A method for fault detection in a carbon dioxide fractional laser therapy device

Abstract

This invention relates to the field of medical device fault diagnosis technology and discloses a fault detection method for a carbon dioxide fractional laser therapy instrument. The method establishes reference driving parameters and reference electrical impedance characteristics during the calibration phase; during detection, it simultaneously acquires scanning coordinates, laser voltage and current, and real-time driving signals. The method first calculates the laser's dynamic impedance based on voltage and current. If the impedance deviates from the reference characteristics, the fault source is determined to be the laser. If the impedance is normal, a compensation intensity scalar is calculated based on the difference between the real-time driving signal and the reference parameters, and the emitted test pulse train is controlled to calculate the power modulation slope. Finally, based on the spatial distribution of the compensation intensity scalar with the scanning radius and the power modulation slope, the fault type is determined to be a geometric alignment fault, a thermal contamination fault, or a global optical path attenuation. This invention can effectively distinguish between electrical system faults and optical system faults and accurately identify the specific attributes of optical path faults.

Description

Technical Field

[0001] This invention relates to the field of medical device fault diagnosis technology, and in particular to a fault detection method for a carbon dioxide fractional laser therapy device. Background Technology

[0002] Fractional carbon dioxide laser therapy devices utilize a galvanometer scanning system to control the laser beam to form a micropore array on the target tissue, and are widely used in skin reconstruction and surgical procedures. During long-term operation, these devices inevitably experience output energy attenuation or beam quality degradation due to limitations in the lifespan of core components and the operating environment. To ensure clinical accuracy of treatment dosage, existing device control systems are typically equipped with a power closed-loop control mechanism based on photoelectric feedback. When the terminal output energy is detected to be lower than a preset standard, the control algorithm automatically increases the laser power supply's drive duty cycle or power command to compensate for the energy loss.

[0003] However, while this compensatory mechanism, which relies solely on end-effector energy feedback, can maintain short-term equipment operation, it masks the root cause of energy decay to some extent. When the control system significantly increases the drive intensity to maintain standard output, existing detection logic struggles to accurately identify whether the fault originates from electrical aging of the laser itself or from physical losses in the external optical transmission path. Laser faults typically involve abnormal radio frequency discharge or gas aging, manifesting as an overall decrease in electro-optical conversion efficiency; while optical path faults involve the transmission efficiency of reflectors, light guides, and focusing lenses. The required repair strategies and costs for these two types of faults are drastically different, and confusion regarding the fault source can easily lead to the incorrect replacement of high-value components.

[0004] Furthermore, current technologies lack sophisticated fault diagnosis capabilities for optical transmission systems. The main causes of optical path loss include geometric alignment deviations due to mechanical wear of the light guide arm joints, and thermal lensing effects caused by contaminants adhering to the optical lens surfaces. Both types of faults may manifest as energy attenuation at the edge of the scanning field of view in static testing at a single moment, but their physical mechanisms are fundamentally different. Mechanical deviations are static spatial position errors, typically not fluctuating with light emission time; while thermal contamination is accompanied by heat accumulation due to laser energy absorption, exhibiting significant time-varying characteristics. Current airborne inspection procedures often lack joint analysis methods combining time-domain thermal response and spatial distribution characteristics, making it impossible to distinguish between mechanical faults and optical contamination without disassembling the equipment. This usually requires specialized engineers to use external instruments for segment-by-segment troubleshooting, increasing equipment maintenance difficulty and downtime. Summary of the Invention

[0005] The problem solved by this invention is that the prior art has difficulty in distinguishing between laser body faults and optical path transmission faults, and it is difficult to locate the specific physical properties of optical path faults.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] The first aspect of this invention provides a fault detection method for a carbon dioxide fractional laser therapy device. This method operates within the control system of the carbon dioxide fractional laser therapy device and includes the following main steps:

[0008] During the calibration phase, the scanning handpiece is controlled to perform scanning actions, the reference drive parameters required to maintain standard output are recorded, and the reference electrical impedance characteristics are established.

[0009] During the testing process, the real-time scanning coordinates of the scanning handpiece, the voltage and current data of the laser, and the real-time drive signal are collected simultaneously.

[0010] Subsequently, the dynamic impedance of the laser is calculated based on the voltage and current data. If the deviation between the dynamic impedance and the reference electrical impedance characteristics exceeds a preset threshold, the fault source is determined to be the laser. If the deviation does not exceed the preset threshold, it indicates that the electrical system is normal. Then, the compensation intensity scalar is calculated based on the difference between the real-time drive signal and the reference drive parameters, and the laser is controlled to emit a test pulse train to calculate the power modulation slope.

[0011] Finally, based on the spatial distribution of the compensation intensity scalar with the scanning radius and the power modulation slope, the fault type is determined to be geometric alignment fault, thermal contamination fault, or global optical path attenuation.

[0012] Furthermore, to establish an accurate reference, a discretized mesh strategy is employed in the process of recording the reference driving parameters. A mesh containing multiple radius nodes from the center of the scanning field to the maximum scanning radius is defined. For each radius node, the driving signal of the laser power supply is adjusted using a PID algorithm until the actual output energy stabilizes at the standard target value. The system records the power command and duty cycle command at this time as reference driving parameters, constructing a radius-driving parameter reference lookup table. Before calculating the compensation intensity scalar in real time, the real-time radius is calculated using the real-time scanning coordinates, and the theoretical reference driving parameters corresponding to the current position are calculated in the reference lookup table using an interpolation model, thereby eliminating the inherent optical path transmission differences under different scanning radii.

[0013] To ensure temporal consistency of the data and accurately reflect the laser load characteristics, a hardware synchronization mechanism is employed during data acquisition. A hardware synchronization trigger source based on the PWM cycle is configured, and the delay time between the sampling moment and the rising edge of the PWM signal is set. This delay time is configured to be greater than the rise time of the laser power supply output current, ensuring that the sampling point falls within the flat-top steady-state region of the current waveform. When the sampling moment is reached, the analog-to-digital converter, galvanometer position interface, and optical path feedback unit are triggered in parallel to generate a synchronized data frame containing a timestamp at the same moment, real-time scan coordinates, voltage and current data, and real-time drive signals.

[0014] To screen for laser-related faults, this invention employs dynamic impedance analysis. A time-sliding window containing multiple sets of historical voltage and current values ​​is constructed. The least squares method is used to perform linear regression on the voltage and current data within the window, and the regression slope is extracted as the dynamic impedance. The deviation of the dynamic impedance from the reference electrical impedance characteristics is calculated. If the deviation exceeds a preset percentage of the reference impedance mean, the fault source is determined to be the laser. The process of establishing the reference electrical impedance characteristics involves continuously acquiring voltage and current data for multiple cycles at different scanning positions, calculating the average impedance and standard deviation, and setting the normal fluctuation range of the reference impedance based on the 3σ statistical principle.

