation direction of the magnet. When the output level of the Hall element changes from low to high, it is determined that the water tank is in a rapidly inverted, installed state. Based on this state, the atomizing plate's data acquisition sequence is activated. This acquisition sequence is limited to the initial power-on interval after the humidifier receives the start command, obtaining a start-up stage trigger marker corresponding to the inverted signal. Based on the start-up stage trigger marker, the current and voltage sample values ​​of the atomizing plate's driving circuit are retrieved. These values ​​are multiplied point-by-point within the initial power-on interval to obtain the initial driving power of the atomizing plate. A corresponding timestamp sequence is marked for this initial driving power, determining the power input curve of the initial driving power changing with the acquisition sequence. The sampling window of the signal channel is defined according to the power input curve. The transient oscillation signal of the high-frequency vibration of the atomizing plate transmitted back through the water is picked up by the piezoelectric vibration sensor attached to the wall of the atomizing chamber. After the transient oscillation signal is filtered to remove power frequency interference by the signal channel, it is aligned with the power input curve according to the acquisition sequence to obtain the start-up transient oscillation signal that synchronously corresponds to the initial driving power during the start-up phase.

[0021] Ultrasonic humidifiers frequently start immediately after water replenishment. When a user quickly inverts the water tank and presses the switch before the device has settled, air is drawn into the atomizing chamber and bubbles accumulate directly above the atomizing plate. This causes the power and vibration during startup to deviate from normal atomization conditions. The following implementation describes the acquisition of the water tank inversion signal, and the subsequent acquisition of initial drive power and transient oscillation signals. In one implementation, the hardware interface is a pre-reserved switch input port on the humidifier's main control board, connected to the detection element on the water tank base. A permanent magnet is embedded in the bottom of the water tank, and a Hall element is installed at the corresponding position on the base. The Hall element outputs a level related to the polarity of the sensed magnetic field based on the Hall effect: when the water tank is properly seated, the magnet faces the Hall element's sensing surface with a specific polarity, and the output remains low; when the user quickly inverts the water tank and then reattaches it to the base, the magnet's orientation relative to the sensing surface flips, the polarity of the magnetic field sensed by the Hall element changes accordingly, and the output jumps from low to high.

[0022] Specifically, the main control board continuously polls the output of the Hall element, and when it detects a rising edge that changes from low to high, it determines that the water tank is in the installed position after rapid inversion, thereby generating a water tank inversion signal.

[0023] It should be noted that the atomizing plate's acquisition timing is activated only when the rising edge occurs and the humidifier subsequently receives a start command. The start point of the acquisition timing is aligned with the instant the atomizing plate drive circuit is energized, and the end point is limited to the end of the initial energization interval, thereby obtaining the start-up stage trigger mark corresponding to the inverted signal.

[0024] Specifically, the current and voltage sample values ​​at the same sampling time are multiplied point by point to obtain the instantaneous power actually consumed by the atomizing plate at that time, denoted as p(t) = u(t) × i(t), where u(t) is the voltage sample value at time t, i(t) is the current sample value at time t, and p(t) is the initial driving power at time t. The initial driving power is marked point by point with the timestamp sequence corresponding to the sampling time sequence. The power values ​​at each time are arranged according to the timestamp to determine the power input curve of the initial driving power changing with the sampling time sequence. This power input curve depicts the contour of the change in electrical energy input from small to large until it stabilizes during the startup phase. Under the bubble retention condition, its contour will deviate from the smooth upward shape during normal atomization.

[0025] Preferably, the power input curve undergoes derectification ripple processing to remove the residual 100Hz and its low-order harmonic components from the power supply rectification stage, retaining only the power change trend related to the vibration of the atomizing plate itself. This processing targets the rectified residuals on the DC bus side, which are different from the power frequency notch filtering performed on the vibration signal channel in S101 for 50Hz AC coupling. They do not overlap in processing location, processing purpose, or the signals being processed. Furthermore, the sampling window of the signal channel is defined based on the derectified power input curve. The sampling window begins when the power value first jumps from the static background and exceeds 1.5 times the average background value. The end time is determined by the following quantitative criteria to identify the first stabilization point: The power change rate between adjacent sampling points is calculated in 0.5 ms increments. When the absolute value of the change rate of at least 20 consecutive sampling points is less than 2% of the peak power, and the average power fluctuation within that segment does not exceed ±3%, the starting point of that continuous segment is taken as the first stabilization point. This delineates the transition segment from oscillation to steady state, concentrating vibration pickup within the initial transition process. If a stabilization segment meeting the above conditions is not detected within 50 ms, a timeout mechanism is used, with 50 ms as the end point of the window, and resampling is triggered. In one embodiment, a piezoelectric vibration sensor is attached to the inner wall of the atomization chamber. When the high-frequency vibration of the atomizing plate is conducted back to the chamber wall via water, it generates a charge signal related to the vibration phase and amplitude, i.e., a transient oscillation signal.

[0026] Specifically, the transient oscillation signal is first filtered by the signal channel to remove power frequency interference and spurious components introduced by the power supply and peripheral circuits at 50 Hz and its harmonics. Then, it is aligned with the power input curve at the same timestamp according to the acquisition sequence, so that the vibration amplitude at each moment corresponds one-to-one with the initial drive power at that moment, thus obtaining the start-up transient oscillation signal that synchronously corresponds to the initial drive power during the start-up phase.

[0027] It is understandable that the above hardware acquisition method is not limited to small-capacity humidifiers in bedrooms and baby rooms, but is also applicable to large-capacity models in offices and medical care environments. Only the specifications of the magnet and the length of the sampling window need to be adjusted according to the size of the water tank and cavity. The power input curve and start-up transient oscillation signal obtained in this acquisition process provide synchronously corresponding raw data benchmarks for identifying bubble retention and air cushion formation.

[0028] Step S102: Perform frequency domain conversion on the start-up transient oscillation signal to obtain the frequency domain response distribution of the atomizing cavity, and determine the acoustic radiation pressure intensity in the atomizing cavity based on the initial driving power and the frequency domain response distribution of the atomizing cavity.

[0029] The transient oscillation signal at startup is acquired, and a Fast Fourier Transform (FFT) is used to perform frequency domain transformation on the signal segment by segment along the acquisition time sequence. The amplitude and phase at each frequency point are extracted, and the amplitudes at each frequency point are arranged from low to high to obtain the frequency domain response distribution of the atomizing cavity. The frequency domain response distribution of the atomizing cavity indicates the energy accumulation pattern at each frequency point after the high-frequency vibration of the atomizing plate is transmitted back through the water. The resonant frequency point of the atomizing cavity is determined based on the frequency point with the highest amplitude in the frequency domain response distribution. The spectral peak value of the frequency domain response distribution of the atomizing cavity in the neighborhood of the resonant frequency point is extracted. The amplitude of the spectral peak value is aligned with the initial driving power along the same time stamp to obtain the sound pressure mapping amount of the initial driving power converted into the vibration amplitude of the water at the resonant frequency point. Based on the square of the amplitude accumulated at each frequency point in the neighborhood of the resonant frequency point, the energy density carried by a unit volume of water in the atomizing cavity is determined. For the impedance interface formed by gas and liquid on the surface of the atomizing plate, the reflection ratio and transmission ratio of sound wave at the impedance interface are determined based on the energy density and the acoustic impedance difference on both sides of the impedance interface. The energy density is converted into radiation intensity pointing upward to the liquid layer according to the transmission ratio, and the acoustic radiation pressure intensity of the area directly above the atomizing plate is obtained.

[0030] When an ultrasonic humidifier is started immediately after water replenishment, air bubbles gather directly above the atomizing plate. The transient oscillation signal generated by the water vibration and transmitted back through the cavity wall exhibits distinct energy distributions at different frequencies. The following embodiments describe the frequency domain transformation of this oscillation signal and the determination of the acoustic radiation pressure intensity within the atomizing cavity. In one embodiment, the transient oscillation signal is a time-varying charge waveform picked up by a piezoelectric vibration sensor. Since it exists in the time domain, it is difficult to visually identify where the energy is concentrated. A Fast Fourier Transform is used to divide this waveform into several equal-length segments along the acquisition sequence. A frequency domain transformation is performed on each segment, decomposing the time-fluctuating waveform into a set of sinusoidal components at different frequencies. The amplitude and phase at each frequency are then extracted.

[0031] Specifically, the input to the Fast Fourier Transform (FFT) is a time-domain sampling sequence, and the output is the amplitude and phase corresponding to each frequency point. The amplitude indicates the strength of the sinusoidal component at that frequency point, and the phase indicates its offset relative to the starting point. The amplitudes of each frequency point are arranged from low to high to obtain the frequency domain response distribution of the atomizing cavity. This frequency domain response distribution, with frequency points on the x-axis and amplitude on the y-axis, describes the energy accumulation pattern in the spectrum after the high-frequency vibration of the atomizing plate is transmitted back through the water.

[0032] It should be noted that the atomizing plate is driven to vibrate at a fixed high frequency. The vibration structure formed by the water and the cavity responds most strongly at a certain frequency, where the amplitude is the highest in the frequency domain response distribution of the atomizing cavity. The resonant frequency of the atomizing cavity is determined based on the frequency with the highest amplitude, reflecting the location where the water vibration energy is most concentrated. Under bubble retention conditions, the bubble clusters on the surface of the atomizing plate change the vibration conditions of the water. The amplitude distribution near the resonant frequency shifts and broadens compared to normal atomization. This morphological difference carries the original information for subsequent identification of the air cushion. For the resonant frequency, the spectral peak value of the frequency domain response distribution of the atomizing cavity in the neighborhood of that frequency is extracted. The neighborhood is a narrow band formed by several adjacent frequency points on both sides of the resonant frequency, and the spectral peak value is the set of amplitudes within this narrow band. The amplitude of the spectral peak value is aligned with the initial driving power at the same timestamp to obtain the sound pressure mapping amount of the initial driving power converted into the water vibration amplitude at the resonant frequency.

[0033] Specifically, the sound pressure mapping quantity characterizes the equivalent sound pressure response corresponding to water vibration at the resonant frequency under unit driving conditions. At the same timestamp, the initial driving power is denoted as p, in watts; the voltage amplitude at the peak of the spectrum at the resonant frequency is denoted as u, in volts; the electroacoustic calibration coefficient of the piezoelectric sensor is denoted as k, in Pascals per volt, and its value is obtained from factory calibration or comparison in a standard sound field, with a commonly used range of 50 to 500 Pascals per volt; the sound pressure mapping quantity... U vib Let be the voltage amplitude at the peak of the spectrum at the resonant frequency, P be the initial driving power, and represent the equivalent sound pressure amplitude obtained by taking the square root of each unit power. Since the radiated sound pressure is proportional to the square root of the input electrical power under linear acoustic conditions, is used... When an air cushion layer blocks sound energy transmission, the equivalent sound pressure on the water side decreases at the same power, and the sound pressure mapping decreases accordingly. Under normal atomization conditions, s is in the range of 0.8 to 3.0 Pa per watt. Once it falls below 0.4 Pa per watt, it is considered that an air cushion layer is present. In one embodiment, the squares of the sound pressure amplitudes at each frequency point in the neighborhood of the resonant frequency are summed and then divided by the volume of the water participating in the vibration to obtain the sound energy density carried by a unit volume of water in the atomization cavity, in joules per cubic meter, which is used to characterize the amount of sound energy carried by a unit volume of water in that neighborhood. When an air cushion layer is present, this energy is difficult to transfer to the upper liquid layer, and the energy density is higher in the region below the atomizing plate. Furthermore, the bubble group forms a boundary between the gas and liquid on the surface of the atomizing plate, which is the impedance interface. The ease with which sound waves propagate in different media is characterized by acoustic impedance, which is the product of the medium density and the speed of sound. It is understandable that the acoustic impedance of a gas is much smaller than that of a liquid. When a sound wave reaches the impedance interface from the liquid side, due to the significant difference in acoustic impedance between the two sides, most of the sound energy is reflected back into the water, while a small portion passes through the interface and enters the gas side. Let the acoustic impedance on the liquid side be Z1, and the acoustic impedance on the gas side be Z2. For water at normal temperature and pressure, Z1 is taken as approximately 1.5 × 10⁻⁶. 6 Pascals per second per meter, Z2 of air is approximately 4.3 × 10⁻⁶. 2 Pascals per second per meter. Based on the energy density and the acoustic impedance difference across the impedance interface, determine the reflection ratio R and transmission ratio T of the sound wave at the interface, where... T = 1 - R. Substituting the typical values ​​above, we get R is approximately 0.9989 and T is approximately 0.0011, meaning that about 99.89% of the incident sound energy is reflected back to the liquid, with only about 0.11% penetrating to the gas side. When changes in liquid surface temperature, gas content, or salinity cause Z1 to exceed 1.4 × 10⁻⁶, the effect is significant. 6 Up to 1.6×10 6 When the sound energy density fluctuates within the range of Pascals per meter, R and T are updated in real time according to the above formula, and then multiplied by the energy density to obtain the sound energy density reflected back to the water and permeated into the gas, which can be used for subsequent sound field allocation calculations.

[0034] For example, the energy density is converted into radiation intensity pointing upwards towards the liquid layer based on the transmission ratio, that is, the portion of sound energy that actually crosses the bubble boundary and is transmitted upwards. Then, it is converted into the force per unit area based on the sound-receiving area directly above the atomizing plate to obtain the sound radiation pressure intensity in the region directly above the atomizing plate. The more complete the air cushion layer coverage, the lower the transmission ratio, the smaller the radiation intensity beyond the boundary, and the sound radiation pressure intensity directly above the atomizing plate deviates from the value during normal atomization.

