Sparkling water machine remote monitoring and fault prediction method and system

Abstract

The present invention relates to the field of fault prediction technology, and in particular to a method and system for remote monitoring and fault prediction of a bubble water machine. The method comprises the following steps: obtaining the historical operation log and real-time working monitoring image of the bubble water machine; performing multi-stage operating state feature analysis based on the historical operation log of the bubble water machine, and generating multi-stage operating state time series change characteristic data; performing end-to-end change tracking of normalized operation based on the multi-stage operating state time series change characteristic data, and performing global dynamic portrait evolution to construct a normalized operation global portrait; performing multi-time-frequency decomposition of the bubble water machine operation vibration audio signal and bubble water machine operation state evolution, thereby obtaining the operation state vibration evolution characteristics; performing current operation stage identification on the real-time working monitoring image, thereby obtaining the carbonization state characteristics of the current stage. The present invention realizes efficient and accurate bubble water machine fault prediction and improves fault diagnosis performance.

Description

Technical Field

[0001] The present invention relates to the technical field of fault prediction, and in particular to a method and system for remote monitoring and fault prediction of a sparkling water machine. Background Art

[0002] As an important equipment widely used in beverage manufacturing, food processing and other industries, the sparkling water machine is mainly used to produce sparkling water or carbonated beverages. Its stability and reliability directly affect production efficiency and product quality. During the operation of the sparkling water machine, key components of the equipment such as the compressor, gas mixing system, control system, etc., are responsible for the driving and control tasks required to produce sparkling water. With the continuous improvement of the industry's requirements for production efficiency and product quality, the degree of automation of the sparkling water machine is constantly improving, and more and more sparkling water machines are widely used in large-scale production lines and intelligent factories.

[0003] However, during the long-term operation of the sparkling water machine, the equipment may be affected by a variety of internal and external factors, such as mechanical wear, improper operation, environmental changes, sensor failure, etc. These factors may cause equipment failure or performance degradation, affecting the continuity and stability of production. Traditional sparkling water machine fault diagnosis and maintenance methods mainly rely on manual inspections, regular maintenance, and on-site inspections, which have defects such as late problem discovery and slow maintenance response, and lack of real-time monitoring and prediction capabilities of equipment operation status.

[0004] Therefore, in order to ensure that the sparkling water machine can operate continuously and stably, and to detect and handle the risk of failure in a timely manner, a more intelligent remote monitoring and fault prediction method for the sparkling water machine is urgently needed. Summary of the invention

[0005] In order to solve the above technical problems, the present invention proposes a method and system for remote monitoring and fault prediction of a sparkling water machine to solve at least one of the above technical problems.

[0006] To achieve the above object, the present invention provides a method for remote monitoring and fault prediction of a sparkling water machine, comprising the following steps: Step S1: Obtain the historical operation log and real-time working monitoring image of the bubble water machine; perform multi-stage operation state characteristic analysis based on the historical operation log of the bubble water machine to generate multi-stage operation state time series change characteristic data; Step S2: Based on the multi-stage operation status time series change characteristic data, end-to-end change tracking of normalized operation is performed, and global dynamic portrait evolution is performed to construct a global portrait of normalized operation; Step S3: monitoring the bubble water machine operation vibration audio signal in real time based on the sensor; performing multi-time-frequency decomposition on the bubble water machine operation vibration audio signal and the bubble water machine operation state evolution, thereby obtaining the operation state vibration evolution characteristics; Step S4: identifying the current operation stage of the real-time working monitoring image, and performing a staged carbonization state analysis based on the vibration evolution characteristics of the operation state, so as to obtain the carbonization state characteristics of the current stage; Step S5: Based on the global portrait of normalized operation, the bubble state evolution of the current carbonization state characteristics is simulated in all stages, and multi-period rolling fault prediction is performed to generate fault prediction data for multiple periods; Step S6: Calculate the fault probability of each time period for the fault prediction data of multiple time periods, make dynamic fault diagnosis decisions, and build an intelligent fault prediction optimization model.

[0007] The present invention can obtain comprehensive operation data of the equipment, including the behavior performance and working environment of the bubble water machine at different stages, through the collection of historical operation logs and real-time monitoring images. These data provide a basis for subsequent analysis and fault prediction. By performing multi-stage state characteristic analysis on historical data, characteristic changes of the bubble water machine at different working stages can be revealed, such as fluctuations over time of parameters such as pressure, temperature, and flow. This analysis helps to clarify the state boundary between normal operation and potential failure of the equipment, laying the foundation for subsequent state tracking and fault prediction. By normalizing the multi-stage operation data, the normal operating mode of the equipment can be better tracked, and the standard behavior of the equipment at different stages can be identified, which is very helpful for detecting deviations of the equipment from the normal state. It provides a basis for the state and timely identification of potential faults, and constructs a dynamic operation portrait, which can continuously reflect the behavioral changes of the bubble water machine in long-term operation. This not only helps to timely identify the dynamic changes in equipment operation, but also provides long-term trend reference data for future fault prediction. The sensor collects the vibration audio signal of the bubble water machine in real time, and the working state of the equipment can be monitored in real time. The vibration signal can reflect the mechanical state of the equipment, such as the operating conditions of components such as gears and bearings, and the possible source of equipment failure can be identified early. Time-frequency analysis of the vibration audio signal can more accurately capture the characteristics of different frequency bands in the signal, and then identify different working states. This is important for judging the changes in the operating state of the equipment and discovering potential faults in advance. Significance: By analyzing the real-time working monitoring images, the system can accurately identify the current operating stage of the bubble water machine. Combined with the vibration evolution characteristics, it can further identify whether there are physical or chemical changes such as carbonization, which helps to timely grasp the specific working status of the equipment, thereby improving the accuracy of monitoring. The carbonization state of the bubble water machine is usually related to the performance and life of the equipment. Through the staged carbonization state analysis, the loss of equipment can be warned in advance, so as to take necessary maintenance or replacement measures to prevent the occurrence of failures. By simulating the changes in the bubble state of the equipment at different operating stages, the health status of the equipment can be comprehensively evaluated. This simulation provides accurate input data for subsequent fault predictions and helps identify potential risks of equipment at different stages. By performing multi-period rolling fault prediction, the failure risk of equipment can be predicted at different time points. This method can provide more detailed fault warning information to ensure timely measures to avoid sudden failures. Fault probability calculation of fault prediction data for each period helps to quantify the future failure risk of equipment. This method can provide maintenance personnel with specific decision-making basis to help them formulate the best maintenance strategy. By establishing an intelligent fault prediction model, the system can continuously optimize the accuracy of fault prediction. Combined with historical data, real-time monitoring signals and dynamic analysis results, the model can continuously improve the accuracy of prediction and reduce false alarms and missed alarms. During the operation of the equipment, the fault diagnosis decision model can be adjusted in real time to adapt to changes in the status of the equipment.Provide flexible fault diagnosis and decision support, which makes fault prediction and diagnosis more accurate, dynamic and intelligent.

[0008] Preferably, step S1 comprises the following steps: Step S11: Obtaining the historical operation log and real-time working monitoring image of the bubble water machine; Step S12: performing a multi-stage operation status characteristic analysis based on the historical operation log of the bubble water machine, thereby obtaining the operation status characteristics of each working stage; Step S13: performing logical correlation analysis of upper and lower stage operations based on the historical operation log of the bubble water machine to generate logical rules for operation between stages; Step S14: Analyze the operation status time series changes of the operation status characteristics of each working stage according to the operation logic rules between stages, so as to generate multi-stage operation status time series change characteristic data.

[0009] The present invention can obtain all key parameters and operating environment data of the long-term operation of the equipment by collecting the historical operation logs and real-time working monitoring images of the bubble water machine. The historical operation logs record the operating status, fault records and maintenance status of the equipment. The real-time monitoring images provide the current working status of the equipment, especially the external working conditions and the physical status of the equipment. The historical operation logs and real-time monitoring images can provide a rich source of information for subsequent data analysis. The comprehensiveness of these data provides an important basis for subsequent status analysis, fault diagnosis and trend prediction. By analyzing the historical operation logs of the bubble water machine, the operation process of the equipment can be divided into multiple different stages (such as the startup stage, the operation stage, the shutdown stage, etc.). The characteristics of each stage (such as pressure, temperature, flow, power, etc.) can reveal the operating rules and performance of the equipment in a specific stage. The multi-stage operating status feature analysis helps to identify abnormal changes in each stage. For example, abnormal fluctuations in the operating characteristics of a certain stage may mean that there are faults or potential problems in this stage, thereby achieving more accurate fault diagnosis. Through the logical correlation analysis of the upper and lower stage operations, the mutual dependence and influence rules between different working stages can be revealed. The state of the equipment in one stage may affect the operating performance of the next stage. Especially the relationship between key stages such as startup, operation and shutdown. By generating the operation logic rules between stages, the behavior patterns of the equipment in each stage can be more clearly understood. This analysis helps to predict the abnormal conditions that may occur in the equipment in the future stage, and take measures early to prevent equipment failures. Understanding the logic rules between stages helps to formulate more reasonable maintenance plans. For example, by predicting the load or possible failures in the next stage, maintenance work can be arranged in advance to avoid unplanned downtime. By performing time-series change analysis on the operation logic rules between stages, the dynamic change process of the equipment in each stage can be revealed. This time-series analysis not only focuses on the current state of the equipment, but also can foresee future state changes through historical data trends. Time-series change analysis helps to compare between different working stages and understand the performance differences of the equipment in each stage. For example, the load change of the equipment in a certain stage may cause abnormalities in the next stage. Through this change trend, problems that the equipment may encounter in the future stage can be predicted early. By analyzing the time-series change characteristic data of each stage, more refined fault diagnosis can be achieved. For example, when the vibration or temperature change in a certain stage exceeds the normal range, the system can immediately issue an early warning to notify the operator to take corresponding maintenance measures.

[0010] Preferably, the specific steps of step S14 are: Identify multiple regular working stages of the bubble water machine based on the historical operation log of the bubble water machine; the multiple regular working stages of the bubble water machine include a water source preparation stage, a gas injection stage, a pressure loading stage, and a storage and water discharge stage; Monitor the working status parameters of each subsystem during the water source preparation stage and extract multiple subsystem status monitoring parameters, wherein the status monitoring parameters include temperature, pressure and water flow; The fluctuation range of multiple subsystem status monitoring parameters is calculated to obtain the status fluctuation range characteristics of the water source preparation stage; During the gas injection stage, the CO2 injection pressure in the gasification chamber is counted and the injection pressure parameters are extracted; Performing pressure variation analysis on injection pressure parameters to generate injection pressure variation data; The water flow rate is calculated during the gas injection stage to obtain the water flow rate characteristics during the gas injection stage; Based on the injection pressure change data and the water flow characteristics during the gas injection stage, dynamic balance evolution is performed to generate dynamic balance characteristics during the gas injection stage; Calculate the real-time pressure of the compression pump based on the pressure loading stage; Fitting the time-series pressure change of the real-time pressure of the compression pump to obtain the time-series pressure change curve in the pressure loading stage; Based on the storage and discharge stages, the transient flow change analysis of the water outlet valve is carried out to obtain the transient flow change characteristics of the water outlet valve; Calculate the water source temperature during the storage and water discharge stages, and extract the water outlet temperature of the bubble machine; Perform regular water outlet state detection on the transient flow change characteristics of the water outlet valve and the water outlet temperature of the bubble machine, thereby generating regular water outlet state characteristics; Based on the state fluctuation range characteristics of the water source preparation stage, the water flow characteristics of the gas injection stage, the time-series pressure change curve of the pressure loading stage and the conventional water discharge state characteristics, the normalized operation state characteristics of each stage are analyzed to obtain the operation state characteristics of each working stage.

[0011] The present invention helps to decompose the operation process of the bubble water machine into multiple detailed stages by identifying the water source preparation stage, gas injection stage, pressure loading stage, and storage and water discharge stage. By accurately distinguishing each stage, the specific functions and possible sources of failure in each stage can be better understood, thereby optimizing the monitoring and maintenance strategies. Different working stages correspond to different operating requirements and potential risks. By clarifying the division of each stage, monitoring and failure prediction can be implemented in a targeted manner. By monitoring the temperature, pressure, and water flow of the subsystems in the water source preparation stage, the health status of the water source preparation can be comprehensively evaluated. Each parameter represents the operating condition of the bubble water machine. Temperature and pressure abnormalities may be precursors to equipment failure or reduced efficiency. , water flow rate affects the water inlet and outlet efficiency of the equipment. By monitoring these basic operating parameters, potential problems in the water source preparation stage, such as unstable water flow and pressure fluctuations, can be discovered in time, which helps to take measures in advance to prevent equipment failure. By calculating the fluctuation range of each monitoring parameter in the water source preparation stage, the stability standard of the equipment in normal operation can be obtained, which provides a quantitative basis for judging whether the equipment is operating within the normal range. If the fluctuation range of a certain parameter exceeds the normal range, it may mean that the equipment is abnormal or has potential failure risks. Statistics and change analysis of the carbon dioxide injection pressure in the gas injection stage can reveal the health status of the gas injection system. The fluctuation range of the injection pressure can reflect the injection The stability of the injection system. Too high or too low pressure may indicate system failure, such as valve failure, gas pipeline blockage and other problems. The water flow characteristics of the gas injection stage are calculated to help evaluate the impact of gas injection on the water flow. The stability of the water flow is crucial to the efficiency and performance of the bubble water machine. Too high or too low water flow may cause the bubble water machine to perform poorly or fail. By fitting the time series changes of real-time pressure data, a smooth pressure change curve can be obtained to reveal the dynamic characteristics of the system. The fitting curve helps to understand the normal fluctuation range of the pressure loading stage and promptly discover pressure anomalies and potential failures. The transient flow changes of the outlet valve in the storage and outlet stages directly affect the final output quality of the equipment. By analyzing flow changes, unstable factors in the water outlet process of the equipment can be discovered in time, such as valve failure or inaccurate flow control. By combining the transient flow characteristics of the water outlet valve and the water outlet temperature of the bubble water machine for routine water outlet status detection, it is possible to verify whether the equipment is in normal water outlet mode. This detection helps to ensure the stability of the equipment throughout the water outlet process and prevent equipment failures due to water outlet problems. By comprehensively analyzing the fluctuation range characteristics of the water source preparation stage, the water flow characteristics of the gas injection stage, the pressure change curve of the pressure loading stage and the routine water outlet status characteristics, the normalized operating status characteristics of the bubble water machine in multiple working stages can be analyzed, which provides a basis for long-term and comprehensive monitoring of equipment status.

[0012] Preferably, the specific steps of step S2 are: Step S21: performing full-process operation dependency analysis based on the historical operation log of the bubble water machine to obtain full-process operation dependency data; Step S22: performing end-to-end change tracking of normalized operation based on the full-process operation dependency data, thereby generating a normalized operation state change trajectory; Step S23: performing state-related multivariate regression modeling on the multi-stage operation state time series change characteristic data based on the normalized operation state change trajectory to obtain an operation state time series change network; Step S24: Evolve the global dynamic profile of the operation status time series change network and construct a global profile of normalized operation.

[0013] The present invention can clarify the dependency between various working stages by analyzing the historical operation log of the bubble water machine, which helps to understand the interaction between the various subsystems of the equipment, such as how the gas source system, water flow system, compression system, etc. work together, and how the failure of a subsystem may affect the operation of other subsystems. The full-process dependency analysis can reveal key links and nodes, and understand which system or component anomalies will cause global operation problems. These dependency points provide important clues for subsequent fault prediction and early warning systems. