Vacuum furnace heating temperature uniformity detection method and system

Abstract

The present application relates to the technical field of vacuum heat treatment equipment monitoring, and particularly relates to a vacuum furnace heating temperature uniformity detection method and system, comprising: executing a sweep frequency pulse sequence to a furnace under a uniform temperature reference state, obtaining a power response of each heating zone to generate a reference response spectrum; reproducing the sequence to generate a real-time response spectrum during routine diagnosis; performing frequency domain transformation on the two groups of response spectra, extracting frequency spectrum features of each frequency band and constructing a frequency domain response synchronism index of adjacent zones; obtaining an external environment electric signal and eliminating a deviation component caused by the external environment electric signal to generate a net deviation feature set; and fusing the net deviation feature set and the synchronism index to generate a comprehensive diagnosis index to determine a temperature uniformity state. The present application fills a diagnostic blind area between a passive monitoring means and a macroscopic temperature difference overrun in thermal symmetry deviation, and provides a more forward-looking early warning capability for vacuum furnace temperature uniformity.

Description

Technical Field

[0001] This invention relates to the field of monitoring technology for vacuum heat treatment equipment, and in particular to a method and system for detecting the uniformity of heating temperature in a vacuum furnace. Background Technology

[0002] Vacuum furnaces are key heat treatment equipment in high-end manufacturing, and the uniformity of temperature distribution within the furnace directly determines the consistency of workpiece microstructure, deformation, and overall mechanical properties. To ensure that temperature uniformity meets process requirements, the industry commonly employs furnace temperature uniformity testing (TUS), which involves deploying multiple thermocouples within the furnace's working area, simulating the process temperature curve under no-load conditions, recording the temperature at each measuring point, and calculating the maximum temperature difference. Furthermore, to achieve online evaluation without shutting down the furnace, existing technologies propose using multi-source sensor arrays within the furnace to collect multi-dimensional signals such as temperature and pressure, constructing a three-dimensional feature field, and performing anomaly diagnosis. However, these methods are all passive monitoring methods, relying on signal changes under natural furnace conditions for analysis. They are insufficient for effectively warning of latent thermal symmetry shifts caused by early degradation of heating elements or slight detachment of local thermal shielding layers before macroscopic temperature differences become apparent. Summary of the Invention

[0003] To overcome the above shortcomings, this invention provides a method and system for detecting the uniformity of heating temperature in a vacuum furnace, aiming to improve the problem that existing passive monitoring methods are unable to achieve early diagnosis between the development of implicit thermal symmetry shifts in the uniformity of vacuum furnace temperature and the macroscopic temperature difference exceeding the limit.

[0004] In a first aspect, the present invention provides the following technical solution: a method for detecting the uniformity of heating temperature in a vacuum furnace, comprising: S1. Under the uniform temperature reference state of the vacuum furnace, a sweeping pulse sequence is executed into the furnace to simultaneously acquire the power response signal of each heating zone and generate a multi-zone reference response spectrum. S2. During routine diagnosis, the frequency sweeping pulse sequence is reproduced under the same temperature conditions to synchronously acquire the power response signal of each heating zone and generate a multi-zone real-time response spectrum. S3. Perform frequency domain transformation on the multi-zone reference response spectrum and the multi-zone real-time response spectrum, extract the spectral characteristics of each heating zone in different frequency bands, and construct the frequency domain response synchronization index between adjacent heating zones. S4. Acquire the electrical signal of the external environment of the furnace body, extract its frequency domain features, and remove the deviation components caused by external environmental fluctuations from the spectrum features of step three to generate a net deviation feature set. S5. Integrate the net deviation feature set and the frequency domain response synchronization index to generate a comprehensive diagnostic index, and determine the current temperature uniformity status based on the comprehensive diagnostic index.

[0005] Secondly, the present invention provides the following technical solution: a vacuum furnace heating temperature uniformity detection system, the system comprising: The reference response generation module is used to execute a sweeping pulse sequence into the furnace when the vacuum furnace is in a uniform temperature reference state, and simultaneously acquire the power response signals of each heating zone to generate a multi-zone reference response spectrum. The real-time response generation module is used to reproduce the frequency sweeping pulse sequence under the same temperature conditions during routine diagnosis, synchronously acquire the power response signals of each heating zone, and generate a multi-zone real-time response spectrum. The frequency domain feature extraction module is used to perform frequency domain transformation on the multi-zone reference response spectrum and the multi-zone real-time response spectrum, extract the spectral features of each heating zone in different frequency bands, and construct the frequency domain response synchronization index between adjacent heating zones. The noise removal module is used to acquire the electrical signal of the external environment of the furnace body, extract its frequency domain features, and remove the deviation components caused by external environmental fluctuations from the spectrum features extracted by the frequency domain feature extraction module to generate a net deviation feature set. The diagnostic judgment module is used to integrate the net deviation feature set and the frequency domain response synchronization index to generate a comprehensive diagnostic index, and to determine the current temperature uniformity status based on the comprehensive diagnostic index.

[0006] The present invention has the following beneficial effects: 1. In this invention, by actively applying a frequency-sweeping pulse sequence as an excitation signal and performing frequency domain transformation and adjacent zone cohesion analysis on the power response of each heating zone, the symmetry change of dynamic thermal coupling between each heating zone can be actively detected. The implicit deterioration trend of thermal symmetry can be captured before the macroscopic temperature difference is manifested, filling the diagnostic blind spot of passive monitoring methods between thermal symmetry shift and macroscopic temperature difference exceeding the limit, and providing a more forward-looking early warning capability for the temperature uniformity of vacuum furnaces.

[0007] 2. In this invention, a frequency-sweeping pulse sequence is used as a perturbation excitation. The thermal response characteristics at different time scales are detected by frequency changes. Continuous testing can be performed without waiting for the furnace to fully stabilize between pulses, which significantly shortens the online diagnostic time while ensuring the richness of diagnostic information.

[0008] 3. In this invention, by collecting the external environmental electrical signals of the furnace body and performing frequency domain noise discrimination, the deviation components caused by external environmental fluctuations are removed from the spectral characteristics of each heating zone, which effectively avoids false alarms of temperature uniformity caused by external interference such as power grid voltage fluctuations and improves the reliability of the diagnostic results.