[0035] It should be noted that the above frequency domain conversion and acoustic radiation pressure calculation method is not limited to small-capacity humidifiers in bedrooms and baby rooms; it is also applicable to large-capacity models in offices and medical care environments, adjusting only the narrowband width of the adjacent domain according to the water volume and the position of the resonant frequency. The obtained acoustic radiation pressure intensity characterizes the actual force of the sound energy directly above the atomizing plate during the start-up phase, providing a quantitative basis for identifying the degree of bubble compression.

[0036] Step S103: Threshold evaluation of acoustic radiation pressure intensity is performed to screen out the frequency range in which the acoustic radiation pressure intensity exceeds the critical value for bubble compression initiation. The high-frequency waste component dissipated by bubble compression within the frequency range is determined according to the frequency domain response distribution of the atomizing cavity. The degree of bubble compression on the surface of the atomizing plate is identified based on the high-frequency waste component.

[0037] The acoustic radiation pressure intensity is obtained, and compared with a pre-established bubble compression initiation threshold value at each frequency point. The bubble compression initiation threshold value indicates the lower limit at which the acoustic radiation pressure presses the bubble against the surface of the atomizing plate, making it difficult for it to detach and float. If the acoustic radiation pressure intensity at a certain frequency point exceeds the bubble compression initiation threshold value, that frequency point is retained. Adjacent frequency points that continuously exceed the threshold value are merged into a frequency range to obtain the frequency range in which the acoustic radiation pressure fixes the bubble compression. For the frequency range, the amplitude values ​​of each frequency point falling into the range are extracted from the frequency domain response distribution of the atomizing cavity. The reference amplitude value in the same range during normal water atomization is subtracted from the current amplitude at each frequency point to obtain the amplitude difference of the bubble that is blocked by the gas interface and does not transform into droplets during the bubble compression process. The amplitude difference is accumulated at each frequency point to determine the high-frequency waste component dissipated by bubble compression in the frequency range. The high-frequency waste component indicates the proportion of sound energy lost during bubble compression due to idling. Based on the proportion of the high-frequency waste component within the frequency range, and combined with the spectral line broadening caused by bubble compression to diffuse the amplitude of the resonant frequency neighborhood to both sides, the degree to which the surface of the atomizing plate is covered by bubbles is determined. If the proportion of the high-frequency waste component increases and the spectral line broadening increases, the bubble compression on the surface of the atomizing plate intensifies, thus obtaining the degree of bubble compression on the surface of the atomizing plate.

[0038] Under bubble retention conditions, the acoustic radiation pressure intensity in the area directly above the atomizing plate varies at different frequencies. At some frequencies, the acoustic energy fails to break up the water into mist, instead suppressing the bubbles. The following implementation describes a threshold assessment of this acoustic radiation pressure intensity and the identification of the degree of bubble compression on the atomizing plate surface. In one implementation, a bubble compression initiation critical value is pre-established. This critical value indicates a lower limit of force: below this value, the bubbles detach and float due to buoyancy; above this value, the bubbles are pressed against the atomizing plate surface by the acoustic radiation pressure and are difficult to detach. This critical value is pre-calibrated based on water viscosity and bubble size and stored in the main control board. During calibration, the critical acoustic radiation pressure P0 = k × (u / r) × C_ is obtained by fitting a force balance relationship. scale+c, where C_ scale The proportionality constant is obtained through calibration and refitting. Under the same conditions, when the radius is increased to 100 micrometers, P0 decreases to 18 Pa. The calibration process is as follows: the atomizing plate is immersed in the water to be tested in a transparent water tank. A high-speed camera is used to aim at a single bubble on the surface of the atomizing plate. The driving voltage is gradually increased in multiple levels to make the acoustic radiation pressure directly above linearly sweep from low to high. Each level is stabilized for 2 seconds, and the actual acoustic radiation pressure is converted by a hydrophone. The critical driving amplitude at which the bubble flips from floating to sticking and stationary is recorded, and the acoustic radiation pressure corresponding to that level is taken as P0. The water temperature is changed to cover the range of 5 to 40 degrees Celsius to obtain different u values. Different microbubble generators with different apertures are used to obtain different r values. Each u and r combination is repeated 3 to 5 times and the average value is taken. Finally, two-dimensional lookup table data is generated with a u step size of 0.1 mPascals per second and a r step size of 10 micrometers and written into the main control board storage area. During operation, the corresponding P0 is retrieved by looking up the table based on the actual water temperature measured by the temperature sensor and the estimated dominant bubble size in the spectrum.

[0039] Specifically, the acoustic radiation pressure intensity in the region directly above the atomizing plate is compared one by one along the frequency points with the critical value for initiating bubble compression. If the acoustic radiation pressure intensity at a certain frequency point is higher than the critical value, that frequency point is retained; if it is lower than the critical value, it is discarded. Adjacent frequency points that consecutively exceed the critical value are merged into a segment to obtain the frequency range in which the acoustic radiation pressure compresses and fixes the bubble. The frequency range indicates at which frequencies the sound energy is redirected to suppress the bubble rather than atomize it.

[0040] It should be noted that the acoustic radiation pressure intensity at the same frequency point differs significantly between normal atomization and bubble retention conditions. Therefore, the position and width of the frequency range vary with the degree of bubble aggregation. For the aforementioned frequency range, the amplitude of each frequency point falling within this range is extracted from the frequency domain response distribution of the atomization cavity to obtain the amplitude distribution of this range under the current operating condition. The frequency domain response distribution of the atomization cavity, with frequency point as the abscissa and amplitude as the ordinate, indicates the aggregation pattern of water vibration energy in the frequency spectrum.

[0041] Specifically, the frequency response distribution is pre-collected under conditions of stable atomization and no bubble accumulation in the water. The amplitude of each frequency point within the same frequency range is extracted and recorded one by one as a reference amplitude. The reference amplitude indicates the vibration intensity that should exist in that range when all sound energy is converted into droplets. The reference amplitude is subtracted from the current amplitude at each frequency point to obtain the amplitude difference at that frequency point that is blocked by the gas interface and not converted into droplets. The heavier the bubble compression and the stronger the gas interface blockage, the larger the amplitude difference. To maintain consistency with the definition in S102 that vibration energy is proportional to the square of the amplitude, the amplitude difference at each frequency point is first squared, and then accumulated along all frequency points within the frequency range to obtain the total amount of sound energy lost during idling at bubble compression, denoted as the high-frequency idling component E, i.e., E=Σ(A0-A). 2Where A0 is the reference amplitude at this frequency point, and A is the current amplitude at this frequency point, the summation range covers all frequency points within the selected frequency interval of 20 kHz to 40 kHz. The high-frequency idle component indicates the proportion of sound energy lost due to idling at the bubble compression point without performing atomization work; the larger the value, the more sound energy is intercepted by the bubble. For example, if the reference spectrum has a cumulative energy of 100 energy units in this interval, and the current spectrum has an E of 25 energy units calculated using the above formula on the same frequency point set, then it is determined that approximately 25% of the sound energy is dissipated due to idling at the bubble compression point, triggering subsequent bubble removal or frequency adjustment actions.

