Fault prediction method and device for high-low temperature box, storage medium and electronic equipment

Abstract

The invention discloses a fault prediction method and device for a high and low temperature box, a storage medium and electronic equipment, and the method comprises the steps: obtaining the original operation data of the high and low temperature box, and the original operation data at least comprises the state information, humidity and temperature information and multi-dimensional vibration information of the high and low temperature box; based on the original operation data, the temperature box working condition state of the high-low temperature box is determined according to working condition identification logic; and based on a set matching rule, matching different building model algorithms for different temperature box working condition states to perform root mean square calculation so as to output health state scores of the high and low temperature box in the temperature box working condition states. According to the embodiment of the invention, the method also has the capability of real-time monitoring and dynamic adjustment while carrying out intelligent fault prediction, not only can prejudge the equipment abnormality in advance, but also can dynamically respond and intervene potential risks in the operation process, achieves the integrated cooperation of intelligent prediction and operation and maintenance control, and remarkably improves the operation efficiency of the equipment and the toughness of a production system.

Description

Technical Field

[0001] This disclosure relates to the technical field of fault detection, and in particular to a fault prediction method, apparatus, storage medium and electronic equipment for high and low temperature chambers. Background Technology

[0002] High and low temperature chambers are essential environmental testing equipment widely used in industries such as electronics, materials, and automobiles. They are primarily used to test the adaptability of products or components to different temperature and humidity environments. Their operational stability directly affects the continuity and reliability of testing, thus impacting the entire product development and quality verification process. However, in practical applications, due to factors such as prolonged operation, frequent start-ups and shutdowns, and environmental stress, high and low temperature chambers are prone to malfunctions such as compressor abnormalities, temperature control failures, and sensor drift. In severe cases, these malfunctions can even lead to test interruptions and sample damage.

[0003] Currently, fault management in high and low temperature chambers mainly relies on periodic manual inspections and experience-based maintenance strategies. These methods suffer from drawbacks such as high latency, high misjudgment rates, and high maintenance costs, making it difficult to achieve early identification and predictive maintenance of potential faults. Furthermore, traditional systems often lack the ability to comprehensively collect and analyze multi-dimensional operational data, failing to integrate and deeply analyze information such as temperature and humidity data, vibration signals, and equipment operating status in real time, thus limiting the possibility of intelligent equipment condition diagnosis and accurate prediction.

[0004] In addition, although some existing monitoring solutions have integrated sensor acquisition or remote monitoring modules, their data processing capabilities are still limited to threshold alarms or static rule judgments. They cannot adapt to the complexity and variability of equipment operating conditions, nor do they have the ability to model and learn from historical data. They are slow to respond when faced with changes in operating conditions or equipment aging trends, making it difficult to meet the needs of modern manufacturing for high availability and intelligent management of equipment. Summary of the Invention

[0005] This disclosure aims to at least partially address one of the technical problems in the related art.

[0006] Therefore, the first objective of this disclosure is to propose a fault prediction method for high and low temperature chambers, comprising: acquiring raw operating data of the high and low temperature chamber, wherein the raw operating data includes at least the state information, humidity and temperature information, and multidimensional vibration information of the high and low temperature chamber; determining the chamber operating condition state of the high and low temperature chamber based on the raw operating data according to operating condition identification logic; and matching different model building algorithms to different chamber operating conditions based on set matching rules to perform root mean square calculation to output a health status score of the high and low temperature chamber under the chamber operating condition state.

[0007] In some embodiments, after acquiring the raw operating data of the high and low temperature chamber, the method further includes: performing preprocessing operations on the raw operating data, wherein the preprocessing operations include at least data filtering, data cleaning, data standardization and structuring.

[0008] In some embodiments, before determining the temperature chamber operating condition status of the high and low temperature chamber based on the original operating data according to the operating condition identification logic, the method further includes: identifying different operating processes of the high and low temperature chamber and classifying them into typical operating condition statuses of predetermined types according to the identification results.

[0009] In some embodiments, identifying different operating processes of the high and low temperature chamber and classifying them into predetermined types of typical operating conditions based on the identification results includes: sliding window segmentation; feature extraction and construction of operating condition feature vectors; operating condition feature segmentation; and real-time operating condition identification and classification.