[0009] 4. In this invention, the integrated diagnostic stage simultaneously introduces global deviation measurement and local extreme value penalty, which not only reflects the overall temperature deviation level of the furnace, but also prevents the severe deterioration of individual heating zones from being masked by the overall average value, so that the comprehensive diagnostic index can more comprehensively characterize the true state of temperature uniformity. Attached Figure Description

[0010] Figure 1 This is a schematic flowchart of a method for detecting the uniformity of heating temperature in a vacuum furnace proposed in this invention. Figure 2 This is a schematic diagram of the frequency domain feature extraction and synchronization index construction process of a vacuum furnace heating temperature uniformity detection method proposed in this invention. Figure 3 This is a schematic diagram of the architecture of a vacuum furnace heating temperature uniformity detection system proposed in this invention. Detailed Implementation

[0011] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] Example 1 In a first embodiment of the present invention, the present invention provides a method for detecting the uniformity of heating temperature in a vacuum furnace, such as... Figures 1-2 As shown, it includes the following steps: S1. Under the condition of uniform temperature reference in the vacuum furnace, a sweeping pulse sequence is executed into the furnace chamber to simultaneously acquire the power response signal of each heating zone and generate a multi-zone reference response spectrum.

[0013] Furthermore, in S1, a sweep-frequency pulse sequence is executed into the furnace to simultaneously acquire the power response signals of each heating zone, generating a multi-zone reference response spectrum, including: After the vacuum furnace passes the standard furnace temperature uniformity test, it is kept at a constant temperature under no-load at the target diagnostic temperature until it reaches a stable state, which is then determined as the uniform temperature reference state. High-purity inert gas is periodically introduced into the furnace. The pressure peak value of each gas injection pulse is set to the same constant value. The duration and width of each gas injection pulse and the interval between adjacent gas injection pulses vary according to a preset frequency sweeping rule to form a frequency sweeping pulse sequence. The power response time-domain signal of each heating zone during the execution of the frequency sweep pulse sequence is recorded synchronously, and after normalization, a multi-zone reference response spectrum is generated.

[0014] Specifically, after the vacuum furnace completes the standard furnace temperature uniformity test and the results are satisfactory, the temperature is raised to the target diagnostic temperature and kept constant under no-load until the fluctuation amplitude of the temperature control thermocouples in each heating zone does not exceed a preset range for a continuous preset time. This state is defined as the uniform temperature reference state. A feasible setting is: within 5 consecutive minutes, the fluctuation amplitude of the temperature control thermocouple measurement values ​​in each heating zone does not exceed ±1℃. The target diagnostic temperature is selected as the typical process temperature of the vacuum furnace, such as 850℃ or 1020℃. Under the uniform temperature reference state, a frequency sweep pulse sequence is started. The frequency sweep pulse sequence is achieved through the coordinated action of the gas flow controller and the vacuum valve group, periodically charging high-purity argon into the furnace. The pressure peak value of each charging pulse is set to the same constant value, a feasible value is 200Pa. The duration width of each charging pulse and the interval time of adjacent charging pulses vary according to the preset frequency sweep rules: the pulse duration width gradually decreases from the first width value to the second width value, and the interval time of adjacent pulses gradually decreases from the first interval value to the second interval value. One specific setting is as follows: the pulse width decreases from 10 seconds to 2 seconds, and the interval time decreases from 20 seconds to 5 seconds. Each combination of width and interval executes one pulse, forming a pulse sequence with a frequency varying from low to high. The total duration of the entire sweeping pulse sequence is controlled within 180 seconds.

[0015] During the execution of the sweep-frequency pulse sequence, the power compensation value of each heating zone is synchronously recorded over time. The recording method is as follows: using the power feedback signal of each heating zone as the data source, with a sampling frequency of no less than 10Hz, the entire process data is recorded from several seconds before the start of the first inflation pulse to several seconds after the end of the last inflation pulse. After recording, the time-domain power response signals of each curve are normalized. The normalization method is to subtract the average power value of each curve in the steady-state stage before inflation from the power value of each curve, and then divide by the average value to obtain a dimensionless relative power deviation sequence. The processed curves of all heating zones together constitute a multi-zone reference response spectrum and are stored in the data storage module. The calculation formula for the above normalization process is: ; in, Let be the power compensation value for the i-th heating zone at time t. Let be the average power of the i-th heating zone during the steady-state phase before inflation. The normalized relative power deviation is used as the response value of the heating zone at time t in the multi-zone reference response spectrum.

[0016] This baseline response spectrum is used for comparative analysis with the multi-zone real-time response spectra obtained during subsequent routine diagnostics, in order to extract frequency domain features reflecting the temperature uniformity status.

[0017] S2. During routine diagnosis, the sweeping pulse sequence is reproduced under the same temperature conditions, and the power response signal of each heating zone is acquired synchronously to generate a multi-zone real-time response spectrum.

[0018] Furthermore, in S2, a multi-region real-time response spectrum is generated, including: During routine diagnostics, the vacuum furnace is heated to the same temperature as the target diagnostic temperature and kept at that temperature. A frequency sweep pulse sequence is executed into the furnace again. The rules for the change of pressure peak value, duration width, and interval time of the frequency sweep pulse sequence are the same as those in step S1. The power response time-domain signal of each heating zone during the execution of the frequency sweep pulse sequence is recorded synchronously, and after normalization processing, a multi-zone real-time response spectrum is generated.

[0019] Specifically, during routine diagnostics, the vacuum furnace is heated to the same target diagnostic temperature as when acquiring the multi-zone reference response spectrum and held at that temperature. The temperature is held constant if the fluctuation amplitude of the temperature-controlled thermocouples in each heating zone does not exceed a preset range for a continuously preset duration. Under constant temperature conditions, a frequency-sweeping pulse sequence is executed again into the furnace. All parameters of this frequency-sweeping pulse sequence are identical to those used when acquiring the multi-zone reference response spectrum: the peak pressure is maintained at 200 Pa, and the duration and width of each inflation pulse and the interval between adjacent inflation pulses vary according to the same frequency sweep rule—that is, the pulse width gradually decreases from 10 seconds to 2 seconds, and the interval time gradually decreases from 20 seconds to 5 seconds, with the total duration controlled within 180 seconds.

[0020] During the execution of the sweep-frequency pulse sequence, the power compensation value of each heating zone is synchronously recorded over time at the same sampling frequency. The recording period extends from several seconds before the start of the first inflation pulse to several seconds after the end of the last inflation pulse. The acquired power response time-domain signals are processed using the same normalization method as when acquiring the multi-zone reference response spectrum, i.e.: ; in, This represents the power compensation value of the i-th heating zone at time t during this diagnosis. This represents the average power of the i-th heating zone during the steady-state phase before inflation, as of the time of this diagnosis. The normalized relative power deviation is used as the response value of the heating zone at time t in the multi-zone real-time response spectrum. The processed curves of all heating zones together constitute the multi-zone real-time response spectrum and are stored in the data storage module for subsequent frequency domain comparison analysis with the multi-zone reference response spectrum.