[0042] Preferably, the amplitude difference at each frequency point within the aforementioned frequency range is squared point by point and then summed, denoted as e. The proportion of the high-frequency waste component in this range is obtained as r = E / Σ(A) using the sum of the squared amplitudes at each frequency point within the range as the denominator. 2 ), where r is a dimensionless quantity indicating the proportion of idling loss sound energy in the total sound energy of the interval. The higher the proportion, the greater the proportion of sound energy trapped by the bubble. Furthermore, bubble compression also causes the amplitude of the resonant frequency neighborhood to spread to both sides. The energy originally concentrated at the resonant frequency is squeezed by the bubble vibration and distributed to a wider frequency band. This diffusion width is the spectral broadening amplitude, which is taken as the frequency distance l corresponding to the point where the amplitude on both sides of the resonant frequency drops to half of the peak value. The wider the bubble coverage, the larger l is. To eliminate the dimensional influence of l, it is first normalized by the nominal half-width l0 of the resonant frequency of this model, resulting in the relative broadening b = l ÷ l0, so that b and r are on the same dimensionless scale. When determining the gear position in S103, the two factors are combined using a weighted sum to form a comprehensive cavitation index k = 0.6 × r + 0.4 × b. Here, r, which directly reflects the energy interception ratio, has a higher weight, while b serves as supplementary evidence. A value less than 0.15 indicates slight cavitation, between 0.15 and 0.35 indicates moderate cavitation, and greater than 0.35 indicates severe cavitation. If there is an inconsistency between r and b (one increasing and the other decreasing), the value of k is used, with an additional single-item over-limit check. That is, if r is greater than 0.4 or b is greater than 2.0, the gear position is directly increased by one level to ensure the closure and enforceability of the determination rule under multi-variable conditions.

[0043] Understandably, the degree to which the surface of the atomizing plate is covered by bubbles is determined based on the proportion *r* of the high-frequency waste component and the spectral broadening amplitude *Δf*. Here, *r* is the ratio of the power of the high-frequency waste component to the total driving power, and *Δf* is the increment of the baseline full width at half maximum (FWHM) relative to the main resonance peak under clean operating conditions, in kilohertz. The baseline FWHM is collected and stored during factory self-testing or each power-on no-load steady-state phase as a reference for subsequent comparisons. In actual judgment, *r* and *Δf* are substituted into preset grading intervals. When r is less than 0.05 and Δf is less than 0.3, it is judged as no compression, and there are basically no bubbles attached to the surface of the atomizing plate. When r is between 0.05 and 0.12 and Δf is between 0.3 and 0.8, it is judged as light compression, corresponding to scattered small bubbles. When r is between 0.12 and 0.20 and Δf is between 0.8 and 1.5, it is judged as medium compression, and the bubbles have covered part of the vibration area. When r is greater than 0.20 or Δf is greater than 1.5, it is judged as heavy compression, and the surface of the atomizing plate is compressed by a large area of ​​air film, and the atomization efficiency is significantly reduced. If r and Δf fall into different levels, the heavier level is merged into the output to avoid misjudgment caused by fluctuation of a single indicator. At the same time, the level is officially confirmed to be consistent within 3 consecutive sampling cycles to suppress transient disturbances. This gives the degree of bubble compression on the surface of the atomizing plate, and triggers subsequent control actions such as pulse backflushing, frequency reduction drive, or shutdown prompt.

[0044] For example, a low proportion and small spectral broadening correspond to bubbles only locally attached; a higher proportion and significantly increased broadening correspond to bubbles forming sheets pressed against the surface of the atomizing plate. This degree of bubble compression characterizes the degree to which bubbles on the surface of the atomizing plate are fixed by acoustic radiation pressure, providing a quantitative basis for identifying the coverage morphology of the surface air cushion.

[0045] It should be noted that the above threshold assessment and bubble compression degree identification methods are not limited to small-capacity humidifiers in bedrooms and baby rooms. They are also applicable to large-capacity models in offices and medical care environments. The calibration value of the bubble compression start-up threshold is adjusted only according to the power of the atomizing plate and the water volume.

[0046] Step S104: Identify the coverage pattern of the air cushion on the surface of the atomizing plate based on the degree of bubble compression; extract the thickness of the surface air cushion based on the coverage pattern of the air cushion on the surface of the atomizing plate; and calculate the sound energy blocking rate for sound energy transmission to the upper liquid layer based on the thickness of the surface air cushion.

[0047] The degree of bubble compression is obtained. Based on the adhesion distribution of the bubble compression on the surface of the atomizing plate, the edge contour of the air cushion is delineated. The boundary of the transition from the densely compressed area to the sparsely attached area of ​​the air cushion is determined by the edge contour. The coverage shape of the air cushion is determined by the proportion of the area defined by the edge contour to the sound-receiving surface of the atomizing plate. The coverage shape determines whether the air cushion is locally patchy or continuously sealed on the surface of the atomizing plate. The transmission path of sound energy through the air cushion is delineated based on the coverage shape. The sound path length during which the amplitude of the sound energy drops to half of its original value when passing through the air cushion along the transmission path is extracted. This sound path length is converted into the longitudinal distance of the gas filling along the normal direction of the air cushion, and the surface air cushion thickness is extracted. The transmission margin of the sound energy after passing through the air cushion along the transmission path is determined based on the surface air cushion thickness. The proportion of sound energy intercepted by the air cushion is obtained by subtracting the transmission margin from the sound energy pointing upwards to the liquid layer before passing through the air cushion. This proportion is divided by the sound energy before passing through the air cushion to obtain the sound energy blocking rate of sound energy transmission to the liquid layer above.

[0048] After bubbles are fixed to the surface of the atomizing plate by acoustic radiation pressure, their aggregation intensity is characterized by the degree of bubble compression. The greater the bubble compression, the more bubbles are pressed into the same area to form a layer, creating a gas interlayer between the atomizing plate and the water. The following embodiments describe how the coverage pattern of the gas cushion is identified by the degree of bubble compression, and how the acoustic energy blocking rate for sound energy transmission to the upper liquid layer is deduced from this. In one embodiment, the degree of bubble compression is not uniform along the sound-receiving surface of the atomizing plate; the compression is heavier at the center of the resonant region and lighter at the outer edges. The sound-receiving surface of the atomizing plate is divided into several small grids according to the resonant region, and the degree of bubble compression is read grid by grid.

[0049] Specifically, the air bubbles in adjacent small squares are arranged from most compressed to least compressed. The inflection points where the compression abruptly drops from dense compression to sparse attachment are identified. The small squares at these inflection points are connected to delineate the edge contour of the air cushion. This edge contour defines the boundary between the densely compressed and sparsely attached areas of the air cushion; the area within the contour is a densely packed area with continuous gas, while the area outside the contour is a sparsely attached area with scattered air bubbles. Further, the coverage pattern of the air cushion is determined based on the proportion of the area defined by the edge contour to the total sound-receiving surface area of ​​the atomizing plate. This coverage pattern determines the spreading morphology of the air cushion on the surface of the atomizing plate.

[0050] It should be noted that if the defined area has a low specific gravity and a scattered outline, the air cushion will appear as localized patches; if the specific gravity is high and the outline is continuous, the air cushion will appear as a continuous, sheet-like seal. These two types of cover patterns correspond to different path conditions where sound energy is obstructed to varying degrees. The sound energy transmission path from the surface of the atomizing sheet through the air cushion towards the upper liquid layer is defined based on the cover pattern. In the patchy attachment pattern, sound energy can still bypass the gaps between the patches and travel upwards; in the continuous seal pattern, sound energy must penetrate the entire gas interlayer head-on.

[0051] Specifically, the amplitude of sound energy attenuates along the path after entering the gas-filled area, and the absorption of sound energy by the gas causes the amplitude to gradually decrease with the distance traveled. The distance traveled by the sound energy amplitude from the point of entry into the air cushion to half of its original value is taken along the transmission path, i.e., the half-value sound path length. The thicker the gas filling, the longer the sound energy requires to cross the same amplitude reduction. This half-value sound path length is projected along the normal of the air cushion and converted into the longitudinal distance of the gas filling to extract the thickness of the surface air cushion. The surface air cushion thickness indicates the thickness of the gas interlayer on the surface of the atomizing sheet in the vertical direction. The greater the thickness, the deeper the sound energy is absorbed in the gas. This thickness extraction avoids the difficulty of directly measuring the inside of the air cushion, and infers the longitudinal dimension of the gas interlayer by the half-value attenuation of the sound energy itself.

[0052] Preferably, under a continuous sealing morphology, the half-value sound path length is taken along multiple normal directions, and the larger one is selected to match the thickest part of the air cushion's obstruction of sound energy. Further, the transmission margin remaining after sound energy passes through the air cushion is determined based on the surface air cushion thickness. Sound energy attenuates exponentially in gas, and less remains after passing through a thicker air cushion. The sound energy pointing upwards towards the liquid layer before passing through is denoted as w, and the transmission margin remaining after passing through is denoted as wt. Both vary with the air cushion thickness; the thicker the air cushion, the smaller wt is relative to w.

[0053] Understandably, subtracting the acoustic energy *w* pointing upwards towards the liquid layer before passing through the air cushion from the transmission margin *wt* yields the portion that was intercepted and not transmitted upwards. Dividing this portion by the acoustic energy before passing through the air cushion gives the acoustic energy blocking rate (*b*) for transmitting sound energy to the liquid layer above, where *b* = (*w* - *wt*) ÷ *w*, and *w* is the acoustic energy pointing upwards towards the liquid layer before passing through the air cushion, *wt* is the transmission margin remaining after passing through the air cushion, and *b* is the acoustic energy blocking rate. The closer the blocking rate is to 1, the more likely the air cushion is intercepting almost all the acoustic energy, with the liquid layer above receiving almost no additional acoustic energy.

[0054] For example, in the patchy attachment morphology, the air cushion is thin and has a large transmission margin, resulting in a low sound energy blocking rate, allowing the fog column to barely rise; in the continuous sealing morphology, the air cushion is thick and has a small transmission margin, resulting in a high sound energy blocking rate, with sound energy almost completely blocked on the surface of the atomizing sheet. This sound energy blocking rate characterizes the degree to which the air cushion layer obstructs the transmission of sound energy, providing a quantitative basis for subsequently measuring the loss of atomization caused by the air cushion.

[0055] It should be noted that the above-mentioned methods for identifying coverage patterns and estimating sound energy blocking rates are not limited to small-capacity humidifiers in bedrooms and baby rooms. They are also applicable to large-capacity models in offices and medical care environments, with the density of the small grids on the sound-receiving surface adjusted only according to the size of the atomizing plate's sound-receiving surface.

[0056] Step S105: Identify the portion of the sound energy blocking rate that exceeds the preset blocking threshold, extract the target driving power when the atomizing sheet generates a blocking effect exceeding the threshold, and determine the degree of atomization loss caused by the air cushion layer based on the target driving power and the thickness of the surface air cushion.

[0057] The acoustic energy blocking rate is obtained, and then compared point by point along the acquisition time sequence with a preset blocking threshold. The preset blocking threshold indicates the upper limit at which the sound energy transmission is blocked but the fog column can still barely rise. Continuous points where the acoustic energy blocking rate exceeds the preset blocking threshold are merged into an over-threshold segment. The target driving power applied to the atomizing plate by the atomizing plate driving circuit within the over-threshold segment is extracted. The target driving power indicates the amount of electrical energy input required to lift the air cushion when the air cushion obstruction is increased. Based on the difference between the target driving power and the driving power consumed to maintain an equal amount of fog column during normal atomization, the power redundancy consumed by the atomizing plate to lift the air cushion is obtained. Based on the power redundancy and the surface air cushion thickness, the proportion of fog droplets that fall when the air cushion thickness increases is obtained, thus determining the degree of atomization loss caused by the air cushion layer.

[0058] The degree to which the air cushion layer obstructs the transmission of sound energy is characterized by the sound energy blocking rate. When the sound energy blocking rate is too high, most of the sound energy is trapped by the air cushion on the surface of the atomizing plate, making it difficult for the mist column to rise. The following implementation describes the threshold evaluation of the sound energy blocking rate and the resulting loss in atomization caused by the air cushion layer. In one implementation, a preset blocking threshold is established in advance. The preset blocking threshold indicates an upper limit: when the sound energy blocking rate is below this value, the sound energy passing through the air cushion is still sufficient to lift the mist column, and the atomizing plate can still barely produce mist; when it is above this value, the sound energy passing through the air cushion is insufficient, and the mist column is intermittent or even stops. This threshold is pre-calibrated according to the power level of the atomizing plate and stored in the main control board.

[0059] Specifically, the acoustic energy blocking rate, obtained point by point along the time sequence, is compared with the preset blocking threshold along the time axis. If the acoustic energy blocking rate is higher than the threshold at a certain moment, the point is marked as an over-threshold point; if it is lower than the threshold, it is discarded. Time-continuous over-threshold points are merged into a segment to obtain an over-threshold segment. The over-threshold segment indicates the time interval during which the air cushion's blocking is intensified and the fog is pressurized. Further, the current and voltage sampling values ​​applied to the atomizing plate within the over-threshold segment are read from the atomizing plate drive circuit, multiplied point by point, and their average value within the segment is obtained to extract the target drive power of the atomizing plate within the over-threshold segment. The target drive power indicates the amount of electrical energy input to the drive circuit to push open the air cushion when the air cushion's blocking is intensified. The thicker the air cushion, the more electrical energy is invested to push open the air cushion, and the target drive power increases accordingly.

[0060] It should be noted that the target driving power only indicates the amount of electrical energy currently invested and cannot be directly converted into the amount of fog that is not generated. It must be compared with the power consumed to maintain an equal amount of fog column during normal atomization.

[0061] Specifically, a baseline driving power is pre-calibrated for the atomizing plate to maintain a rated mist column when there is no air cushion obstruction. Subtracting this baseline driving power from the target driving power yields the additional power redundancy consumed by the atomizing plate to open the air cushion, denoted as d, where d = pm - pb, where pm is the target driving power within the threshold range, pb is the baseline driving power consumed to maintain an equal mist column during normal atomization, and d is the power redundancy. Power redundancy indicates the additional electrical energy the atomizing plate is forced to invest under the same mist column target due to air cushion obstruction. A larger power redundancy indicates that the air cushion consumes more electrical energy to open the air bubbles rather than to break up the water.

[0062] In one embodiment, the amount of fog that falls is estimated based on the power redundancy and the surface air cushion thickness. The surface air cushion thickness indicates the thickness of the gas interlayer; the greater the thickness, the more acoustic energy is absorbed by the air cushion and does not reach the water surface under the same power redundancy.

[0063] Specifically, the surface air cushion thickness h is converted into energy transmittance η using the sound energy attenuation coefficient α. The calculation formula is η = e^(-αh), where α ranges from 0.15 to 0.25 millimeters, and h is in millimeters. Multiplying the power redundancy d by the transmittance loss fraction 1 minus η yields the sound power loss value Ploss = d × (1 - η) absorbed by the air cushion layer. Then, based on the atomization efficiency β of the atomizing plate, defined as the droplet mass generated per watt of sound power per second (typically 0.8 to 1.2 milligrams per watt-second), multiplying the sound power loss value Ploss by the atomization efficiency β and the action time t yields the amount of atomized mist Mloss = Ploss × β × t. This mist loss value represents the degree of atomization loss caused by the air cushion layer. For example, when the power redundancy is 5 watts, the air cushion thickness is 3 mm, the attenuation coefficient is 0.2 per mm, the atomization efficiency is 1.0 mg / W / s, and the action time is 10 seconds, the transmittance η is approximately 0.55, the power loss is 2.25 watts, and the amount of fog falling into the air is 22.5 mg.

[0064] For example, a short overthreshold segment, low power redundancy, and thin air cushion thickness result in only a slight reduction in fog volume; conversely, a long overthreshold segment, high power redundancy, and thick air cushion thickness result in a significant drop in fog volume or even complete interruption of flow. This degree of fog volume loss provides a quantitative basis for subsequently classifying the severity level of air cushion layer retention.

[0065] It should be noted that the above threshold assessment and atomization loss estimation methods are not limited to small-capacity humidifiers in bedrooms and baby rooms. They are also applicable to large-capacity models in offices and medical care environments. The preset blocking threshold and the calibrated value of the reference drive power are adjusted according to the power level of the atomizing plate.

[0066] Step S106: Based on the degree of atomization loss and bubble compression, the severity level of air cushion layer retention is determined. The severity level covers three levels: mild local bubble aggregation, moderate air cushion sheet coverage, and severe air cushion complete sealing.

[0067] The degree of atomization loss and the degree of bubble compression are obtained. The degree of atomization loss is divided into small, medium and large segments from small to large. The degree of bubble compression is divided into light, medium and heavy segments from light to heavy. The segments into which the two fall are combined in pairs to obtain the combined values ​​that characterize the retention state of the air cushion layer. The gear to which each combined value belongs is determined according to the pre-established judgment boundary between the segments and the gear. Based on the combined values ​​falling within the range defined by the judgment boundary, the severity level of air cushion layer retention is determined. If the atomization loss is in the small range and the bubble compression is in the light range, it is classified as a light bubble local aggregation state; if the atomization loss is in the medium range and the bubble compression is in the medium range, it is classified as a moderate air cushion sheet coverage state; if the atomization loss is in the large range and the bubble compression is in the heavy range, it is classified as a heavy air cushion full coverage state, thus obtaining the severity level of air cushion layer retention covering three levels.

[0068] The drag on atomization caused by the air cushion layer has been characterized by two separate indicators: the degree of atomization loss indicates the severity of the atomization volume reduction caused by the air cushion, and the degree of bubble compression indicates the degree of bubble coverage on the surface of the atomizing plate. The following implementation describes the combined filing of these two indicators and the determination of the severity level of air cushion layer retention. In one implementation, the degree of atomization loss and the degree of bubble compression each express one aspect of air cushion retention, and judging based on only one indicator is prone to bias. For example, if the bubbles are densely attached but the thickness is still thin, the atomization loss may not be significant; therefore, the two indicators are used for juxtaposition.

[0069] Specifically, the degree of atomization loss is divided into three segments—small, medium, and large—according to an increasing value. The small segment corresponds to a slight reduction in mist volume, the medium segment to a significant reduction, and the large segment to a substantial drop in mist volume or even complete interruption of flow. The boundary values ​​between each segment are pre-calibrated based on the rated mist output of the atomizing plate and stored in the main control board. Similarly, the degree of bubble compression is divided into three segments—light, medium, and heavy—according to an increasing value. The light segment corresponds to scattered bubbles, the medium segment to compressed bubbles, and the heavy segment to bubbles covering the entire surface of the atomizing plate. The boundary values ​​for each segment are pre-calibrated based on the sound-receiving surface of the atomizing plate. Furthermore, the segments into which the degree of atomization loss falls and the segments into which the degree of bubble compression falls are paired to obtain combined values ​​characterizing the retention state of the air cushion layer. The combined values ​​are arranged with one segment as the row and the other segment as the column, falling into a cell in the two-dimensional segment table. Each cell indicates a combination of the two indicators with varying degrees of importance.

[0070] It should be noted that a determination boundary is pre-established between the segments and gear positions. This determination boundary divides the two-dimensional segment table into three connected regions: the region with the smaller diagonal portion is assigned to light gear, the central portion to medium gear, and the region with the larger diagonal portion to heavy gear. The pairs at the boundaries are merged into adjacent gears according to the nearest principle. This determination boundary is calibrated once and stored on the main control board, and can be directly retrieved during subsequent comparisons.

[0071] In one embodiment, the severity level of air cushion layer retention is determined by the combination of values ​​falling within the range defined by the determination boundary. If the atomization loss is in the small range and the bubble compression is in the light range, the combination of values ​​falls into the light range, belonging to the light bubble local aggregation state, where the corresponding bubbles are only sporadically attached to the surface of the atomizing plate, and atomization can still be barely maintained.

[0072] Specifically, if both the atomization loss and bubble compression are in the medium range, the combined value falls into the medium range, classifying it as a moderate air cushion-like coverage state. This corresponds to bubbles pressing against the surface of the atomizing sheet, resulting in a significant reduction in atomization volume. If both the atomization loss and bubble compression are in the heavy range, the combined value falls into the heavy range, classifying it as a severe air cushion complete coverage state. This corresponds to the air cushion completely covering the surface of the atomizing sheet, with sound energy almost completely blocked, and atomization nearly stagnant. This yields the severity levels of air cushion layer retention at three levels: light, medium, and heavy.

[0073] It is understandable that for the cross-combination of two indicators located in different hot and cold zones, such as fog loss in the medium zone and bubble compression in the heavy zone, the adjacent gear should be merged according to the principle of proximity of the judgment boundary to avoid omission and misjudgment.

[0074] It should be noted that the above-mentioned segmentation and severity level classification method is not limited to small-capacity humidifiers in bedrooms and baby rooms; it also applies to large-capacity models in offices and medical care environments. The boundary values ​​for each segment are adjusted only based on the rated mist output of the atomizing plate and the size of the sound-receiving surface. The resulting severity level depicts the level of air cushion layer retention from light to heavy, providing a quantitative basis for subsequent graded warning outputs.

[0075] Step S107: Output graded warning signals for the startup phase according to the severity level. The graded warning signals correspond to the humidifier's atomization startup status and drive power compensation requirements.

[0076] The severity level of the air cushion layer retention is obtained. Based on the pre-established mapping between severity levels and warning code values, a warning code value is assigned to each of the following states: mild localized bubble aggregation, moderate sheet-like air cushion coverage, and severe complete air cushion coverage, resulting in a graded warning signal for the startup phase. The atomization startup status indicator and drive power compensation requirement are retrieved based on the warning code value carried by the graded warning signal. The atomization startup status indicator is marked as normal startup, delayed startup, or startup obstruction. The drive power compensation requirement is marked as the compensation level from low to high. The corresponding graded warning signal indicating the humidifier's atomization startup status and drive power compensation requirement is output.

[0077] The severity levels of air cushion layer retention have been categorized into three levels: mild localized bubble aggregation, moderate sheet-like air cushion coverage, and severe complete air cushion coverage. The following implementation describes the output of graded warning signals for the initiation phase based on these severity levels. In one implementation, a mapping table between severity levels and warning code values ​​is pre-stored on the main control board. Mild localized bubble aggregation corresponds to code value 1, moderate sheet-like air cushion coverage corresponds to code value 2, and severe complete air cushion coverage corresponds to code value 3. After reading the severity level of air cushion layer retention, the corresponding warning code value is retrieved according to the mapping table to obtain the graded warning signal for the initiation phase. The graded warning signal consists of code values ​​from level 1 to level 3, which vary with the severity level; a higher value indicates greater air cushion obstruction.

[0078] Specifically, the warning code value carried by the graded warning signal is associated with a lookup table, which contains two preset items for each code value: atomization start-up status indication and drive power compensation requirement. Code value 1 indicates normal start-up and low compensation level, and the atomizing plate can produce mist by maintaining the original drive power; code value 2 indicates delayed start-up and medium compensation level, indicating that the drive power should be appropriately increased for the air cushion-like coverage; code value 3 indicates hindered start-up and high compensation level, indicating that the drive power should be significantly increased or even the mist production should be temporarily suspended for complete air cushion coverage.

[0079] It should be noted that the above mapping and retrieval methods are not limited to small-capacity humidifiers in bedrooms and baby rooms; they are also applicable to large-capacity models in offices and medical care environments, only requiring adjustments to the compensation values ​​corresponding to each code value based on the model's power level. After retrieving the carried code value, a graded warning signal is output, indicating the humidifier's atomization start-up status and drive power compensation requirements, for retrieval during the start-up phase.

[0080] It should be noted that the above examples are merely some specific embodiments of the present invention. Obviously, the present invention is not limited to the above embodiments and many variations are possible. All variations that can be directly derived or conceived by those skilled in the art from the content disclosed in this invention should be considered within the scope of protection of this invention.

Claims

1. A method for intelligent identification of abnormal states in a humidifier, characterized in that, The method includes: The water tank inversion signal is obtained through the hardware interface, and the initial driving power and transient oscillation signal of the atomizing plate during the start-up phase are collected based on the water tank inversion signal. The transient oscillation signal at startup is converted into the frequency domain to obtain the frequency domain response distribution of the atomizing cavity. The acoustic radiation pressure intensity in the atomizing cavity is determined based on the initial driving power and the frequency domain response distribution of the atomizing cavity. Threshold evaluation of acoustic radiation pressure intensity is performed to screen out the frequency range where the acoustic radiation pressure intensity exceeds the critical value for bubble compression initiation. The high-frequency loss component dissipated by bubble compression within the frequency range is determined based on the frequency domain response distribution of the atomizing cavity. The degree of bubble compression on the surface of the atomizing plate is identified based on the high-frequency loss component. The coverage pattern of the air cushion on the surface of the atomizing plate is identified based on the degree of bubble compression. Based on the coverage pattern of the air cushion on the surface of the atomizing plate, the thickness of the surface air cushion is extracted. Based on the thickness of the surface air cushion, the sound energy blocking rate for sound energy transmission to the upper liquid layer is calculated. Identify the portion of the sound energy blocking rate that exceeds the preset blocking threshold, extract the target driving power when the atomizing sheet generates over-threshold blocking, and determine the degree of atomization loss caused by the air cushion layer based on the target driving power and the surface air cushion thickness. Based on the degree of atomization loss and bubble compression, the severity level of air cushion layer retention is classified, including mild local bubble aggregation, moderate air cushion sheet coverage, and severe air cushion complete coverage. Based on the severity level, a graded warning signal is output for the startup phase. The graded warning signal corresponds to the humidifier's atomization startup status and drive power compensation requirements.

2. The intelligent identification method for abnormal states of a humidifier according to claim 1, characterized in that, The process of acquiring the water tank inversion signal via a hardware interface, and collecting the initial drive power and transient oscillation signal of the atomizing plate during the startup phase based on the water tank inversion signal, includes: The inverted signal output by the Hall element at the bottom of the water tank is received through the hardware interface. When the output level of the Hall element changes from low to high, the acquisition timing of the atomizing plate is activated to obtain the start-up phase trigger mark. The current and voltage sampling values ​​of the atomizing plate driving circuit are retrieved according to the trigger flag of the startup phase, and multiplied point by point in the initial power-on interval to obtain the initial driving power and power input curve. The transient oscillation signal is picked up by a piezoelectric vibration sensor. After filtering out power frequency interference, it is aligned with the power input curve according to the acquisition timing to obtain the start-up transient oscillation signal.

3. The intelligent identification method for abnormal states of a humidifier according to claim 1, characterized in that, The step of performing frequency domain conversion on the start-up transient oscillation signal to obtain the frequency domain response distribution of the atomizing cavity, and determining the acoustic radiation pressure intensity within the atomizing cavity based on the initial driving power and the frequency domain response distribution of the atomizing cavity, includes: The fast Fourier transform is used to perform frequency domain transformation on the start-up transient oscillation signal segment by segment along the acquisition time sequence, extract the amplitude and phase of each frequency point, and arrange the frequency points from low to high to obtain the frequency domain response distribution of the atomizing cavity. The resonant frequency point of the atomizing cavity is determined for the highest amplitude frequency point. The peak values ​​of the spectrum in the neighborhood of the resonant frequency point are extracted, and the amplitude and the initial driving power are aligned along the same time stamp. The square of the amplitude is accumulated point by point to obtain the energy density of the water body. Based on the difference in acoustic impedance between the two sides of the energy density and impedance interface, the radiation intensity pointing upwards to the liquid layer is converted according to the transmission ratio, thus obtaining the acoustic radiation pressure intensity in the region directly above the atomizing plate.

4. The intelligent identification method for abnormal states of a humidifier according to claim 1, characterized in that, The threshold evaluation of acoustic radiation pressure intensity, screening out frequency ranges where the acoustic radiation pressure intensity exceeds the critical value for bubble compression initiation, determining the high-frequency waste component dissipated by bubble compression within the frequency range based on the frequency domain response distribution of the atomizing cavity, and identifying the degree of bubble compression on the surface of the atomizing plate based on the high-frequency waste component, includes: The acoustic radiation pressure intensity is compared with the bubble compression start-up critical value at each frequency point, and the frequency points that exceed the critical value are retained and the adjacent frequency points that exceed the critical value are merged into the frequency range. The amplitude values ​​of each frequency point falling within the frequency range are extracted from the frequency domain response distribution of the atomizing cavity. The reference amplitude value in the same range during normal atomization is subtracted from the current amplitude point by point, and the amplitude difference is accumulated according to the frequency point to obtain the high-frequency waste component. Based on the proportion of the high-frequency waste component in the frequency range, and combined with the spectral line broadening amplitude of the resonant frequency neighborhood amplitude spreading to both sides, the degree of bubble compression on the surface of the atomizing plate is obtained.

5. The intelligent identification method for abnormal states of a humidifier according to claim 1, characterized in that, The process of identifying the coverage pattern of the air cushion on the surface of the atomizing plate based on the degree of bubble compression, extracting the thickness of the surface air cushion based on the coverage pattern, and calculating the sound energy blocking rate for sound energy transmission to the upper liquid layer based on the surface air cushion thickness includes: The edge contour of the air cushion is delineated based on the adhesion distribution of the bubble compression degree on the surface of the atomizing plate. The boundary of the transition from the dense compression area to the sparse adhesion area of ​​the air cushion is determined by the edge contour. The coverage shape of the air cushion is determined based on the proportion of the area enclosed by the edge contour to the sound-receiving surface of the atomizing plate. The coverage shape is determined to show whether the air cushion is locally patchy or continuously covered on the surface of the atomizing plate. The sound energy transmission path through the air cushion is defined according to the coverage shape. The sound energy path length along the transmission path is extracted and converted into the longitudinal distance of gas filling along the normal of the air cushion to obtain the surface air cushion thickness. The transmission margin of sound energy after passing through the air cushion along the transmission path is determined based on the thickness of the surface air cushion. The proportion of sound energy intercepted by the air cushion is obtained by subtracting the transmission margin from the sound energy pointing upwards to the liquid layer before passing through the air cushion. The sound energy blocking rate of sound energy transmission to the liquid layer upwards is obtained by dividing the sound energy before passing through the air cushion.

6. The intelligent identification method for abnormal states of a humidifier according to claim 1, characterized in that, The portion of the sound energy blocking rate exceeding a preset blocking threshold is identified, the target driving power at which the atomizing sheet generates over-threshold blocking is extracted, and the degree of atomization loss caused by the air cushion layer is determined based on the target driving power and the surface air cushion thickness, including: The acoustic energy blocking rate is compared with a preset blocking threshold point by point along the acquisition time sequence, and consecutive points where the acoustic energy blocking rate exceeds the preset blocking threshold are merged into an over-threshold segment; Extract the target driving power applied to the atomizing plate by the atomizing plate driving circuit within the over-threshold range. The target driving power indicates the amount of electrical energy input that is raised when the air cushion is pushed open when the air cushion barrier is increased. Based on the difference between the target driving power and the driving power consumed to maintain an equal amount of mist column during normal atomization, the extra power redundancy consumed by the atomizing plate to open the air cushion is obtained. Based on the power redundancy and the surface air cushion thickness, the proportion of droplets that fall out when the air cushion thickness increases is obtained, and the degree of atomization loss caused by the air cushion layer is determined.

7. The intelligent identification method for abnormal states of a humidifier according to claim 1, characterized in that, The severity level of air cushion layer retention is determined based on the degree of atomization loss and bubble compression. This severity level covers three grades: mild localized bubble aggregation, moderate sheet-like air cushion coverage, and severe complete air cushion coverage. The degree of atomization loss is divided into small, medium and large segments from small to large, and the degree of bubble compression is divided into light, medium and heavy segments from light to heavy. The segments into which the two fall are combined in pairs to obtain the combined values ​​that characterize the retention state of the air cushion layer. Based on the pre-established judgment boundaries between sections and levels, the combination of falling into a small section and a light section is classified as a light bubble local aggregation state, the combination of falling into a medium section and a medium section is classified as a moderate air cushion sheet coverage state, and the combination of falling into a large section and a heavy section is classified as a severe air cushion full coverage state, thus obtaining the severity level of air cushion layer retention.

8. The intelligent identification method for abnormal states of a humidifier according to claim 1, characterized in that, The step of outputting a graded warning signal for the startup phase based on the severity level, wherein the graded warning signal corresponds to indicating the humidifier's atomization startup status and drive power compensation requirements, including: Based on the pre-established mapping between severity levels and warning code values, a warning code value is assigned to each of the following states: mild localized bubble aggregation, moderate air cushion sheet coverage, and severe air cushion complete coverage, thus obtaining a graded warning signal for the initiation phase. Based on the warning code value carried by the graded warning signal, the atomization start-up status indication and drive power compensation requirement are retrieved. The atomization start-up status indication indicates normal start-up, delayed start-up, or start-up obstruction. The drive power compensation requirement indicates the compensation level from low to high. The graded warning signal corresponding to the humidifier atomization start-up status and drive power compensation requirement is output.

Images