[0010] In some embodiments, the feature extraction and construction of the operating condition feature vector involves extracting time-domain feature indicators, frequency-domain feature indicators, and time-frequency joint features from key monitoring parameters within each sliding window and constructing a feature vector for operating condition identification. The time-domain feature indicators include at least one of waveform factor, impulse factor, kurtosis factor, and margin factor. The frequency-domain feature indicators include at least one of spectral features based on Fast Fourier Transform, characteristic frequency components, frequency band energy distribution, and dominant frequency amplitude. The feature vector includes at least one of temperature change rate, humidity change trend, root mean square of vibration signal, kurtosis, margin, frequency domain energy distribution, and equipment status flag.

[0011] In some embodiments, the step of matching different model building algorithms to different incubator operating conditions based on set matching rules to perform root mean square (RMS) calculation and output the health status score of the high and low temperature chamber under the incubator operating conditions includes: a preset operating condition-model mapping rule table; performing RMS feature calculation on the collected multidimensional vibration information under different operating conditions; and dynamically calling the matched prediction model to perform real-time inference based on the RMS features within the current window according to the currently identified incubator operating conditions, and outputting the health status score.

[0012] In some embodiments, the method further includes: periodically summarizing historical running data and model output results, re-extracting features and retraining the model, and achieving dynamic optimization of model parameters.

[0013] A second aspect of this disclosure is to provide a fault prediction device for high and low temperature chambers, comprising:

[0014] The acquisition module is used to acquire the raw operating data of the high and low temperature chamber, which includes at least the status information, humidity and temperature information and multidimensional vibration information of the high and low temperature chamber.

[0015] The determination module is used to determine the temperature chamber operating status of the high and low temperature chamber based on the original operating data and the operating condition identification logic;

[0016] The output module is used to match different model building algorithms to different incubator operating conditions based on the set matching rules, and to perform root mean square calculation to output the health status score of the high and low temperature chamber under the incubator operating conditions.

[0017] To achieve the above objectives, a third aspect of this disclosure provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the detection method described in the first aspect.

[0018] To achieve the above objectives, a fourth aspect of this disclosure provides an electronic device, including an electronic device comprising at least a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the program, it implements the detection method described in the first aspect above.

[0019] This disclosure utilizes the high-dimensional features of a network model to distinguish between lane lines and road edges, significantly improving the accuracy of target detection and providing support for subsequent precise and rapid perception. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this disclosure or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of one of the steps of a fault prediction method for a high and low temperature chamber according to an embodiment of the present disclosure. Detailed Implementation

[0022] Embodiments of this disclosure are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting this disclosure.

[0023] The first embodiment of this disclosure provides a fault prediction method for a high and low temperature chamber. The high and low temperature chamber is a device used to simulate different temperature environments to test the adaptability and reliability of products under high and low temperature conditions. This embodiment is used to predict faults in the high and low temperature chamber. Figure 1 As shown, the fault prediction method includes the following steps:

[0024] S101, acquire the raw operating data of the high and low temperature chamber, the raw operating data including at least the status information, humidity and temperature information and multidimensional vibration information of the high and low temperature chamber.

[0025] In this step, the raw operating data of the high and low temperature chamber is first collected and / or gathered. At least the data in the controller of the high and low temperature chamber can be obtained, such as the equipment status information of the high and low temperature chamber. Different operating information of the high and low temperature chamber can also be collected from multiple key monitoring points through different acquisition devices. For example, the temperature and humidity information inside the high and low temperature chamber can be collected through a humidity and temperature sensor, and the multi-dimensional vibration information of the compressor in the high and low temperature chamber can be collected through a vibration sensor. The multi-dimensional vibration information here includes, for example, vibration velocity, acceleration, displacement and other information.

[0026] In this embodiment, the main raw operating data of the high and low temperature chamber during operation are collected, gathered, and organized. This raw operating data includes, but is not limited to, the status information of the high and low temperature chamber, its temperature and humidity information, and multidimensional vibration information within the chamber. Specific methods may include: acquiring the status information of the high and low temperature chamber in real time from its controller based on a communication protocol, such as start / stop flags, control modes, and operating time; calling the interface of a temperature and humidity sensor to collect temperature and humidity information within the chamber, as well as the trend information of these changes under different operating conditions; collecting multi-axis vibration data of the compressor within the high and low temperature chamber during operation using a high-frequency vibration acquisition device, wherein the sampling frequency of the high-frequency vibration acquisition device is 100Hz; and storing the collected raw operating data in a database or a data analysis framework such as Pandas as the basis for subsequent model building and predictive analysis.