[0021] The same normalization method as under the baseline condition is used here to ensure that the two sets of response spectra are comparable on the amplitude scale. If there are workpieces loaded in the furnace during routine diagnostics, the average power in the steady-state stage before gas charging will automatically reflect the change in the current furnace heat capacity. The normalization operation can partially eliminate the impact of loading differences on subsequent analysis.

[0022] S3. Perform frequency domain transformation on the multi-zone reference response spectrum and the multi-zone real-time response spectrum, extract the spectral characteristics of each heating zone in different frequency bands, and construct the frequency domain response synchronization index between adjacent heating zones.

[0023] Furthermore, in S3, the spectral characteristics of each heating zone in different frequency bands are extracted, including: Fourier transforms were performed on the power response time-domain signals of each heating zone in the multi-zone reference response spectrum and the multi-zone real-time response spectrum to obtain the reference frequency domain data and real-time frequency domain data of each heating zone. The dividing frequency for frequency band division is determined based on the range of variation in duration and width and the range of variation in interval time of the swept pulse sequence. Based on the boundary frequency, low-frequency amplitude features and high-frequency amplitude features are extracted from the reference frequency domain data and real-time frequency domain data of each heating zone, respectively, as the spectral features of each heating zone in different frequency bands.

[0024] Furthermore, in S3, a frequency domain response synchronization index is constructed between adjacent heating zones, including: Two heating zones that are physically adjacent within the furnace are selected as a pair of areas to be evaluated. Based on the reference frequency domain data of each heating zone, the amplitude square coherence function of the region to be evaluated in the full frequency band or a specified frequency band is calculated to obtain the coherence coefficient curve as a function of frequency, and the integral area of ​​the coherence coefficient curve in the corresponding frequency band is calculated as the reference cohesion. Based on the real-time frequency domain data of each heating zone, the amplitude square coherence function of the region to be evaluated in the full frequency band or a specified frequency band is calculated to obtain the coherence coefficient curve that varies with frequency. The integral area of ​​the coherence coefficient curve in the corresponding frequency band is calculated as the real-time cohesion. The rate of decrease in real-time cohesion relative to baseline cohesion is used as an indicator of frequency domain response synchronicity.

[0025] Specifically, Fourier transforms are performed on the time-domain power response signals of each heating zone in the multi-zone reference response spectrum and the time-domain power response signals of each heating zone in the multi-zone real-time response spectrum. The Fourier transform is implemented using the Fast Fourier Transform algorithm. A Hanning window is applied to the time-domain signal before the transform to reduce spectral leakage. The window length of the Hanning window is set to half the total number of sampling points within the total duration of the swept pulse sequence to achieve a balance between frequency resolution and time resolution. If the sampling frequency is 10Hz and the total duration of the swept pulse sequence is 180 seconds, then the total number of sampling points is 1800, and the window length is 900 points. After the Fourier transform of the time-domain signal of the i-th heating zone, the reference frequency domain data of that heating zone is obtained. and real-time frequency domain data Where f represents the frequency. The dividing frequency for the frequency band is determined based on the range of duration and interval variation of the swept-frequency pulse sequence. The pulse width range of the swept-frequency pulse sequence is 2 to 10 seconds, corresponding to an excitation frequency range of approximately 0.1 Hz to 0.5 Hz; the pulse interval range is 5 to 20 seconds, corresponding to an excitation frequency range of approximately 0.05 Hz to 0.2 Hz. Accordingly, the dividing frequency is set to 0.1 Hz, dividing the entire frequency band into a low-frequency band (0.05 Hz to 0.1 Hz) and a high-frequency band (0.1 Hz to 0.5 Hz). The principle for selecting the dividing frequency is that the low-frequency band mainly corresponds to the slow thermal response excited by the change in pulse interval time, while the high-frequency band mainly corresponds to the fast thermal response excited by the change in pulse width.

[0026] Based on the boundary frequency, reference frequency domain data from each heating zone and real-time frequency domain data Spectral features are extracted from each of the following data points. For the reference frequency domain data, the squared amplitude values ​​of each frequency point within the low-frequency band are extracted and summed to obtain the low-frequency band amplitude features of the i-th heating zone. The amplitude characteristics of the i-th heating zone are obtained by summing the squared amplitudes of each frequency point within the high-frequency band. For real-time frequency domain data, the corresponding low-frequency amplitude features are extracted using the same method. and high-frequency amplitude characteristics The amplitude characteristics of each heating zone in the low-frequency and high-frequency bands mentioned above together constitute the spectral characteristics of each heating zone in different frequency bands.

[0027] After extracting spectral features, a frequency domain response synchronization index is constructed between adjacent heating zones. Two heating zones physically adjacent within the furnace are selected as a pair of regions to be evaluated. For example, the upper heating zone and the middle heating zone form one pair, the middle heating zone and the lower heating zone form another pair, and adjacent heating zones within the same height layer also form separate pairs. For each pair of regions to be evaluated, the baseline cohesion and real-time cohesion are calculated. Taking the pair of regions to be evaluated consisting of the i-th heating zone and the j-th heating zone as an example, the baseline cohesion is calculated based on the baseline frequency domain data. and Calculate the squared coherence function of the two values. : ; in, Let be the cross-power spectral density of the reference frequency domain data between the i-th heating zone and the j-th heating zone. Let be the auto-power spectral density of the reference frequency domain data for the i-th heating zone. Let f be the autopower spectral density of the reference frequency domain data for the j-th heating zone, and f be the frequency. The amplitude squared coherence function... The value ranges from 0 to 1; the closer the value is to 1, the stronger the correlation between the responses of the two heating regions at that frequency. Cross-power spectral density. and self-power spectral density , All estimations were performed using the Welch-averaged periodogram method. During estimation, the time-domain signal of each heating zone was divided into several equal-length segments according to a preset segment length. A preset overlap rate was set between adjacent segments. After applying a Hanning window to each segment, a Fourier transform was performed, and then the periodograms of all segments were averaged. The segment length was set to one-quarter of the total duration of the swept-frequency pulse sequence, the overlap rate was set to 50%, and the number of Fourier transform points was set to twice the segment length and rounded up to the nearest power of 2. This estimation method can obtain statistically stable power spectrum estimates while ensuring frequency resolution.

[0028] For squared coherence functions Integrating across the entire frequency band yields the baseline cohesion of the region to be evaluated. : ; in, The lower limit frequency for the entire frequency band is set to 0.05Hz. The upper limit frequency of the entire frequency band is set to 0.5Hz. The lower limit frequency of the entire frequency band is... It is determined by the reciprocal of the maximum pulse interval of the swept-frequency pulse sequence, i.e. =1 / 20 second = 0.05 Hz; the upper limit frequency of the entire frequency band. It is determined by the reciprocal of the minimum pulse duration of the swept-frequency pulse sequence, i.e. =1 / 2 second = 0.5 Hz.