[0015] In the optical path fault analysis phase, to quantify the degree of compensation for optical path loss by the closed-loop control system, a compensation intensity scalar is calculated. Specifically, the difference between the real-time drive signal and the theoretical reference drive parameters is calculated to obtain a residual vector containing power compensation residuals and duty cycle compensation residuals. Using weighting coefficients determined based on the system's open-loop gain, a weighted Euclidean norm is calculated on the power compensation residuals and duty cycle compensation residuals to obtain a unified compensation intensity scalar.

[0016] To distinguish between optical faults caused by thermal effects and geometric faults caused by mechanical structures, this invention introduces the power modulation slope as a time-domain feature. The scanning handpiece is positioned at the scanning radius corresponding to the detected maximum compensation intensity scalar, and the laser emits a test pulse sequence containing multiple continuous pulses. The energy value of each pulse in the sequence is acquired, constructing a pulse train energy attenuation sequence. Linear regression analysis is then performed on this sequence, and the absolute value of the slope of the regression line is calculated as the power modulation slope.

[0017] This invention determines the fault type by jointly analyzing spatial distribution characteristics and temporal thermal response characteristics. The Pearson correlation coefficient between the compensation intensity scalar and the scanning radius is calculated. If the Pearson correlation coefficient is less than a preset correlation threshold, and the mean value of the compensation intensity scalar exceeds the normal range, it indicates that the loss is spatially uniformly distributed, and is determined to be global optical path attenuation. If the Pearson correlation coefficient is greater than or equal to the preset correlation threshold, it indicates that the loss is concentrated at the edge of the field of view. In this case, the power modulation slope is further considered: if the power modulation slope is greater than a preset thermal response threshold, it indicates the presence of a thermal lensing effect that accumulates over time, and is determined to be a thermal contamination fault in the optical lens; if the power modulation slope is less than or equal to the preset thermal response threshold, it indicates that the loss does not change significantly over time, and is determined to be a geometric alignment fault caused by mechanical deflection in the light guide arm.

[0018] Furthermore, this invention can further locate specific faulty components. The scanning field is divided into four quadrants, and the average compensation intensity in each quadrant is calculated. The asymmetry factor of the difference between the maximum and minimum average compensation intensity relative to the overall average value is calculated. If the asymmetry factor is less than a preset symmetry threshold, it indicates that the edge attenuation is centrally symmetrically distributed, and the faulty component is determined to be an F-theta lens with rotational symmetry; if the asymmetry factor is greater than or equal to the preset symmetry threshold, it indicates that the attenuation is directional, and the direction of the quadrant with the largest average compensation intensity is located, thus locating the faulty component as the galvanometer axis in the corresponding direction.

[0019] A second aspect of the present invention provides a carbon dioxide fractional laser therapy device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method described in the first aspect above.

[0020] This invention achieves hierarchical decoupling of fault sources by constructing a multidimensional feature space that includes electrical impedance, spatial distribution, and thermal response. It utilizes dynamic impedance to eliminate laser-based faults, spatial correlation to distinguish between global and local attenuation, and thermal response slope to differentiate between optical contamination and mechanical deflection. Ultimately, it achieves accurate diagnosis and location of fault mechanisms in the optical path system, effectively improving equipment maintenance efficiency and safety.

[0021] In summary, the present invention has at least one of the following beneficial technical effects:

[0022] 1. This invention calculates the dynamic impedance by collecting voltage and current data from the laser and compares it with the characteristics of a reference electrical impedance, effectively decoupling electrical faults in the laser itself from external optical path transmission faults. Existing technologies typically rely solely on the attenuation of the terminal output energy to determine the system status, making it difficult to identify whether energy loss stems from abnormal laser discharge, gas aging, or optical path component obstruction. This method introduces electrical impedance analysis at the front end of the detection process, directly assessing the laser's load characteristics. Optical path analysis is then performed only after ruling out excitation source faults, avoiding ineffective repairs due to incorrect fault source identification.

[0023] 2. This invention combines spatial distribution patterns with temporal thermal response characteristics to solve the technical challenge of distinguishing between thermal contamination of optical lenses and mechanical deflection of the light guide arm. By utilizing the correlation between the compensation intensity scalar and the scanning radius, non-uniform losses at the edge of the field of view can be identified. Furthermore, the power modulation slope of the test pulse train can quantify the thermal absorption effect of the optical component. Since the geometric loss caused by mechanical deflection usually does not change with the pulse duration, while the thermal lensing effect generated by contaminant absorption of laser light causes the loss to dynamically increase with heat accumulation, this method utilizes this difference in physical characteristics to accurately determine the nature of the fault, thereby guiding maintenance personnel to perform targeted lens cleaning or mechanical calibration.

[0024] 3. This invention employs a hardware synchronous triggering mechanism based on the PWM cycle, ensuring the timing consistency and accuracy of multi-dimensional detection data. By configuring delay parameters, the sampling action occurs within the flat-top steady-state region of the laser power supply output current, avoiding transient fluctuation interference from the current rising edge and improving the signal-to-noise ratio of impedance calculation. Simultaneously, the parallel triggering mechanism guarantees a strict correspondence between the galvanometer position coordinates, drive signals, and electrical parameters at the same microsecond level, eliminating spatial and temporal errors caused by system communication delays or asynchronous sampling, and providing reliable data support for the quantitative calculation of weak optical path losses. Attached Figure Description

[0025] Figure 1 This is a system hardware structure block diagram of the carbon dioxide fractional laser therapy device provided in the embodiments of the present invention;

[0026] Figure 2 This is a flowchart illustrating the overall steps of the carbon dioxide fractional laser therapy device fault detection method provided in this embodiment of the invention.

[0027] Figure 3 This is a flowchart of step S100 in this embodiment of the invention, which describes the construction of the baseline state space.

[0028] Figure 4 This is a flowchart illustrating the specific process of collecting real-time multi-source data in step S200 of this embodiment of the invention.

[0029] Figure 5 This is a flowchart of step S300 in this embodiment of the invention, which calculates dynamic impedance and screens for source faults.

[0030] Figure 6 This is a flowchart illustrating the specific steps of step S400 in this embodiment of the invention, which involves performing scanning field uniformity detection.

[0031] Figure 7 This is a flowchart illustrating the specific steps of step S500 in this embodiment of the invention for extracting the thermal response features of the pulse train.

[0032] Figure 8 This is a flowchart of step S600 in this embodiment of the invention for determining the fault type and locating the fault.

[0033] Figure 9 This is a schematic diagram comparing the dynamic impedance characteristics (VI trajectory) of the laser in normal state and gas aging state in an embodiment of the present invention;

[0034] Figure 10 This is a thermal map showing the spatial distribution of the scanning field compensation residual when the X-axis galvanometer of the scanning hand has a mechanical deflection fault in an embodiment of the present invention.

[0035] Figure 11This is a comparison chart of the temporal evolution trend of pulse train energy under three states in the embodiments of the present invention: normal optical path, mechanical blockage fault, and thermal contamination fault.

[0036] Among them, 100 is the main control unit; 200 is the laser power supply; 210 is the sampling circuit; 300 is the laser; 400 is the light guide arm; 410 is the scanning handpiece; and 500 is the feedback unit. Detailed Implementation

[0037] To enable those skilled in the art to better understand the technical solutions in the embodiments of this application, and to make the above-mentioned objectives, features and advantages of the embodiments of this application more apparent and understandable, the technical solutions in the embodiments of this application will be further described in detail below with reference to the accompanying drawings.

[0038] See attached document Figure 1 This invention provides a fault detection method for a carbon dioxide fractional laser therapy device, which operates in the control system of the carbon dioxide fractional laser therapy device.

[0039] The carbon dioxide fractional laser therapy device mainly includes a main control unit 100, a laser power supply 200, a laser 300, a light guide arm 400, a scanning handpiece 410 located at the end of the light guide arm 400, and a feedback unit 500. The main control unit 100 is connected to the laser power supply 200, the scanning handpiece 410, and the feedback unit 500.

[0040] The main control unit 100 is equipped with a processor for executing control algorithms and fault detection logic. The main control unit 100 sends digital signals containing power commands and duty cycle commands to the laser power supply 200.

[0041] The laser power supply 200 outputs excitation energy to the laser 300 based on the received digital signal. The laser power supply 200 integrates a sampling circuit 210. The sampling circuit 210 is configured to collect the voltage signal and loop current signal at both ends of the laser 300, and convert the collected signals into digital signals to be sent to the main control unit 100.

[0042] Laser 300 generates a laser beam after receiving excitation energy. The laser beam is coupled into light guide arm 400. Light guide arm 400 is composed of multiple arm segments connected in series via rotary joints and is used to transmit the laser beam.

[0043] The scanning handpiece 410 integrates a galvanometer motor and a scanning mirror, configured to receive scanning commands from the main control unit 100 and deflect the laser beam to form a preset scanning pattern in the target area. The main control unit 100 can acquire the real-time scanning coordinates of the scanning handpiece 410 or send position control signals to it.

[0044] The feedback unit 500 is located at the end of the light guide arm 400 or at the light outlet of the scanning handpiece 410. The feedback unit 500 includes a detector for monitoring the output laser energy intensity and sending the energy signal to the main control unit 100.

[0045] The main control unit 100 compares the energy signal from the feedback unit 500 with the target energy value, adjusts the digital signal sent to the laser power supply 200, and generates real-time compensation parameters.

[0046] See attached document Figure 2 This invention provides a method for fault detection in a carbon dioxide fractional laser therapy device, comprising the following steps:

[0047] S100, construct the reference state space; during the device initialization or calibration phase, the main control unit 100 controls the scanning handpiece 410 to perform full-range scanning in the scanning field, records the reference driving parameters and reference impedance characteristics required to maintain the standard output, and establishes a reference database of scanning field flatness under different scanning radii.

[0048] S200 collects real-time multi-source data; in the fault detection process, the main control unit 100 synchronously collects the real-time scanning coordinates of the scanning handpiece 410, the voltage and current data output by the sampling circuit 210, and the real-time drive signal generated by the control loop.

[0049] S300, calculate dynamic impedance and screen for source faults; the main control unit 100 calculates the dynamic impedance of the laser 300 based on voltage and current data, and compares the dynamic impedance with the reference impedance characteristics; if the deviation exceeds the impedance threshold, the fault source is determined to be the laser 300, an electrical fault signal is generated and the subsequent steps are terminated; if the deviation is within the allowable range, proceed to step S400.

[0050] S400 performs scanning field uniformity detection; the main control unit 100 controls the scanning handpiece 410 to perform a spiral scanning test that spreads from the center to the edge. Based on the difference between the real-time control input and the reference parameters, it calculates the compensation intensity scalar of the entire field and constructs a residual dataset describing the distribution law of loss with radius, which is used to characterize whether there is geometric aperture truncation in the optical path.

[0051] S500 extracts the thermal response characteristics of the pulse train; the main control unit 100 controls the scanning handpiece 410 to remain in the center zero position, controls the laser 300 to emit a fixed-point test pulse train, analyzes the time evolution trend of the real-time driving signal, calculates the power modulation slope in the pulse train, and quantifies the impact of the thermal lensing effect on the transmission efficiency.

[0052] S600 determines the fault type and locates the fault; the main control unit 100 jointly analyzes the edge attenuation coefficient and power modulation slope. If the edge attenuation coefficient is high but the power modulation slope is low, it is determined that the light guide arm 400 has a geometric alignment fault caused by mechanical deflection; if the power modulation slope is high, it is determined that the optical lens has a thermal contamination fault; if both are normal but the overall energy is low, it is determined to be a global optical path attenuation.

[0053] The following provides a detailed description of each step of the method of the present invention.

[0054] See attached document Figure 3 Step S100 specifically includes the following steps:

[0055] S101 defines the discretized grid of the scan field.

[0056] The main control unit 100 reads the optical parameters stored internally in the scanning handpiece 410, including the maximum scanning radius. .

[0057] Since the scanning handpiece 410 typically employs an F-theta flat lens, its optical path characteristics are theoretically rotationally symmetric. Mechanical deflection or aperture obstruction in the optical path system usually manifests as a nonlinear abrupt change in energy when the light spot reaches a specific radius. Therefore, this embodiment abandons the Cartesian coordinate system meshing and adopts a radial discretization strategy based on polar coordinates to reduce data dimensionality.

[0058] In this embodiment, a one-dimensional discrete variable is defined. It represents the radial distance between the center of the light spot and the center of the scanning field.

[0059] Set radial sampling step size , construct a containing A one-dimensional discrete set of nodes .in:

[0060] , representing the geometric center point of the scanned field;

[0061] , indicating the maximum scan radius;

[0062] , indicating the first The radius value of each sampling node.

[0063] Sampling step size The value is set based on the spot diameter. To ensure coverage, the value is usually set to 50% to 100% of the spot diameter, for example, 0.5mm to 1.0mm.

[0064] S102, acquires reference drive parameters.

[0065] During the factory calibration phase or when the equipment is confirmed to be in normal operating mode, the system establishes benchmark data for comparison. At this time, the light guide arm 400 is in a naturally extended state, and the optical path system is free from mechanical deformation or optical contamination.

[0066] During this process, the main control unit 100 controls the deflection of the galvanometer motor in the scanning handpiece 410, causing the laser beam to be projected sequentially onto the collection. Each radius node On the corresponding circular trajectory.

[0067] For each radius node The main control unit 100 performs the following closed-loop control operations:

[0068] 1. Control laser 300 to emit laser light and set the initial drive signal;

[0069] 2. Read the actual output energy detected by the feedback unit 500;

[0070] 3. Adjust the drive signal of the laser power supply 200 using a PID algorithm until the actual output energy stabilizes at the preset standard target value (e.g., 10 watts).

[0071] Once the output energy stabilizes, the system records the current driving parameters as the reference values ​​for that radius. Typically, due to the variation of lens transmittance with the incident angle (Fresnel loss) and vignetting effect, the reference driving parameters at the edge radius will be slightly higher than those at the center radius; this is an inherent optical characteristic of the system.

[0072] Recorded benchmark drive parameters Including reference power command and base duty cycle command .

[0073] S103, establish the reference electrical impedance characteristics.

[0074] While acquiring the reference driving parameters, the voltage of the laser circuit is also acquired simultaneously. With current .

[0075] The principle behind this step is that the electrical impedance of a laser (especially an RF-excited CO2 laser) is primarily affected by the gas state within the cavity, and is independent of external optical path transmission. By establishing an electrical reference, laser-related faults (such as gas leaks or RF power supply failures) can be decoupled from optical path transmission faults during subsequent fault detection.

[0076] In this embodiment, the reference gas impedance under this state is calculated. Specifically, the main control unit 100 continuously collects data at each sampling location. Voltage and current data for each cycle ( (Typically, values ​​are taken from 50 to 100), to calculate the average impedance. and impedance standard deviation .

[0077] Based on the 3σ statistical principle, the normal fluctuation range of the reference impedance is set. :

[0078] ;

[0079] in, The confidence coefficient is calculated based on electromagnetic interference in the field to avoid false alarms. The value range is set to 2.5 to 3.0.

[0080] S104, Construct the baseline lookup table.

[0081] The radius nodes obtained in the above steps Its corresponding reference drive parameters Each parameter is mapped one-to-one and stored in the non-volatile memory of the control system, forming a "radius-drive parameter" reference lookup table. This lookup table quantifies the inherent transmission efficiency distribution of the normal optical path system under different scanning radii.

[0082] S105, Configure the interpolation model for non-grid point states.

[0083] In actual fault detection, the scanning handpiece 410 can perform spiral scanning or random pattern scanning, which actually involves the radius of the light spot position. Often not equal to discrete sets The integer node values ​​in the space. To obtain an accurate reference at any location in continuous space, a numerical interpolation model needs to be constructed.

[0084] In this embodiment, a linear interpolation algorithm is used. For any given time... scan coordinates The main control unit 100 first calculates its real-time radius:

[0085] ;

[0086] Next, in the set Search and locate two adjacent grid nodes. and This makes the conditions satisfied. .

[0087] Finally, the theoretical baseline value is calculated based on the relative position of the current radius within the interval. :

[0088] ;

[0089] in:

[0090] To find the radius stored in the table Reference drive parameters at the location;

[0091] The interpolation weighting coefficients are defined as follows:

[0092] ;

[0093] This formula ensures the continuity of the reference data throughout the entire scanning field, enabling the system to calculate the theoretical energy required at any scanning radius, providing an accurate comparison benchmark for subsequent calculations.

[0094] See attached document Figure 4 Step S200 specifically includes the following steps:

[0095] S201 defines the system state vector and control input vector.

[0096] The operation of a carbon dioxide fractional laser therapy device is influenced by three dimensions: electrical drive, optical transmission, and mechanical motion. Specifically, the output power of the laser 300 depends on the current electrical load; while the energy density reaching the target surface is affected by the transmittance distribution of the flat-field lens in the scanning handpiece 410, which exhibits inherently nonlinear differences with the beam deflection position. Therefore, a coupled state model needs to be constructed.

[0097] In this embodiment, discrete time points are defined. Construct the system state vector as follows:

[0098] ;

[0099] in, This indicates that the galvanometer system in the scanning handpiece 410 is at time... The two-dimensional scanning coordinate vector corresponds to the geometric position of the laser beam on the working surface; The sampling circuit 210 collects the real-time load voltage across the laser 300. The sampling circuit 210 collects the real-time loop current flowing through the laser 300; The real-time laser energy intensity value collected by the feedback unit 500.

[0100] At the same time, define the control input vector. This indicates the drive command applied by the main control unit 100:

[0101] ;

[0102] in, This is the power command value. This is the duty cycle instruction value.

[0103] S202, configures a hardware synchronous trigger source based on the PWM cycle.

[0104] The principle of this step is based on the electrical response characteristics of radio frequency (RF) excited CO2 lasers. Since lasers are typically driven by high-frequency PWM modulation, their loop current is not instantaneously established. At the rising edge of the PWM signal, limited by the inductive effect of the RF power supply and the setup time of the impedance matching network, the current waveform exhibits a rising process followed by a brief oscillation period. If sampling is performed during this period, the obtained impedance data cannot accurately reflect the steady-state characteristics of the gas discharge.

[0105] In this embodiment, the main control unit 100 uses an internal timer to generate a PWM signal and uses the counter comparison match event as a global synchronization trigger signal. Set the sampling time. Delay time relative to the rising edge of the PWM . The calculation formula is:

[0106] ;

[0107] in, To represent the typical rise time of the output current of the laser power supply 200, this parameter is determined by the laser power supply hardware manual. For conventional 30W to 60W RF lasers, its value is usually between 10μs and 30μs. To ensure stability and avoid waveform overshoot, and to ensure that the sampling point falls within the current flat-top region, The value range is set to 5μs to 10μs.

[0108] By configuring a timer, the system ensures that only after the current reaches a steady state... Data collection is triggered at any time.

[0109] S203 performs parallel data latching and conversion.

[0110] When the global synchronization trigger signal Upon generation, the main control unit 100 triggers the following three modules in parallel through a hardware linkage mechanism to ensure the alignment of all state data on the timeline:

[0111] First, trigger the analog-to-digital converter (ADC): sample and hold the analog voltage and current signals and initiate the conversion to obtain... and ;

[0112] Second, trigger the galvanometer position interface: the main control unit 100 immediately reads and latches the real-time position data fed back by the galvanometer driver board (or latches the current DAC output command value) to obtain... ;

[0113] Third, trigger the optical path feedback unit 500: synchronously read the output level of the energy detector, and obtain... .

[0114] S204 uses Direct Memory Access (DMA) to transfer data and construct synchronous data frames.

[0115] To avoid timing jitter caused by CPU interrupt response delays, the DMA controller is configured to directly transfer the converted digital data to the memory buffer.

[0116] In this embodiment, when the buffer is ready, the main control unit 100 reads the data collected at the same trigger time and generates an index with a timestamp. Synchronization data frames :

[0117] ;

[0118] in, Indicates the first One PWM control cycle; These are the scan coordinates corresponding to that moment. This data frame forms the basis for subsequent calculations of dynamic impedance and spatial transmission efficiency.

[0119] See attached document Figure 5 Step S300 specifically includes the following steps:

[0120] S301, Construct a time sliding window for electrical characteristic calculation.

[0121] In order to extract the steady-state trend of the gas discharge inside the laser 300 from the sampled data containing high-frequency switching noise, the main control unit 100 processes the synchronization data frame sequence output in step S204. Perform digital filtering.

[0122] In this embodiment, a length of [length missing] is set in the memory. First-in-first-out buffer as a sliding window This window contains information from the current time. Backward Group voltage and current values:

[0123] ;

[0124] Window length The value of is determined based on the relationship between the ADC sampling frequency and the power supply ripple characteristics. It must satisfy the Nyquist sampling theorem and have sufficient smoothing capability. Its calculation formula is:

[0125] ;

[0126] in, The ADC sampling frequency is typically set to 100kHz to 500kHz. The main ripple frequency of the laser power supply output current is 200. For radio frequency power supplies, it usually corresponds to the PWM modulation frequency or its harmonics, ranging from 5kHz to 20kHz. This is a smoothing coefficient used to determine the number of ripple cycles covered; it is used to effectively suppress random noise and ensure the algorithm's dynamic response speed to impedance changes. The value range is set to 3 to 10, which covers 3 to 10 complete ripple cycles.

[0127] S302, Calculate the dynamic impedance spectrum.

[0128] Based on sliding window The data within is used to calculate the dynamic impedance of laser 300 under the current operating conditions. .

[0129] The physical principle behind this step is that radio frequency excited carbon dioxide lasers are nonlinear gas discharge loads. When chronic aging faults occur inside the laser (such as pressure changes due to CO2 gas decomposition, helium leakage, or electrode oxidation), the slope of its current-voltage characteristic curve (i.e., differential resistance) will drift significantly. Conversely, lens contamination or mechanical deflection within the guide arm 400 is an external fault and will not reversely change the discharge physical characteristics of the laser body. Therefore, dynamic impedance is a key decoupling characteristic that distinguishes between source faults and transmission faults.

[0130] In this embodiment, to eliminate the influence of DC operating point drift, the least squares method is used to perform linear regression on the voltage-current data within the window, and the regression slope is extracted as the dynamic impedance.

[0131] ;

[0132] in, and These are the current and voltage sample values ​​within the window, respectively. The formula calculates... Essentially, it is the differential impedance near the operating point. It can reflect changes in the gas discharge state more sensitively than the simple static resistance V / I.

[0133] S303, calculate impedance deviation and determine the source of the fault.

[0134] Read the reference impedance range stored in the reference lookup table in step S100. Calculate real-time dynamic impedance Deviation from the reference interval .

[0135] Deviation The calculation logic uses the absolute deviation method:

[0136] ;

[0137] This logic quantifies the degree to which the current electrical characteristics deviate from the normal baseline.

[0138] S304 executes fault decoupling logic.

[0139] First, set the impedance anomaly threshold. The physical meaning of this threshold corresponds to the tolerance limit of the RF power matching network. When the load impedance changes too much, causing the VSWR to increase, the power efficiency will drop sharply. According to industry experience and power supply manuals, the permissible impedance mismatch range is typically 15% to 25% of the nominal value.

[0140] Therefore, this embodiment is set as follows:

[0141] ;

[0142] in The average reference impedance obtained in step S103.

[0143] In this embodiment, the deviation is compared. With threshold :

[0144] like The system determines that there is a fault in the laser 300 itself (such as gas depletion or RF board failure). At this time, the main control unit 100 generates an electrical fault code and sets the source fault flag. The program logic directly jumps to the safety protection module, terminating subsequent optical path analysis steps. This avoids misjudgment of the optical path status caused by fluctuations in the output of the light source itself.

[0145] like The laser 300 is determined to be in a healthy state. At this point, if the feedback unit 500 detects the final output energy... If the energy is lower than the target value, the energy attenuation must be caused by the optical path transmission link (i.e., the light guide arm 400 or the scanning handpiece 410).

[0146] This step achieves precise decoupling of the fault source, providing a logical premise for the subsequent step S400 to analyze the uniformity of the scanning field and thermal effects.

[0147] See attached document Figure 6 Step S400 specifically includes the following steps:

[0148] S401, acquire real-time control input and theoretical reference input.

[0149] After the system determines that the laser 300 itself is fault-free, i.e., the impedance is normal in step S300, the main control unit 100 enters the optical path transmission efficiency evaluation process. The main control unit 100 reads the real-time control input vector at the current moment. This vector contains the current power command value. and duty cycle command value .

[0150] Meanwhile, the main control unit 100 scans the coordinates in real time. Calculate the scanning radius of the spot center relative to the optical axis center. :

[0151] ;

[0152] Using the interpolation model preset in step S105, Calculate the theoretical baseline control input corresponding to this position for the index. . This represents the standard drive parameters required to achieve the target output energy under ideal conditions where the 410F-theta lens of the scanning handpiece is uncontaminated and the optical path is free of mechanical deflection. This benchmark already incorporates the inherent attenuation characteristics of a normal optical system as the field of view changes.

[0153] S402 defines the compensation residual vector.

[0154] In a closed-loop energy control system, when the optical path transmission efficiency decreases due to a fault (such as a dirty lens or a misaligned optical path), the PID controller will automatically increase the control input (increase the power or duty cycle) to maintain a constant output energy at the end. Although this compensatory mechanism ensures the therapeutic effect, it also masks the early characteristics of optical path faults.

[0155] In this embodiment, to quantify the potential loss compensated by the closed-loop adjustment, a compensation residual vector is defined. It is defined as the difference between the real-time control input vector and the theoretical baseline control input vector:

[0156] ;

[0157] The unfolded form is as follows:

[0158] ;

[0159] in, For power compensation residuals; To compensate for residuals for duty cycle.

[0160] When the above residual is positive, it physically indicates that the system is consuming additional electrical energy to offset abnormal losses in the optical path.

[0161] S403, calculate the scalar value of the compensation strength.

[0162] Since power commands (usually in watts or digital quantization) and duty cycle commands (percentages) have different physical dimensions, direct comparison or superposition lacks physical meaning. Therefore, it is necessary to introduce weighting coefficients based on system sensitivity to normalize the multidimensional residuals into a unified scalar. .

[0163] In this embodiment, The calculation uses the weighted Euclidean norm:

[0164] ;

[0165] in, and These are the normalized weighting coefficients for power and duty cycle, respectively. These two coefficients are determined based on the system's open-loop gain (sensitivity) near its rated operating point, and are calculated as follows:

[0166] ;

[0167] in, and These are the partial derivatives of the output energy with respect to power and duty cycle, respectively. This indicates the value taken near the rated operating point.

[0168] Through this weighting process By unifying the control increments of different dimensions into a metric of equivalent energy loss, the additional regulation intensity applied by the closed-loop system to maintain steady-state output can be objectively reflected.

[0169] S404, construct the radius-residual space distribution model.

[0170] In this embodiment, the main control unit 100 calculates the compensation intensity scalar at each moment. With the corresponding scan radius Perform timestamp alignment and build associated data pairs. .

[0171] As the scanning handpiece 410 operates within the test pattern described in S100, the system continuously accumulates these data pairs, forming a dynamic dataset describing the distribution of loss with radius. .

[0172] This dataset establishes a mapping relationship between geometric spatial location and energy transmission loss, which forms the mathematical basis for subsequent fault classification.

[0173] 1. If The value varies If the increase in the residual is nonlinear and shows a significant increase (i.e., the edge residual is much larger than the center residual), it indicates that there is mechanical deflection or aperture blockage in the optical path, i.e., geometric alignment failure.

[0174] 2. If If the value is uniformly high across all radii, it indicates a global attenuation in the optical path, i.e., an optical contamination fault.

[0175] For data storage, the system uses a circular buffer to record the data of the most recent complete scan cycle in real time, which is then used by the subsequent step S600 for feature extraction and determination.

[0176] See attached document Figure 7 Step S500 specifically includes the following steps:

[0177] S501, configures the pulse train transmission mode and sampling window.

[0178] In this embodiment, when the compensation intensity scalar calculated in step S400 indicates that the system has abnormal energy loss, the system automatically enters the fixed-point thermal effect analysis mode. The purpose of this mode is to distinguish between thermal decay and geometric decay through time-domain analysis.

[0179] The main control unit 100 first controls the galvanometer in the scanning handpiece 410 to position the beam center at the location where the maximum residual value was detected in step S400. The corresponding scan radius Once the position is locked, control laser 300 to emit a set of standard test pulse sequences.

[0180] This sequence is composed of It consists of a series of short pulses, in which The value of needs to be sufficient to cover the thermal absorption response time of the optical lens (usually made of ZnSe material). In this embodiment, it is set to... It ranges from 50 to 100.

[0181] Set the single pulse width to The pulse interval is To prevent damage to the substrate due to excessive heat absorption by contaminants on the lens surface during the testing process, the single pulse energy is limited to 10% to 20% of the rated output energy.

[0182] Simultaneously, an optical path feedback unit 500 is configured for synchronous acquisition. To accurately capture the transient peak value of each pulse, the sampling frequency... Must meet:

[0183] ;

[0184] This ensures that each pulse width contains at least 10 sampling points, and the sampling window length is set to... The system continuously performs ADC conversions and stores data within the window.

[0185] S502, extract the pulse train energy decay sequence.

[0186] In this embodiment, the main control unit 100 performs peak search on the raw data within the sampling window and extracts the peak values. The peak voltage corresponding to each pulse is converted into an energy value, and a pulse train energy decay sequence is constructed. ,in Indicates the first The measured energy of each pulse.

[0187] This step is based on the physical principle of the difference between thermal lensing and geometric truncation:

[0188] 1. If organic contaminants (such as splattered tissue or oil) are present in the optical path, these substances have a high absorption rate for 10.6 μm wavelength laser light. Under the continuous action of the pulse sequence, the contaminants absorb light energy and convert it into heat energy, causing a local temperature rise and a change in refractive index gradient in the lens substrate. This change in refractive index will cause beam wavefront distortion (usually manifested as a shortening of the focal length or an increase in the divergence angle), resulting in a "slope-like" decrease in the energy density reaching the detector target surface over time.

[0189] 2. If the fault originates from geometric obstruction caused by loose mechanical components (such as the beam hitting the edge of the lens barrel), the proportion of the beam that is truncated depends primarily on the geometric location and does not change with heat accumulation over a short period. Therefore, the energy sequence This will manifest as a decrease in overall amplitude, but the relative amplitude between pulses remains constant, exhibiting a "step" response.

[0190] S503 defines the thermal modulation slope.

[0191] In this embodiment, a thermal modulation slope is defined to quantify the trend of energy sequence changes over time. .

[0192] The least squares method was used to analyze the energy decay sequence. Linear regression analysis was performed, and the fitted model was: ,in The pulse number. Energy value The intercept of the regression line. Thermal modulation slope. Take the slope of the regression line The absolute value of is calculated using the following formula:

[0193] ;

[0194] in, It is the sum of the product of the pulse number and the energy. It is the sum of the pulse numbers. The numerical value reflects the rate of thermal decay of energy per unit pulse interval (unit: mJ / pulse).

[0195] In addition, a thermal response determination threshold is set. This threshold was determined using statistical methods: during the equipment's factory calibration phase (clean optical path condition), data was collected... For each group (e.g., K=50) of pulse train data, calculate the natural thermal drift slope for each group and take the average value. and standard deviation .

[0196] The threshold is set as follows:

[0197] ;

[0198] Under typical operating conditions, The value range is typically from 0.01 mJ / pulse to 0.05 mJ / pulse.

[0199] S504, establish the mapping relationship between thermal response characteristics and fault types.

[0200] In this embodiment, the main control unit 100 will calculate the thermal modulation slope in real time. With threshold By comparing the results, the specific physical properties of the fault can be determined.

[0201] like The system determined that the energy sequence exhibited significant time-dependent decay. The fault type was identified as thermal contamination. This indicates the presence of highly absorptive contaminants on the surface of the optical components, causing a rapid decrease in transmission efficiency with heat accumulation. At this point, the main control unit 100 generated a maintenance instruction recommending cleaning the optical lenses.

[0202] like The system determines that the energy sequence does not exhibit significant time-dependent decay. Although its absolute energy value may be lower than the standard value (confirmed by step S400), the optical characteristics of the transmission channel remain stable for a short period. The system identifies the fault type as a geometric occlusion fault. This indicates that the fault is caused by a constant proportional loss due to optical path alignment misalignment or mechanical interference. At this point, the main control unit 100 generates a maintenance instruction suggesting calibration of the optical path mechanical structure.

[0203] See attached document Figure 8 Step S600 specifically includes the following steps:

[0204] S601 calculates spatial correlation to distinguish between global and local faults.

[0205] In this embodiment, the main control unit 100 retrieves the radius-residual dataset generated in step S404. ,in This represents the total number of sampling points.

[0206] To distinguish whether the fault source is located inside the scanning handpiece 410 or in the upstream optical path based on geometric distribution, the system analyzes the compensation intensity. With scan radius The changing trend.

[0207] This embodiment uses the Pearson correlation coefficient. To quantify the degree of linear correlation between the two, the calculation formula is as follows:

[0208] ;

[0209] in, and These are the arithmetic mean of the scan radius and the compensation intensity, respectively.

[0210] The physical basis for this step is:

[0211] 1. If the fault originates from the upstream optical path (such as laser aging or a dirty protective mirror at the light guide arm inlet), the resulting energy attenuation is usually spatially uniform (isotropic). In this case, The values ​​are relatively stable and do not change with the scan radius. The changes in these variables exhibit regular fluctuations, showing low correlation.

[0212] 2. If the fault originates within the scanning handpiece 410 (e.g., damage to the coating at the edge of the F-theta lens, interference from the mechanical limit of the galvanometer), its loss will exhibit a significant characteristic of low center and high edge. In this case, With radius The increase is significant, indicating a high positive correlation.

[0213] Set correlation threshold This threshold is determined based on the allowable edge decay rate of the system design, and is typically between 0.6 and 0.8. The determination logic is as follows:

[0214] like and If the value exceeds the normal range, the fault type is determined to be global inlet attenuation.