[0027] It should be noted that the data collected or acquired above covers a variety of structured and semi-structured formats, including real-time sensor access interfaces (such as Modbus / PLC protocol), CSV export file formats, and industrial database structures (such as PostgreSQL or Influx DB), etc., and has good data compatibility and scalability.

[0028] Furthermore, in actual operation, after the collection of raw operating data is completed, the raw operating data is preprocessed based on SQL query language or Pandas data analysis library. The purpose is to generate a dataset with a clear structure, standardized fields, and usable for subsequent modeling and analysis. The above preprocessing operations include at least data filtering, data cleaning, data standardization and structuring, so as to ensure that the input foundation for subsequent working condition identification and predictive analysis is stable and reliable.

[0029] The data filtering described here involves extracting data records relevant to the current task from multi-source data (such as temperature and humidity information, status information, and vibration information) collected by different acquisition devices of the high and low temperature chamber, based on conditions such as task time range, vibration signal, and unique equipment identifier, to initially construct an analytical data subset. The data cleaning described here includes, but is not limited to, outlier removal, virtual data completion, and multi-source data conflict resolution models. The outlier removal model uses the median absolute variance method to improve the efficiency of outlier identification. The virtual data completion model uses an optimized AR model to extrapolate and generate virtual data to fill in missing values. The multi-source data conflict resolution model uses support vector regression to extract a smooth and continuous trend function from the monitoring data, and uses normalized cross-correlation analysis to measure the similarity between trends, excluding invalid monitoring data based on the similarity.

[0030] The cleaned data will be stored in a structured form in a database or intermediate data container (such as PandasDataFrame) to form a standardized preprocessed dataset, providing high-quality input data support for subsequent feature extraction, working condition identification, and fault prediction model training. The data standardization and structuring mentioned here refer to data standardization and field specification. Fields from different sources are processed uniformly to ensure consistent field names and standardized data formats. Through the above operations, a structured preprocessed dataset can be generated.

[0031] S102, Based on the original operating data, determine the temperature chamber operating status of the high and low temperature chamber according to the operating condition identification logic.

[0032] After obtaining the original operating data of the high and low temperature chamber through the above step S101, the original operating data includes at least the status information, humidity and temperature information and multidimensional vibration information of the high and low temperature chamber. In this step, the temperature chamber operating condition is determined based on the original operating data according to the operating condition identification logic. The temperature chamber operating condition is one of the preset typical operating conditions.

[0033] Prior to this step, the different operating processes of the high and low temperature chamber are identified and classified into predetermined types of typical operating conditions based on the identification results. Here, the identification of typical operating conditions is based on the historical operating data of the high and low temperature chamber. Specifically, this includes extracting features from the operating process of the high and low temperature chamber using a sliding window to obtain multi-domain feature samples, and combining this with a random decision tree classification model to dynamically identify and classify the multi-source data acquired during the operation of the high and low temperature chamber, thereby classifying different operating processes into predetermined types of typical operating conditions and setting corresponding operating condition labels.

[0034] Based on the historical operating data and corresponding operating condition labels of the high and low temperature chamber, a random decision tree classification model is trained using multi-domain feature samples. After the random decision tree classification model is deployed, the latest feature input within a real-time sliding window is received during operation, and the inference of operating condition classification is automatically completed, thereby outputting the corresponding operating condition status label. In a specific implementation, according to the set operating condition identification logic and data-driven classification algorithm, the operating process of the high and low temperature chamber is dynamically identified and segmented into operating conditions. By constructing an analysis strategy based on a sliding window and a trained random decision tree classification model, the operating process of the chamber is divided into four typical operating conditions, such as standby condition, stable operation condition, rapid heating condition, and rapid cooling condition.

[0035] The specific steps for the operating condition identification logic to identify the historical operating data of the high and low temperature chamber include:

[0036] The sliding window segmentation process involves segmenting the continuously collected raw operating data (such as temperature, humidity, vibration velocity, acceleration, displacement, etc.) during the operation of the high and low temperature chamber into segments according to a set time window.

[0037] Feature extraction and construction of operating condition feature vectors involve extracting time-domain and frequency-domain feature indicators for key monitoring parameters within each sliding window and constructing feature vectors for operating condition identification. The extracted time-domain feature indicators include typical dimensionless time-domain features such as waveform factor, impulse factor, kurtosis factor, and margin factor. The extracted frequency-domain feature indicators include spectral features, characteristic frequency components, frequency band energy distribution, and dominant frequency amplitude based on Fast Fourier Transform (FFT). Furthermore, to address the characteristics of non-stationary signals, time-frequency analysis methods such as Short-Time Fourier Transform (STFT) and wavelet transform are introduced to extract joint time-frequency features.

[0038] In addition, the feature vectors constructed for operating condition identification include temperature change rate, humidity change trend, root mean square (RMS) of vibration signal, kurtosis, margin and other statistics, frequency domain energy distribution (such as FFT main frequency band amplitude), equipment status flags (such as whether it is started or stopped, operating mode, etc.).

[0039] Operating condition feature segmentation refers to segmenting operating condition features based on existing historical operating data and pre-labeled operating condition types, such as using Random Forest.

[0040] (4) Real-time operating condition identification and classification judgment, that is, based on the combination of feature vectors, such as temperature change rate, vibration characteristics, etc., the judgment is made and the operating condition label representing the operating condition of the temperature chamber in the current window is output.

[0041] S103, based on the set matching rules, different model building algorithms are matched for different incubator operating conditions to perform root mean square calculation and output the health status score of the high and low temperature chamber under the incubator operating conditions.

[0042] After setting the operating condition identification logic through the above step S102 and identifying different operating processes of the high and low temperature chamber according to the operating condition identification logic, and classifying them into typical operating condition states of a predetermined type according to the identification results, in this step, different construction model algorithms are matched to different chamber operating condition states based on the set matching rules to perform root mean square calculation to output the health status score of the high and low temperature chamber under the chamber operating condition state.

[0043] After determining the operating condition of the high and low temperature chamber, this step involves setting multi-condition matching rules based on the identified operating condition, and employing differentiated modeling strategies for different operating conditions. By matching applicable algorithm models, in-depth analysis of key vibration data and fault trend identification of the high and low temperature chamber are achieved. Specifically, this includes:

[0044] (1) A preset working condition-model mapping rule table is used to match the most suitable analysis method and prediction model type based on the operating characteristics and data patterns of various temperature chamber working conditions and the working condition matching algorithm. The working condition matching algorithm module is used to match different working conditions. Based on the matching rules, the optimal model construction strategy is matched for different working conditions. The root mean square feature operation is performed on the three dimensions of the collected multidimensional vibration information, namely vibration velocity, acceleration and displacement, respectively, to aggregate the data within the time period (per second) and select the most suitable machine learning model construction strategy.

[0045] (2) The root mean square feature calculation is performed on the three dimensions of the multidimensional vibration information collected under different working conditions, namely vibration velocity, acceleration and displacement, to aggregate the data within the time period (per second) and select the most suitable machine learning model construction strategy. Among them, the root mean square (RMS) calculation is performed on the multidimensional vibration information collected under each temperature chamber working condition according to the time window to obtain representative indicators reflecting vibration energy.

[0046] Specifically, the preprocessed raw data undergoes root mean square (RMS) calculation. First, through data aggregation and dimensionality reduction, and addressing the needs of multivariate feature extraction, feature selection and fusion tools such as principal component analysis and nonnegative matrix factorization are developed to process feature extraction from continuous time-series signals such as vibration velocity, acceleration, and displacement. The RMS calculation method is as follows:

[0047] For a discrete signal sequence within one sampling period, the root mean square value is calculated according to the following formula:

[0048]

[0049] The sampled signal is segmented by a sliding window, and an RMS value is calculated within each window. This transforms the original waveform signal into a statistical feature sequence that reflects its energy change trend. This feature is suitable for reflecting the energy level changes of vibration signals under different working conditions, which helps to improve the accuracy of fault prediction models.

[0050] In this step, considering the need for multivariate feature extraction from the original running data, principal component analysis is developed for feature extraction and compression.

[0051] In actual engineering applications, the signals obtained often contain noise. To effectively identify the status and faults of the high and low temperature chamber, it is necessary to judge and identify them using feature information from multiple physical quantities. When the number of feature information used for identification is too large, it is equivalent to judging the status or faults of the high and low temperature chamber in a high-dimensional space. To simplify the judgment process, it is necessary to compress multiple identification features to achieve data dimensionality reduction.

[0052] Principal component analysis (PCA) is a data dimensionality reduction technique that transforms a large number of correlated variables into a small set of uncorrelated variables, called principal components. The calculation steps are as follows:

[0053] (1): For the original feature vector After zero-mean normalization, we obtain ;

[0054] (2): Calculate the covariance matrix ;

[0055] (3): Solve for the matrix The characteristic equation is used to obtain the eigenvalues, which are then arranged in descending order. The corresponding feature vector is , , , ;

[0056] (4): Pre-set the compression dimension p, or according to the given cumulative contribution rate requirements. Find the satisfying The minimum value of p is obtained by standardizing the first p eigenvectors. ;

[0057] (5): Finally, the vector obtained after zero mean normalization is... Projecting onto each standardized feature vector yields the compressed feature vector. .

[0058] The first principal component is a weighted combination of k observed variables, which has the greatest explanatory power for the variance of the initial variable set. The second principal component is also a linear combination of the initial variables, which has the second highest explanatory power for the variance and is orthogonal to the first principal component (uncorrelated). Each subsequent principal component maximizes its explanatory power for the variance and is orthogonal to all previous principal components. Considering the goal of using fewer principal components to explain all variables, the feature description of the target object can be completed by selecting a smaller number of earlier components, thereby achieving dimensionality reduction and extracting effective features. The optimal model parameters for the equipment are then set according to different operating conditions.

[0059] (3) Based on the currently identified incubator operating condition, the matching prediction model is dynamically invoked, and real-time inference is performed based on the root mean square features within the current window to output a health status score under the incubator operating condition. The prediction model used here is based on a one-class support vector machine (SVM), which is suitable for anomaly detection and health score calculation under a single operating condition. In addition, the model results under the above multiple operating conditions can be superimposed to generate a unified health assessment report, which serves as the basis for subsequent early warning and maintenance recommendations.

[0060] Specifically, based on the constructed support vector machine (SVM) model as a baseline model, SVM model operations are performed in real time. The prediction results of each SVM model are statistically analyzed, and the health status is calculated in real time to output the health result of the high and low temperature chamber. In this embodiment, by collecting the operating data of the high and low temperature chamber in real time and constructing a model application, a corresponding baseline model is constructed according to different typical operating conditions, thereby feeding the operating data into the model in real time and outputting health parameters.

[0061] This step uses support vector regression as the basic method to calculate and output health parameters in real time. It includes: using support vector regression as the basic tool to establish a trend prediction model for equipment status, forming a trend fitting and time extrapolation technique with nonlinearity, high robustness and small sample learning characteristics, and adjusting the model parameters in combination with monitoring feature performance to obtain good prediction results.

[0062] Specifically, based on the output of the support vector regression module, i.e., the regression function, extrapolation is performed to calculate the possible trends of monitored quantities, thus allowing for advance understanding of the equipment's operating status. Based on the input prediction parameters, interpolation yields the time series data. ,in The initial regression prediction time is... For the regression termination time, To predict the termination time, the time intervals are equal, determined based on the prediction requirements. Inputting the regression function yields the predicted vibration trend, which can then be plotted as a curve. Extrapolation based on the regression function output by the support vector regression module and the prediction curve output by the trend prediction module can yield the vibration amount at any future time. However, if the prediction period is too long, the prediction accuracy obviously decreases, and its reference value becomes limited. Let the prediction period be PredictDay (initial value PredictDay=30), and set the vibration threshold Thr. The trend prediction module calculates and outputs the state values ​​for the future PredictDay time period, and iteratively searches, where i=1 indicates... The first prediction time is when a certain time... When the predicted value of the vibration is greater than or equal to the threshold Thr (1≤i≤PredictDay), output... Then the first time the threshold is reached, FAT is... If i > 30 and the threshold is still not reached, output "Continuing to predict is meaningless".

[0063] Based on the above model principles and methods, the model is constructed to collect data such as vibration velocity, acceleration, and displacement of the incubator equipment in real time. The constructed support vector machine model is used as the benchmark model to perform support vector machine model calculations in real time, and the prediction results of each support vector machine model are statistically analyzed. The health status is calculated in real time and the health result is output.

[0064] In this embodiment, the structure of the high and low temperature chamber needs to be modified. Acquisition devices are installed at key locations within the chamber to collect operational data from the chamber and its internal vacuum furnace and vibration table. In this embodiment, the verified data acquisition methods include PLC acquisition, equipment control program communication acquisition, OCR recognition for video stream acquisition, and shared file acquisition. This allows for real-time acquisition and collection of the chamber's status information, temperature and humidity information, acceleration, displacement, and other multi-dimensional vibration information. Furthermore, the aforementioned multi-channel data acquisition mechanism enables real-time acquisition of key operational data. The acquired data stream passes through a buffer before entering the real-time feature processing pipeline. Specifically, relying on the aforementioned multi-channel acquisition mechanism, comprehensive real-time monitoring of key operational data is possible. The acquired data is first buffered through an IoT low-code platform and then flows into the real-time feature processing pipeline.

[0065] The data processing end involved in this embodiment is based on a low-code visualization platform to build a data analysis process. Through modular configuration, it realizes operations such as real-time data cleaning, sliding window segmentation, RMS operation, and time domain / frequency domain / time-frequency joint feature extraction. At the same time, the data processing process is integrated with a pre-trained machine learning model (such as a support vector machine model) to ensure that feature data can be input into the model in each sampling period to complete fault prediction or health assessment tasks.

[0066] In addition, in this embodiment, the fault prediction method also includes: periodically summarizing historical operating data and model output results, re-extracting features and training the model, realizing dynamic optimization of model parameters, and finally realizing adaptive updating and long-term stable operation of the model. This process supports both automatic iteration and manual intervention mechanisms to ensure the continuous and stable operation of the prediction system. The specific steps include: periodically summarizing and archiving historical operating data, combining the latest real-time data, and carrying out model retraining and continuous iterative optimization; (2) by introducing incremental learning or transfer learning and other technologies, the model's adaptive ability to environmental changes and equipment status fluctuations is improved, realizing dynamic adjustment of model parameters and performance improvement, thereby ensuring the long-term stable operation and high-efficiency performance of the model in complex environments.

[0067] The above process ensures that the model can continuously capture new features and potential anomalies in the system operation, avoid model aging or performance degradation, and thus support the accuracy and reliability of tasks such as fault prediction and health assessment, ensuring the long-term stable operation and continuous optimization of the model in the actual application environment.

[0068] This disclosed embodiment integrates artificial intelligence algorithms, possesses real-time multi-source data acquisition capabilities, and can dynamically identify equipment operating conditions and predict potential faults, thereby improving the operational reliability and maintenance efficiency of high and low temperature chamber equipment, realizing the transformation from "post-event maintenance" to "pre-event prediction," and providing a solid equipment support foundation for intelligent manufacturing.

[0069] Furthermore, in this embodiment, the fault prediction method is not only used for fault prediction modeling, but also undertakes the function of real-time monitoring and dynamic intervention of the operating status of the high and low temperature chamber, so as to enhance the adaptability and robustness to the actual operating environment. The specific steps include: real-time perception of the operating status of the high and low temperature chamber, dynamic diagnosis and health update, prediction-driven maintenance suggestions and scheduling intervention, and linkage with the IoT industrial Internet of Things platform to realize closed-loop linkage from prediction and judgment to production, thereby improving the overall production safety and continuity.

[0070] This disclosed embodiment not only performs intelligent fault prediction, but also has the ability of "real-time monitoring + dynamic adjustment". It can not only predict equipment abnormalities in advance, but also dynamically respond to and intervene in potential risks during operation, realizing the integrated collaboration of intelligent prediction and operation and maintenance control, and significantly improving equipment operating efficiency and the resilience of the production system.

[0071] Based on the same inventive concept as the first embodiment described above, the second embodiment of this disclosure provides a target detection device, which includes a mutually coupled acquisition module, a determination module, and an output module, wherein:

[0072] The acquisition module is used to acquire the raw operating data of the high and low temperature chamber, which includes at least the status information, humidity and temperature information, and multidimensional vibration information of the high and low temperature chamber.

[0073] The determining module uses the original operating data and operating condition identification logic to determine the operating condition status of the high and low temperature chamber.

[0074] The output module is used to match different model building algorithms to different incubator operating conditions based on the set matching rules, perform root mean square calculation, and output the health status score of the high and low temperature chamber under the incubator operating conditions.

[0075] Furthermore, it also includes a preprocessing module, which is used to preprocess the raw running data. The preprocessing operations include at least data filtering, data cleaning, data standardization and structuring.

[0076] Furthermore, it also includes a classification module, which is used to identify different operating processes of the high and low temperature chamber and classify them into typical operating conditions of a predetermined type based on the identification results.

[0077] Furthermore, the partitioning module includes: a first processing unit for sliding window segmentation; a second processing unit for feature extraction and construction of working condition feature vectors; a third processing unit for working condition feature segmentation; and a fourth processing unit for real-time working condition identification and classification.

[0078] Furthermore, the second processing unit is specifically used to extract time-domain feature indicators, frequency-domain feature indicators, and time-frequency joint features from key monitoring parameters within each sliding window and construct a feature vector for operating condition identification. The time-domain feature indicators include at least one of waveform factor, impulse factor, kurtosis factor, and margin factor. The frequency-domain feature indicators include at least one of spectral features based on fast Fourier transform, characteristic frequency components, frequency band energy distribution, and dominant frequency amplitude. The feature vector includes at least one of temperature change rate, humidity change trend, root mean square of vibration signal, kurtosis, margin, frequency domain energy distribution, and equipment status flag bit.

[0079] Furthermore, the output module includes: a fifth processing unit for presetting a working condition-model mapping rule table; a sixth processing unit for performing root mean square feature calculation on the collected multidimensional vibration information according to working conditions; and a seventh processing unit for dynamically calling the matched prediction model based on the currently identified temperature chamber working condition to perform real-time inference based on the root mean square features within the current window and outputting a health status score.

[0080] Furthermore, it also includes an eighth processing unit, which is used to periodically summarize historical running data and model output results, re-extract features and re-train the model, and realize dynamic optimization of model parameters.

[0081] This disclosed embodiment not only performs intelligent fault prediction, but also has the ability of "real-time monitoring + dynamic adjustment". It can not only predict equipment abnormalities in advance, but also dynamically respond to and intervene in potential risks during operation, realizing the integrated collaboration of intelligent prediction and operation and maintenance control, and significantly improving equipment operating efficiency and the resilience of the production system.

[0082] The third embodiment of this disclosure provides a storage medium, which is a computer-readable medium storing a computer program. When executed by a processor, the computer program implements the method provided in the first embodiment of this disclosure, including the following steps S11 to S13:

[0083] S11, acquire the raw operating data of the high and low temperature chamber, the raw operating data including at least the status information, humidity and temperature information and multidimensional vibration information of the high and low temperature chamber;

[0084] S12, Based on the original operating data, determine the temperature chamber status of the high and low temperature chamber according to the operating condition identification logic;

[0085] S13, based on the set matching rules, different model building algorithms are matched for different incubator operating conditions to perform root mean square calculation and output the health status score of the high and low temperature chamber under the incubator operating conditions.

[0086] Furthermore, when the computer program is executed by the processor, it implements other methods provided in the first embodiment of this disclosure.

[0087] This disclosed embodiment not only performs intelligent fault prediction, but also has the ability of "real-time monitoring + dynamic adjustment". It can not only predict equipment abnormalities in advance, but also dynamically respond to and intervene in potential risks during operation, realizing the integrated collaboration of intelligent prediction and operation and maintenance control, and significantly improving equipment operating efficiency and the resilience of the production system.

[0088] A fourth embodiment of this disclosure provides an electronic device, which includes at least a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program in the memory, it implements the method provided in any embodiment of this disclosure. Exemplarily, the steps of the electronic device's computer program are as follows: S21 to S23:

[0089] S21, acquire the raw operating data of the high and low temperature chamber, the raw operating data including at least the status information, humidity and temperature information and multidimensional vibration information of the high and low temperature chamber;

[0090] S22, Based on the original operating data, determine the temperature chamber status of the high and low temperature chamber according to the operating condition identification logic;

[0091] S23, based on the set matching rules, different model building algorithms are matched for different incubator operating conditions to perform root mean square calculation and output the health status score of the high and low temperature chamber under the incubator operating conditions.

[0092] Furthermore, the processor also executes the computer program described in the third embodiment above.

[0093] This disclosed embodiment not only performs intelligent fault prediction, but also has the ability of "real-time monitoring + dynamic adjustment". It can not only predict equipment abnormalities in advance, but also dynamically respond to and intervene in potential risks during operation, realizing the integrated collaboration of intelligent prediction and operation and maintenance control, and significantly improving equipment operating efficiency and the resilience of the production system.

[0094] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0095] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this disclosure, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0096] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of this disclosure includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of this disclosure pertain.

[0097] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0098] It should be understood that various parts of this disclosure can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0099] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0100] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0101] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present disclosure have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present disclosure. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present disclosure.

Claims

1. A fault prediction method for high and low temperature chambers, characterized in that, include: Acquire raw operating data of the high and low temperature chamber, wherein the raw operating data includes at least the state information, humidity and temperature information, and multidimensional vibration information of the high and low temperature chamber; Based on the original operating data, the operating condition status of the high and low temperature chamber is determined according to the operating condition identification logic; Based on the set matching rules, different model building algorithms are matched to different incubator operating conditions to perform root mean square calculation and output the health status score of the high and low temperature chamber under the incubator operating conditions.

2. The fault prediction method according to claim 1, characterized in that, After obtaining the raw operating data of the high and low temperature chamber, the method further includes: performing preprocessing operations on the raw operating data, the preprocessing operations including at least data filtering, data cleaning, data standardization and structuring.

3. The fault prediction method according to claim 1, characterized in that, Before determining the temperature chamber operating condition status of the high and low temperature chamber based on the original operating data and the operating condition identification logic, the method further includes: identifying different operating processes of the high and low temperature chamber and classifying them into typical operating condition statuses of predetermined types according to the identification results.

4. The fault prediction method according to claim 3, characterized in that, The process of identifying different operating processes of the high and low temperature chamber and classifying them into predetermined types of typical operating conditions based on the identification results includes: sliding window segmentation; feature extraction and construction of operating condition feature vectors; operating condition feature segmentation; and real-time operating condition identification and classification judgment.

5. The fault prediction method according to claim 4, characterized in that, The feature extraction and construction of the operating condition feature vector involves extracting time-domain feature indicators, frequency-domain feature indicators, and time-frequency joint features from key monitoring parameters within each sliding window and constructing a feature vector for operating condition identification. The time-domain feature indicators include at least one of waveform factor, impulse factor, kurtosis factor, and margin factor. The frequency-domain feature indicators include at least one of spectral features based on Fast Fourier Transform, characteristic frequency components, frequency band energy distribution, and dominant frequency amplitude. The feature vector includes at least one of temperature change rate, humidity change trend, root mean square of vibration signal, kurtosis, margin, frequency domain energy distribution, and equipment status flag.

6. The fault prediction method according to claim 1, characterized in that, The method of matching different model building algorithms based on the set matching rules for different incubator operating conditions to perform root mean square calculation and output the health status score of the high and low temperature chamber under the incubator operating conditions includes: Preset working conditions - model mapping rule table; Root mean square feature calculation is performed on the collected multidimensional vibration information under different working conditions. Based on the currently identified incubator operating condition, the matching prediction model is dynamically invoked to perform real-time inference based on the root mean square features within the current window, and a health status score is output.

7. The fault prediction method according to claim 6, characterized in that, Also includes: Periodically summarize historical operating data and model output results, re-extract features and retrain the model to achieve dynamic optimization of model parameters.

8. A fault prediction device for a high and low temperature chamber, characterized in that, include: The acquisition module is used to acquire the raw operating data of the high and low temperature chamber, which includes at least the status information, humidity and temperature information and multidimensional vibration information of the high and low temperature chamber. The determination module is used to determine the temperature chamber operating status of the high and low temperature chamber based on the original operating data and the operating condition identification logic; The output module is used to match different model building algorithms to different incubator operating conditions based on the set matching rules, and to perform root mean square calculation to output the health status score of the high and low temperature chamber under the incubator operating conditions.

9. A storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the detection method as described in any one of claims 1-8.

10. An electronic device, characterized in that, The device includes an electronic device, which includes at least a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements the detection method as described in any one of claims 1-8.

Images