[0029] Using the same method, based on real-time frequency domain data and Calculate the real-time amplitude squared coherence function and integrate it within the same frequency band to obtain the real-time cohesion of the region pair to be evaluated. .

[0030] Then, the rate of decrease in real-time cohesion of the region to be evaluated relative to the baseline cohesion is calculated. : ; in, The rate of decrease is represented by a value that indicates a more severe deterioration in the thermal response synchronicity of the adjacent heating zones compared to the baseline state.

[0031] The above calculations were performed on all pairs of regions to be evaluated, and the rate of decline for each pair of regions to be evaluated was calculated. Together, they constitute the frequency domain response synchronization index, which is stored in the data storage module for subsequent net deviation elimination and comprehensive diagnostic index generation.

[0032] S4. Obtain the electrical signal of the external environment of the furnace body, extract its frequency domain features, and remove the deviation components caused by external environmental fluctuations from the spectrum features of step three to generate a net deviation feature set.

[0033] Furthermore, in S4, a net deviation feature set is generated, including: The voltage signal connected to the power supply circuit of the vacuum furnace is collected as the electrical signal of the external environment of the furnace body; Fourier transform is performed on the electrical signal of the external environment of the furnace body, and the frequency components whose amplitude exceeds the preset threshold are extracted to form the main frequency set of external noise. In the spectral characteristics of each heating zone, the component whose frequency is the same as any frequency component in the set of dominant external noise frequencies is marked as the noise deviation component. The noise deviation component is removed from the spectral characteristics of each heating zone, and the remaining spectral characteristics after removal are used as the net deviation feature set.

[0034] Specifically, after completing the spectral feature extraction and frequency domain response synchronization index construction, an external noise removal step is performed to eliminate spurious deviations introduced by external environmental factors such as grid voltage fluctuations.

[0035] The voltage signal connected to the power supply circuit of the vacuum furnace is collected as the external environmental electrical signal of the furnace. The voltage signal is collected at the output terminal of the voltage transformer of the main power supply bus of the vacuum furnace. The sampling frequency is consistent with the sampling frequency of the power response signal, not lower than 10Hz, and the collection period is synchronized with the recording period of the multi-zone real-time response spectrum. A Fourier transform is performed on the collected voltage signal to obtain its frequency domain data. Frequency components with amplitudes exceeding a preset threshold are extracted from the frequency domain data of the voltage signal to form the external noise dominant frequency set. One feasible value for the preset threshold is one-tenth of the largest amplitude value in the voltage signal frequency domain data. Each frequency component in the external noise dominant frequency set represents the presence of external power grid fluctuation interference at that frequency.

[0036] The spectral characteristics of the real-time frequency domain data extracted from each heating zone in step S3 are compared with the main frequency set of external noise. For each frequency component in the main frequency set of external noise, it is first determined whether the amplitude of the real-time frequency domain data of each heating zone at that frequency shows statistical consistency with the amplitude of the real-time frequency domain data of the other heating zones at that frequency. If the amplitude of a certain heating zone at a certain frequency deviates significantly from the statistical distribution of the amplitudes of the other heating zones, it is determined that there is superposition of the real thermal response signal and noise at that frequency. At this time, a noise contribution subtraction operation is performed: the estimated noise contribution value is subtracted from the real-time amplitude of the heating zone at that frequency, and the real thermal response signal component is retained. The remaining amplitude after subtraction is taken as the net amplitude at that frequency. If the amplitudes of all heating zones at a certain frequency are basically consistent, it is determined that there is only external noise at that frequency, and the amplitude component at that frequency is directly set to zero.

[0037] The noise contribution estimate is determined as follows: Pure noise frequency points that do not contain the actual thermal response signal are selected from the set of dominant external noise frequencies. The average amplitude of each heating zone at these frequency points is calculated as the noise contribution estimate at that noise frequency. If all frequency points in the set of dominant external noise frequencies are determined to have superimposed actual thermal response signals, the amplitude at that frequency in the voltage signal frequency domain data is multiplied by a preset coupling coefficient as the noise contribution estimate. One feasible value for this coupling coefficient is 0.3.

[0038] Based on the above judgments and operations, for the low-frequency amplitude characteristics of the i-th heating zone, the net low-frequency amplitude characteristics obtained after the elimination operation are... for: When the kth frequency in the set of dominant external noise frequencies When judged as pure noise: ; When the kth frequency in the set of dominant external noise frequencies When judged as a superposition of real signal and noise: ; in, The real-time low-frequency amplitude features of the i-th heating zone extracted in step S3 are... This is the set of dominant frequencies of external noise. The set of dominant frequencies of external noise that is determined to be the superposition of real signal and noise. This refers to the k-th noise frequency in the low-frequency range within the set of dominant external noise frequencies. For the real-time frequency domain data of the i-th heating zone in frequency The value at that location. For frequency The noise contribution estimate at that location. Similarly, the same elimination operation is performed on the high-frequency amplitude characteristics to obtain the net high-frequency amplitude characteristics.

[0039] After performing the above judgments and operations in each heating zone, the remaining net low-frequency amplitude characteristics and net high-frequency amplitude characteristics together constitute the net deviation feature set, which is stored in the data storage module. The net deviation feature set eliminates the power response fluctuation component caused by external power grid fluctuations and retains only the true deviation information caused by changes in the internal thermal characteristics of the furnace, providing a more reliable data foundation for the subsequent generation of comprehensive diagnostic indices.

[0040] S5. Integrate the net deviation feature set and frequency domain response synchronization index to generate a comprehensive diagnostic index, and determine the current temperature uniformity status based on the comprehensive diagnostic index.

[0041] Furthermore, in S5, a comprehensive diagnostic index is generated, including: The net deviation values ​​of each heating zone in the net deviation feature set are arranged according to their spatial location to construct the spatial distribution vector of furnace thermal deviation, and the weighted Euclidean norm of the spatial distribution vector of furnace thermal deviation is calculated as the overall deviation metric. Arrange the synchronicity degradation rate of each adjacent heating zone pair in the frequency domain response synchronicity index according to spatial location, construct the thermal coupling degradation spatial distribution vector, and calculate the weighted Euclidean norm of the thermal coupling degradation spatial distribution vector as the overall metric of coupling degradation. The maximum value of the net deviation in each heating zone in the net deviation feature set is taken as the deviation extreme value penalty. The maximum value of the synchronicity decrease rate of each adjacent heating zone pair in the frequency domain response synchronicity index is taken as the coupling extremum penalty. The overall deviation metric, the overall coupling degradation metric, the deviation extreme value penalty, and the coupling extreme value penalty are weighted and summed according to preset weights to obtain the comprehensive diagnostic index.

[0042] Furthermore, in S5, the current temperature uniformity status is determined based on this comprehensive diagnostic index, including: The comprehensive diagnostic index is compared with a preset first threshold. If the comprehensive diagnostic index is less than or equal to the first threshold, the current temperature uniformity is determined to be normal. If the comprehensive diagnostic index is greater than the first threshold and less than or equal to the preset second threshold, the current temperature uniformity is determined to be deteriorating, and an early warning is issued. If the comprehensive diagnostic index is greater than the second threshold, the current temperature uniformity is determined to be abnormal, and an alarm signal is output.

[0043] Specifically, the net deviation values ​​of each heating zone are extracted from the net deviation feature set. For the i-th heating zone, its net low-frequency amplitude features are... With the characteristics of net high frequency band amplitude Add them together to obtain the net deviation value of the heating zone. : ; Arrange the net deviation values ​​of all heating zones according to their physical spatial location within the furnace to construct a spatial distribution vector of furnace thermal deviation. If the vacuum furnace has N heating zones, then This is an N-dimensional vector, with each dimension corresponding to the net deviation value of a heating zone. The arrangement follows a top-to-bottom, front-to-back spatial order. Calculate the spatial distribution vector of the furnace thermal deviation. The weighted Euclidean norm is used as a measure of overall deviation. : ; in, The spatial weighting coefficient is the one corresponding to the i-th heating zone. The value is determined based on the importance of the heating zone's location within the furnace. Weighting coefficient. The general method for determining the weight of the effective working area of ​​the furnace is as follows: Calculate the Euclidean distance between the geometric center of each heating zone and the geometric center of the effective working area of ​​the furnace. Set the weight of the heating zone corresponding to the maximum distance to 0.6, and the weight of the heating zone corresponding to the minimum distance to 1.0. The weights of the remaining heating zones are determined by linear interpolation between the minimum and maximum distances mentioned above. The method for determining the geometric center of the effective working area of ​​the furnace is as follows: Take the midpoint of the area enclosed by reducing the corresponding dimension of the effective working area by 10% along each coordinate axis towards the furnace edge from the geometric center point of the furnace interior space. Let N be the net deviation value of the i-th heating zone, and N be the total number of heating zones.

[0044] The synchronicity degradation rate of each adjacent heating zone pair is extracted from the frequency domain response synchronicity index. The synchronicity degradation rates of all pairs of regions to be evaluated are arranged according to their spatial location to construct a spatial distribution vector of thermal coupling degradation. If there are M pairs of regions to be evaluated, then Let M be a vector, where each dimension corresponds to the rate of synchronicity decrease of a pair of adjacent heating zones. Calculate the spatial distribution vector of thermal coupling degradation. The weighted Euclidean norm is used as a global measure of coupling degradation. : ; in, This is the spatial weighting coefficient for the kth group of adjacent heating zone pairs. The method of setting the value is consistent with the weighting principle of the overall deviation metric. Let M be the rate of synchronicity decrease of the k-th group of adjacent heating zone pairs, and M be the total number of region pairs to be evaluated.

[0045] Then, a local extremum penalty is introduced to prevent severe degradation in individual heating zones from being masked by the overall average. The maximum net deviation value of each heating zone in the net deviation feature set is taken as the deviation extremum penalty. : ; The maximum value of the rate of decrease in synchronicity between adjacent heating zones in the frequency domain response synchronicity index is taken as the coupling extremum penalty. : ; The four metrics mentioned above are weighted and summed according to preset weights to generate a comprehensive diagnostic index. : ; in, , , , Preset weighting coefficients. One feasible weighting setting is: Take 0.35, Take 0.35, Take 0.15, A weight of 0.15 is used. This weight setting ensures that the overall deviation metric and the overall coupling degradation metric both dominate, while reserving an appropriate contribution share for local extremum penalties. The method for determining the above-mentioned preset weight coefficients is as follows: After the vacuum furnace passes the standard furnace temperature uniformity test, the diagnostic process under the baseline state is executed multiple times, and the comprehensive diagnostic index is recorded each time. Taking the minimum mean and minimum variance of the comprehensive diagnostic index as the optimization objective, a set of weight coefficients that minimizes and stabilizes the comprehensive diagnostic index under the baseline state is determined through a grid search method, and this set is used as the preset weight coefficient for this vacuum furnace model.

[0046] After generating the comprehensive diagnostic index, the temperature uniformity status is determined. The comprehensive diagnostic index D is compared with a preset first threshold. Second threshold Comparison, among which First threshold One feasible value is 0.15, the second threshold. One feasible value is 0.35. If If so, the current temperature uniformity is determined to be normal; if If the current temperature uniformity is determined to be deteriorating, a warning will be issued. If the current temperature uniformity is deemed abnormal, an alarm signal will be output. The alarm signal may take the form of a pop-up alarm on the human-machine interface, activation of an audible and visual alarm device, or sending an abnormal message to the production management system. First threshold. Second threshold The general method for determining this is as follows: During the historical healthy operation period of the vacuum furnace after it has passed the standard furnace temperature uniformity test, collect the comprehensive diagnostic index of at least 20 routine diagnoses and calculate the average value of this set of data. and standard deviation The first threshold is set as follows: The second threshold is set to This determination method is based on the principle of statistical process control, which ensures that the probability of the comprehensive diagnostic index exceeding the first threshold when the furnace condition is healthy is lower than the preset false alarm rate, thus ensuring that the alarm logic has statistical basis and reproducibility.

[0047] The values ​​of the first and second thresholds mentioned above can be adjusted according to the specific vacuum furnace model and historical operating data. When using this method for the first time, the recommended values ​​mentioned above can be used; after accumulating several diagnostic data, the thresholds can be recalibrated based on the statistical distribution of the comprehensive diagnostic index of the vacuum furnace in a healthy state.

[0048] Furthermore, after step S5, the following steps are also included: Based on the net deviation values ​​of each heating zone in the net deviation feature set, with each heating zone as the radial axis and the corresponding net deviation value of each heating zone as the radial coordinate value, a real-time contour line of the current diagnostic cycle is drawn to form a furnace cavity thermal health radar map. In the furnace cavity thermal health radar map, the historical contour lines of the most recent historical diagnostic cycles are superimposed and displayed, with each historical contour line distinguished by different colors or line types. When the real-time coordinate value of any radial axis exceeds the preset statistical boundary of the historical coordinate value of that axis, a highlight mark is made on that radial axis; In the central area of ​​the furnace cavity thermal health radar map, the comprehensive diagnostic index and its corresponding temperature uniformity status indicator are displayed in numerical form.

[0049] Specifically, the net deviation values ​​of each heating zone are extracted from the net deviation feature set. Using the spatial location identifier of each heating zone as the radial axis label and the net deviation value as the radial coordinate value, the net deviation values ​​of each heating zone in the current diagnostic cycle are plotted sequentially in a polar coordinate system and connected to form a closed real-time contour line, constituting a furnace cavity thermal health radar chart. The number of radial axes is consistent with the number of heating zones, and the range of the radial coordinates is uniformly set to zero to the preset upper limit of the net deviation value.

[0050] The net deviation feature set of the most recent historical diagnostic cycles is read from the data storage module. Using the same drawing method as the real-time contour line, the contour lines of each historical diagnostic cycle are overlaid and displayed on the same furnace thermal health radar chart. Each historical contour line is distinguished by a different color; for example, the current real-time contour line is a solid red line, the most recent historical contour line is a dashed blue line, the most recent historical contour line is a dashed green line, and so on. Simultaneously, the correspondence between each color and the diagnostic cycle is marked in the legend area.

[0051] For each radial axis, calculate the mean and standard deviation of the historical net deviation values ​​corresponding to that axis, and determine the preset statistical boundaries accordingly. Let the sequence of net deviation values ​​for the i-th radial axis in the most recent H historical diagnostic cycles be denoted as follows. Where H is the total number of historical diagnostic cycles, and the minimum value of H is 5. The average historical net deviation value of this axis. The calculation formula is: ; The standard deviation of the historical net deviation value of this axis The calculation formula is: ; The preset statistical boundary upper and lower limits for this axis are as follows: ; ; When the real-time net deviation value of the radial axis satisfy or When the real-time coordinate value exceeds the preset statistical boundary, a highlight mark is made at the real-time coordinate point of the radial axis, and the highlighting method is marked by a flashing dot. At the same time, the font of the axis label of the radial axis is changed to bold.

[0052] In the central blank area of ​​the furnace thermal health radar chart, the comprehensive diagnostic index generated in the current diagnostic cycle is displayed numerically, and the corresponding temperature uniformity status is indicated by color-coded blocks. A green block indicates a normal state, a yellow block indicates a deterioration trend, and a red block indicates an abnormal state.

[0053] If sufficient historical diagnostic data has not yet been accumulated, i.e., the number of historical diagnostic cycles is less than the preset minimum of 5, then only a real-time outline is drawn, and the comprehensive diagnostic index and status indicator are displayed in the central area, without historical overlay or statistical boundary determination.

[0054] Example 2 The diagnosis of temperature uniformity in vacuum furnaces has long relied on passive monitoring methods, which involve collecting and analyzing signals such as temperature and pressure under natural conditions by deploying thermocouples or sensor arrays within the furnace. However, when heating elements experience only slight impedance drift, local thermal shielding layers show micro-detachment, or insulation materials begin to slowly deteriorate, the dynamic thermal coupling symmetry between heating zones within the furnace will shift before the macroscopic temperature difference. This early, latent degradation manifests as extremely weak thermal response inconsistencies in the natural signal, easily masked by the system's normal closed-loop temperature control compensation behavior. Existing passive monitoring methods, lacking an active excitation mechanism, cannot specifically excite and separate the dynamic thermal coupling characteristics between heating zones, resulting in a diagnostic blind spot between the thermal symmetry shift stage and the macroscopic temperature difference exceeding the limit stage. This leads to a lack of effective early warning mechanisms between the occurrence of latent degradation and the appearance of measurable temperature differences. Typically, furnace abnormalities are only detected after batch-specific quality deviations occur in the workpiece, missing the optimal time for predictive maintenance. To address these issues, this invention provides a vacuum furnace heating temperature uniformity detection system, the structure of which is as follows: Figure 3 As shown. The specific implementation process of this system is as follows: The reference response generation module is used to execute a sweeping pulse sequence into the furnace when the vacuum furnace is in a uniform temperature reference state, and simultaneously acquire the power response signals of each heating zone to generate a multi-zone reference response spectrum. The real-time response generation module is used to reproduce the sweeping pulse sequence under the same temperature conditions during routine diagnosis, synchronously acquire the power response signal of each heating zone, and generate a multi-zone real-time response spectrum. The frequency domain feature extraction module is used to perform frequency domain transformation on the multi-zone reference response spectrum and the multi-zone real-time response spectrum, extract the spectral features of each heating zone in different frequency bands, and construct the frequency domain response synchronization index between adjacent heating zones. The noise removal module is used to acquire the electrical signals of the external environment of the furnace body, extract its frequency domain features, and remove the deviation components caused by external environmental fluctuations from the spectrum features extracted by the frequency domain feature extraction module to generate a net deviation feature set. The diagnostic judgment module is used to integrate the net deviation feature set and the frequency domain response synchronization index to generate a comprehensive diagnostic index, and to determine the current temperature uniformity status based on the comprehensive diagnostic index.

[0055] Specifically, based on the functions of each module in the above-mentioned system, a specific application example is provided below as a second embodiment.

[0056] A certain type of vacuum brazing furnace is equipped with six independently temperature-controlled heating zones: upper front zone, upper rear zone, middle front zone, middle rear zone, lower front zone, and lower rear zone. Each zone has independent temperature-controlled thermocouple feedback, and power regulation is executed by a thyristor frequency controller. During the baseline establishment phase, after the vacuum furnace passes the standard furnace temperature uniformity test, the furnace temperature is raised to the target diagnostic temperature of 950℃ and kept constant under no-load until the temperature fluctuation of each zone's temperature-controlled thermocouple does not exceed ±0.3℃ within 10 minutes. The baseline response generation module initiates a frequency-sweeping pulse sequence, periodically injecting high-purity argon gas into the furnace chamber through a gas flow controller. The peak pressure is fixed at 200Pa, the pulse width gradually decreases from 10 seconds to 2 seconds, and the interval between adjacent pulses gradually decreases from 20 seconds to 5 seconds, for a total duration of 180 seconds. Simultaneously, the power compensation values ​​of the six heating zones are recorded over time at a sampling frequency of 10Hz. After normalization, a multi-zone baseline response spectrum is generated and stored in the data storage module.

[0057] Three days later, routine online diagnostics were performed. The vacuum furnace was heated to 950℃ and kept at that temperature. The real-time response generation module executed the sweep frequency pulse sequence again with the exact same sweep frequency parameters, synchronously recording the power response time-domain signals of the six heating zones and normalizing them to generate multi-zone real-time response spectra.

[0058] The frequency domain feature extraction module receives multi-zone reference response spectra and multi-zone real-time response spectra, performs Fast Fourier Transform on the time-domain signals of each heating zone, and obtains reference frequency domain data and real-time frequency domain data. Based on the excitation frequency range corresponding to pulse widths of 2 to 10 seconds and intervals of 5 to 20 seconds, a dividing frequency of 0.1 Hz is set, dividing the frequency into a low-frequency band (0.05 Hz to 0.1 Hz) and a high-frequency band (0.1 Hz to 0.5 Hz), and extracts the amplitude features of each heating zone in the low-frequency and high-frequency bands respectively. Simultaneously, five pairs of adjacent heating zones were selected: the upper front zone and the upper rear zone, the upper front zone and the middle front zone, the middle front zone and the middle rear zone, the middle front zone and the lower front zone, and the lower front zone and the lower rear zone. Based on the reference frequency domain data and the real-time frequency domain data, the integral area of ​​the amplitude square coherence function of each pair of zones in the entire frequency band was calculated to obtain the reference cohesion and the real-time cohesion. Then, the decrease rate of the real-time cohesion of each pair of zones relative to the reference cohesion was calculated as an index of frequency domain response synchronization.

[0059] The noise removal module collects the voltage signal connected to the vacuum furnace power supply circuit at a sampling frequency of 10Hz, with the acquisition period synchronized with the real-time response spectrum recording period. After performing a Fourier transform on the voltage signal, frequency components with amplitudes exceeding one-tenth of the maximum amplitude are extracted to form the external noise master frequency set. Comparison reveals that the external noise master frequency set contains a 50Hz power frequency interference component. Further analysis of the amplitude consistency of the six heating zones at 50Hz reveals that the amplitudes of five heating zones are basically consistent and positively correlated with the amplitude of the 50Hz component in the voltage signal, while the amplitude of the central front zone at 50Hz is significantly higher than the other five heating zones, indicating that there is a superposition of the true thermal response signal and power frequency noise at 50Hz in the central front zone. Therefore, a noise contribution reduction operation is performed on the central front zone, subtracting 0.3 times the amplitude of the 50Hz component in the voltage signal from the real-time amplitude at this frequency to obtain the noise contribution estimate, while retaining the true thermal response component; for the other five heating zones, the amplitude component at this frequency is directly set to zero. The net deviation feature set is obtained after processing.

[0060] The diagnostic judgment module receives the net deviation feature set and the frequency domain response synchronization index. The net deviation values ​​of the six heating zones are 0.08, 0.12, 0.06, 0.09, 0.05, and 0.07, respectively. Arranged according to their spatial positions, a six-dimensional spatial distribution vector of furnace thermal deviation is constructed. The weighted Euclidean norm is calculated to obtain an overall deviation metric of 0.24. The synchronization degradation rates of five pairs of adjacent heating zones are 8%, 15%, 5%, 11%, and 7%, respectively. A five-dimensional spatial distribution vector of thermal coupling degradation is constructed, and the weighted Euclidean norm is calculated to obtain an overall coupling degradation metric of 0.31. The maximum net deviation value is 0.12, which is used as the deviation extreme value penalty. The maximum synchronization degradation rate is 15%, i.e., 0.15, which is used as the coupling extreme value penalty. Weighted summation is performed using preset weighting coefficients of 0.35, 0.35, 0.15, and 0.15 to obtain the comprehensive diagnostic index: D=0.35×0.24+0.35×0.31+0.15×0.12+0.15×0.15=0.233; The comprehensive diagnostic index of 0.233 was compared with the first threshold of 0.15 and the second threshold of 0.35. Since 0.15 < 0.233 ≤ 0.35, the diagnostic judgment module determined that the current temperature uniformity was showing a deteriorating trend and output a warning to the human-machine interface. Based on the warning information, combined with the net deviation values ​​of each heating zone and the spatial distribution of the rate of decrease in synchronicity in each area, the operator determined that the thermal response of the upper heating zone had significantly declined compared to the baseline state. The operator then scheduled maintenance on the upper heating elements and thermal shielding layer during the next production break, thus eliminating potential hazards before the macroscopic temperature difference exceeded the limit.

[0061] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting the uniformity of heating temperature in a vacuum furnace, characterized in that, include: S1. Under the uniform temperature reference state of the vacuum furnace, a sweeping pulse sequence is executed into the furnace to simultaneously acquire the power response signal of each heating zone and generate a multi-zone reference response spectrum. S2. During routine diagnosis, the frequency sweeping pulse sequence is reproduced under the same temperature conditions to synchronously acquire the power response signal of each heating zone and generate a multi-zone real-time response spectrum. S3. Perform frequency domain transformation on the multi-zone reference response spectrum and the multi-zone real-time response spectrum, extract the spectral characteristics of each heating zone in different frequency bands, and construct the frequency domain response synchronization index between adjacent heating zones. S4. Acquire the electrical signal of the external environment of the furnace body, extract its frequency domain features, and remove the deviation components caused by external environmental fluctuations from the spectrum features of step three to generate a net deviation feature set. S5. Integrate the net deviation feature set and the frequency domain response synchronization index to generate a comprehensive diagnostic index, and determine the current temperature uniformity status based on the comprehensive diagnostic index.

2. The method for detecting the uniformity of heating temperature in a vacuum furnace according to claim 1, characterized in that, In step S1, a sweeping pulse sequence is executed into the furnace to simultaneously acquire the power response signals of each heating zone and generate a multi-zone reference response spectrum, including: After the vacuum furnace passes the standard furnace temperature uniformity test, it is kept at a constant temperature under no-load at the target diagnostic temperature until it reaches a stable state, which is determined as the uniform temperature reference state. High-purity inert gas is periodically introduced into the furnace. The pressure peak value of each gas injection pulse is set to the same constant value. The duration and width of each gas injection pulse and the interval between adjacent gas injection pulses vary according to a preset frequency sweeping rule to form the frequency sweeping pulse sequence. The power response time-domain signal of each heating zone during the execution of the frequency sweep pulse sequence is recorded synchronously, and the multi-zone reference response spectrum is generated after normalization.

3. The method for detecting the uniformity of heating temperature in a vacuum furnace according to claim 2, characterized in that, In step S2, a multi-zone real-time response spectrum is generated, including: During routine diagnostics, the vacuum furnace is heated to the same temperature as the target diagnostic temperature and kept at that temperature. The frequency sweep pulse sequence is executed into the furnace again, and the rules for the variation of the pressure peak value, duration width, and interval time of the frequency sweep pulse sequence are the same as those in step S1. The power response time-domain signal of each heating zone during the execution of the frequency sweep pulse sequence is recorded synchronously, and the multi-zone real-time response spectrum is generated after normalization.

4. The method for detecting the uniformity of heating temperature in a vacuum furnace according to claim 1, characterized in that, In step S3, the spectral characteristics of each heating zone in different frequency bands are extracted, including: Fourier transforms are performed on the power response time-domain signals of each heating zone in the multi-zone reference response spectrum and the multi-zone real-time response spectrum to obtain the reference frequency domain data and real-time frequency domain data of each heating zone. The dividing frequency for frequency band division is determined based on the range of variation in duration and width and the range of variation in interval time of the swept pulse sequence. Based on the boundary frequency, low-frequency amplitude features and high-frequency amplitude features are extracted from the reference frequency domain data and real-time frequency domain data of each heating zone, respectively, as the spectral features of each heating zone in different frequency bands.

5. The method for detecting the uniformity of heating temperature in a vacuum furnace according to claim 4, characterized in that, In step S3, a frequency domain response synchronization index is constructed between adjacent heating zones, including: Two heating zones that are physically adjacent within the furnace are selected as a pair of areas to be evaluated. Based on the reference frequency domain data of each heating zone, the amplitude squared coherence function of the region to be evaluated in the full frequency band or a specified frequency band is calculated to obtain the coherence coefficient curve that varies with frequency, and the integral area of ​​the coherence coefficient curve in the corresponding frequency band is calculated as the reference cohesion. Based on the real-time frequency domain data of each heating zone, the amplitude square coherence function of the region to be evaluated in the full frequency band or a specified frequency band is calculated to obtain the coherence coefficient curve that varies with frequency, and the integral area of ​​the coherence coefficient curve in the corresponding frequency band is calculated as the real-time cohesion. The rate of decrease of the real-time cohesion relative to the baseline cohesion is used as the frequency domain response synchronization index.

6. The method for detecting the uniformity of heating temperature in a vacuum furnace according to claim 1, characterized in that, In step S4, a net deviation feature set is generated, including: The voltage signal connected to the power supply circuit of the vacuum furnace is collected as the external environmental electrical signal of the furnace body; Fourier transform is performed on the electrical signal of the external environment of the furnace body, and the frequency components whose amplitude exceeds a preset threshold are extracted to form the main frequency set of external noise; The component in the spectral characteristics of each heating zone whose frequency is the same as any frequency component in the set of dominant external noise frequencies is marked as a noise deviation component. The noise deviation component is removed from the spectral characteristics of each heating zone, and the remaining spectral characteristics after removal are used as the net deviation feature set.

7. The method for detecting the uniformity of heating temperature in a vacuum furnace according to claim 1, characterized in that, In step S5, a comprehensive diagnostic index is generated, including: Arrange the net deviation values ​​of each heating zone in the net deviation feature set according to their spatial location to construct the furnace thermal deviation spatial distribution vector, and calculate the weighted Euclidean norm of the furnace thermal deviation spatial distribution vector as the overall deviation metric. Arrange the synchronicity degradation rate of each adjacent heating zone pair in the frequency domain response synchronicity index according to spatial position to construct a thermal coupling degradation spatial distribution vector, and calculate the weighted Euclidean norm of the thermal coupling degradation spatial distribution vector as the overall metric of coupling degradation. The maximum value of the net deviation value of each heating zone in the net deviation feature set is taken as the deviation extreme value penalty amount; The maximum value of the synchronicity decrease rate of each adjacent heating zone pair in the frequency domain response synchronicity index is taken as the coupling extreme value penalty. The overall deviation metric, the overall coupling degradation metric, the deviation extreme value penalty, and the coupling extreme value penalty are weighted and summed according to preset weights to obtain the comprehensive diagnostic index.

8. The method for detecting the uniformity of heating temperature in a vacuum furnace according to claim 1, characterized in that, In step S5, the current temperature uniformity status is determined based on the comprehensive diagnostic index, including: The comprehensive diagnostic index is compared with a preset first threshold. If the comprehensive diagnostic index is less than or equal to the first threshold, the current temperature uniformity is determined to be normal. If the comprehensive diagnostic index is greater than the first threshold and less than or equal to the preset second threshold, the current temperature uniformity is determined to be deteriorating, and an early warning is output. If the comprehensive diagnostic index is greater than the second threshold, the current temperature uniformity is determined to be abnormal, and an alarm signal is output.

9. The method for detecting the uniformity of heating temperature in a vacuum furnace according to claim 1, characterized in that, Following step S5, the following is also included: Based on the net deviation values ​​of each heating zone in the net deviation feature set, with each heating zone as the radial axis and the corresponding net deviation value of each heating zone as the radial coordinate value, a real-time contour line of the current diagnostic cycle is drawn to form a furnace cavity thermal health radar map. In the furnace cavity thermal health radar map, the historical contour lines of the most recent historical diagnostic cycles are superimposed and displayed, with each historical contour line distinguished by different colors or line types. When the real-time coordinate value of any radial axis exceeds the preset statistical boundary of the historical coordinate value of that axis, a highlight mark is made on that radial axis; In the central area of ​​the furnace cavity thermal health radar map, the comprehensive diagnostic index and its corresponding temperature uniformity status indicator are displayed in numerical form.

10. A system for detecting the uniformity of heating temperature in a vacuum furnace, characterized in that, A method for detecting the heating temperature uniformity of a vacuum furnace according to any one of claims 1-9, the system comprising: The reference response generation module is used to execute a sweeping pulse sequence into the furnace when the vacuum furnace is in a uniform temperature reference state, and simultaneously acquire the power response signals of each heating zone to generate a multi-zone reference response spectrum. The real-time response generation module is used to reproduce the frequency sweeping pulse sequence under the same temperature conditions during routine diagnosis, synchronously acquire the power response signals of each heating zone, and generate a multi-zone real-time response spectrum. The frequency domain feature extraction module is used to perform frequency domain transformation on the multi-zone reference response spectrum and the multi-zone real-time response spectrum, extract the spectral features of each heating zone in different frequency bands, and construct the frequency domain response synchronization index between adjacent heating zones. The noise removal module is used to acquire the electrical signal of the external environment of the furnace body, extract its frequency domain features, and remove the deviation components caused by external environmental fluctuations from the spectrum features extracted by the frequency domain feature extraction module to generate a net deviation feature set. The diagnostic judgment module is used to integrate the net deviation feature set and the frequency domain response synchronization index to generate a comprehensive diagnostic index, and to determine the current temperature uniformity status based on the comprehensive diagnostic index.

Images