[0215] like The fault type was determined to be field-of-view edge attenuation.

[0216] S602, integrates spatiotemporal features to construct a fault determination logic matrix.

[0217] For cases identified as field-of-view edge attenuation, the main control unit 100 combines the thermal modulation slope calculated in step S500. (Time-domain characteristics) Correlation coefficient with step S601 (Spatial domain features), construct the decision matrix:

[0218] Scenario 1: If and It was determined to be thermal contamination caused by the field lens.

[0219] At this point, the fault characteristics are: spatially concentrated at the large-radius edges, and temporally accompanied by thermal accumulation and decay of energy. The physical mechanism indicates that there are absorptive contaminants on the surface of the F-theta lens, and due to the longer optical path at the edges, the thermal lensing effect is amplified at the edges.

[0220] Scenario 2: If and It was determined to be mechanical deflection of the galvanometer.

[0221] At this point, the fault characteristics are: spatially concentrated at the large radius edge, but temporally stable energy transmission. The physical mechanism indicates that the fault is caused by mirror zero-position drift or mechanical interference, resulting in the beam being physically truncated when deflected at large angles.

[0222] S603, based on spatial symmetry analysis, locates specific components.

[0223] After confirming the presence of field-of-view edge attenuation, in order to further distinguish whether the problem is a lens issue or a single-axis galvanometer issue, the system calculates the symmetry of the residual distribution.

[0224] The scanning field was divided into four quadrants (Q1 to Q4), and the average compensation intensity in each quadrant was calculated. Define the asymmetric factor. :

[0225] ;

[0226] The decision logic is as follows:

[0227] 1. If ( (A symmetry threshold, typically 0.2): This indicates that the edge attenuation is centrally symmetrically distributed, and the fault source is located in the F-theta lens. Because the lens is a rotationally symmetric optical element, its contamination usually leads to isotropic edge attenuation.

[0228] 2. If This indicates that the attenuation has a clear directionality. System search. The direction of the largest quadrant is used to locate the corresponding galvanometer axis. For example, if the residual in the positive X-axis region is significantly larger, it is located as the X-axis galvanometer.

[0229] S604 outputs the final diagnostic report and maintenance strategy.

[0230] Based on the analysis results of S601 to S603 above, the main control unit 100 generates a comprehensive diagnostic code that includes the nature of the fault, the fault mechanism, and the recommended maintenance location.

[0231] For example, the output code ERR-SCAN-LENS-THERM indicates that thermal contamination (THERM) exists in the lens of the scanning handpiece.

[0232] The output code ERR-SCAN-XMIRROR-MECH indicates that there is mechanical deflection (MECH) in the X-axis galvanometer (XMIRROR) of the scanning handpiece.

[0233] The system calls the preset maintenance strategy library based on the diagnostic results:

[0234] For global inlet attenuation, it is recommended to check the light guide arm inlet or the laser.

[0235] For thermal contamination of field lenses, it is recommended to clean the outer surface of the F-theta lens;

[0236] If the galvanometer is mechanically deflected, a prompt will appear indicating that the galvanometer geometry calibration procedure should be performed.

[0237] This step enables the diagnosis from vague energy deficiency to specific physical components (lens / X-axis / Y-axis), avoiding blindly disassembling the equipment.

[0238] To verify the effectiveness of the fault detection method for the carbon dioxide fractional laser therapy device proposed in this invention, an experimental platform was built, comprising a main control unit 100 (based on ARM Cortex-M7), a radio frequency excited CO2 laser (rated power 30W), a high-precision two-dimensional galvanometer scanning handpiece, and an end power meter. The experiment verified three typical fault scenarios: electrical aging, scanning mechanical deflection, and optical thermal contamination.

[0239] Experimental environment setup:

[0240] Benchmark group (traditional method): Only monitors the end output energy A general alarm is triggered when the energy level is below 85% of the target value.

[0241] Experimental group (method of this invention): Enables full-process diagnosis using dynamic impedance monitoring, scanning field residual analysis, and thermal response feature extraction.

[0242] Test objects: The focus is on monitoring the impedance characteristics of the RF power supply and the optical transmission characteristics of the scanning handpiece (including the X / Y galvanometer and the F-theta field lens).

[0243] Scenario 1: Laser gas aging fault detection

[0244] Experimental conditions: After the simulated laser has been used for 5000 hours, a small amount of helium gas leaks in the mixed gas in the discharge cavity, which causes changes in the plasma discharge characteristics, but the output energy has not decreased significantly at this time (it is still maintained at more than 95% of the rated value).

[0245] See the appendix for experimental results. Figure 9 :

[0246] Traditional method: due to output energy The system determined the equipment was normal as it did not fall below the 85% alarm threshold. This resulted in the laser operating under impedance mismatch for an extended period, and the increased reflected power accelerated the loss of the final stage power transistor in the RF power supply.

[0247] The method of this invention: displaying the dynamic trajectory of voltage-current (VI) acquired by the main control unit, and the dynamic impedance of the aging laser. A significant drift occurred. For example... Figure 9 As shown, the VI curve (solid line) under normal conditions exhibits a stable linear slope; while the curve (dashed line) under aging conditions has a larger slope, indicating an increase in differential resistance. The system calculates the impedance deviation. Exceeding the threshold It provides an early warning of electrical aging (gas failure) approximately 200 hours before significant energy decay, prompting users to perform refilling and maintenance.

[0248] Scenario 2: Location of mechanical deflection fault in scanning galvanometer

[0249] Experimental conditions: The mechanical limit of the X-axis galvanometer was manually fine-tuned to adjust its scanning radius corresponding to the deflection angle. Optical path obstruction occurs. The control system performs a full-field spiral scan test.

[0250] See the appendix for experimental results. Figure 10 :

[0251] Traditional methods detect a decrease in average energy during scanning, but cannot distinguish whether it is due to insufficient power supply or an optical path problem, nor can they pinpoint which axis galvanometer is faulty. Repairs typically require replacing the entire tooling set.

[0252] The method of this invention: The system records the compensation strength scalar. With scan radius The distribution relationship. For example... Figure 10 As shown in the scan field residual thermal map, the central region ( The light color indicates low residual, signifying normal transmission; while the edge region along the X-axis ( The color turns dark (representing high residuals), forming a clear asymmetric distribution.

[0253] Diagnostic output: The system calculates the radius correlation coefficient. And asymmetric factors Significantly, the system output a diagnostic report stating "Fault location: X-axis galvanometer of scanning handpiece; Type: Mechanical deflection," instructing maintenance personnel to simply calibrate the X-axis zero position without replacing any parts.

[0254] Scenario 3: Differentiation of thermal contamination in F-theta field lenses

[0255] Experimental conditions: The outer edge region of the F-theta field lens at the end of the scanning handpiece (corresponding to the scanning radius) A small amount of petroleum jelly is applied to the area to simulate tissue splatter contamination during clinical treatment. The system then positions the beam at this high-loss region and fires a test sequence containing 50 pulses.

[0256] See the appendix for experimental results. Figure 11 :

[0257] Traditional methods only detect that the average energy at that point is slightly lower than the standard value (about 90%), which can easily be misjudged as an optical path alignment deviation, leading to ineffective mechanical adjustments by maintenance personnel.

[0258] The method of this invention: Extracting the temporal characteristics of the energy of a pulse train. For example... Figure 11 As shown:

[0259] Normal lens (solid line): High energy amplitude and remains stable and flat within 50 pulses.

[0260] Mechanical blockage (dashed line): This is characterized by a low overall energy amplitude (step decrease), but the relative amplitude between pulses remains stable with no attenuation trend.

[0261] Contaminated lens (dotted line): This is characterized by a significant linear decrease in single-pulse energy as the pulse sequence continues (a ramp response caused by thermal lensing effect).

[0262] Diagnostic output: The system calculates the thermal modulation slope. =0.15mJ / pulse, far exceeding the threshold. The system was identified as thermal contamination, prompting the user to clean or replace the field lens directly, thus avoiding misdiagnosis.

[0263] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A fault detection method for a carbon dioxide dot laser treatment instrument, characterized in that, Includes the following steps: During the calibration phase, the scanning handpiece is controlled to scan and record the reference drive parameters that maintain the standard output, and the reference electrical impedance characteristics are established. Simultaneously acquire real-time scanning coordinates of the scanning handpiece, voltage and current data of the laser, and real-time drive signals; The dynamic impedance of the laser is calculated based on voltage and current data. If the deviation between the dynamic impedance and the reference electrical impedance characteristics exceeds a preset threshold, the fault source is determined to be the laser. If the deviation does not exceed the preset threshold, the compensation intensity scalar is calculated based on the difference between the real-time drive signal and the reference drive parameters, and the laser is controlled to emit a test pulse train to calculate the power modulation slope. Based on the spatial distribution of the compensation intensity scalar with the scanning radius and the power modulation slope, the fault type is determined to be geometric alignment fault, thermal contamination fault, or global optical path attenuation. The control of the laser to emit test pulse trains to calculate the power modulation slope specifically includes: The scanning handpiece is positioned to the scanning radius corresponding to the detected maximum compensation intensity scalar, and the laser is controlled to emit a test pulse sequence containing M consecutive pulses; Collect the energy value of each pulse in the test pulse sequence to construct a pulse train energy decay sequence; Linear regression analysis is performed on the pulse train energy attenuation sequence, and the absolute value of the slope of the regression line is calculated as the power modulation slope; The method of determining the fault type based on the spatial distribution of the compensation intensity scalar with the scanning radius and the power modulation slope specifically includes: Calculate the Pearson correlation coefficient between the compensation intensity scalar and the scanning radius; If the Pearson correlation coefficient is less than the preset correlation threshold and the mean of the compensation intensity scalar exceeds the normal range, it is determined to be a global optical path attenuation. If the Pearson correlation coefficient is greater than or equal to a preset correlation threshold, and the power modulation slope is greater than a preset thermal response threshold, it is determined to be a thermal pollution fault. If the Pearson correlation coefficient is greater than or equal to a preset correlation threshold, and the power modulation slope is less than or equal to a preset thermal response threshold, it is determined to be a geometric alignment fault.

2. The method of claim 1, wherein the method further comprises: The reference driving parameters that maintain the standard output are specifically included in the recording: Define a one-dimensional discretized grid for the scan field, the grid comprising multiple radius nodes from the center of the scan field to the maximum scan radius; For each radius node, the drive signal of the laser power supply is adjusted by the PID algorithm until the actual output energy stabilizes at the standard target value. The power command and duty cycle command at this time are recorded as the reference drive parameters, and a radius-drive parameter reference lookup table is constructed. Before calculating the compensation intensity scalar based on the difference between the real-time drive signal and the reference drive parameters, the method further includes: calculating the real-time radius using the real-time scan coordinates, and calculating the theoretical reference drive parameters corresponding to the current position in the reference lookup table using an interpolation model.

3. The method of claim 1, wherein the method further comprises: The synchronous acquisition of the scanning handpiece's real-time scanning coordinates, laser voltage and current data, and real-time drive signals specifically includes: Configure a hardware synchronous trigger source based on the PWM cycle, and set the delay time of the sampling time relative to the rising edge of the PWM signal. The delay time is configured to be greater than the rise time of the laser power supply output current and to make the sampling point fall in the current flat-top region. When the sampling time is reached, the analog-to-digital converter, the galvanometer position interface, and the optical path feedback unit are triggered in parallel to generate a synchronous data frame containing the timestamp of the same moment, the real-time scanning coordinates, the voltage and current data, and the real-time drive signal.

4. The method of claim 1, wherein the method further comprises: The calculation of the laser's dynamic impedance based on voltage and current data specifically includes: Construct a time-sliding window containing N sets of historical voltage and current values; The least squares method is used to perform linear regression on the voltage and current data within the time sliding window, and the regression slope is extracted as the dynamic impedance. The specific steps for determining that the fault source is a laser are as follows: calculate the deviation of the dynamic impedance from the reference electrical impedance characteristic; if the deviation is greater than a preset percentage of the average value of the reference impedance, the fault source is determined to be a laser.

5. The method of claim 2, wherein the method further comprises: The calculation of the compensation intensity scalar based on the difference between the real-time drive signal and the reference drive parameters specifically includes: Calculate the difference between the real-time drive signal and the theoretical reference drive parameters to obtain a residual vector that includes power compensation residual and duty cycle compensation residual; Using the weighting coefficients determined based on the system open-loop gain, the power compensation residual and the duty cycle compensation residual are weighted Euclidean norms to calculate the compensation intensity scalar.

6. The method of claim 1, wherein the method further comprises: The determination of a geometric alignment failure refers to a geometric alignment failure caused by mechanical deflection of the light guide arm; the determination of a thermal contamination failure refers to a thermal contamination failure of the optical lens.

7. The fault detection method for a carbon dioxide fractional laser therapy device according to claim 1, characterized in that, After determining whether the fault is caused by thermal contamination or geometric alignment, the method further includes the step of locating the specific faulty component: The scanning field is divided into four quadrants, and the average compensation intensity in each quadrant is calculated. Calculate the asymmetry factor of the difference between the maximum and minimum average compensation strength relative to the overall average value; If the asymmetry factor is less than a preset symmetry threshold, the faulty component is determined to be an F-theta lens; If the asymmetry factor is greater than or equal to the preset symmetry threshold, find the quadrant direction with the largest average compensation intensity and locate the faulty component as the galvanometer axis in the corresponding direction.

8. The fault detection method for a carbon dioxide fractional laser therapy device according to claim 1, characterized in that, The established reference electrical impedance characteristics include: Voltage and current data for multiple cycles are continuously acquired at different scanning positions. The average impedance and standard deviation are calculated, and the normal fluctuation range of the reference impedance is set based on the 3σ statistical principle.

Images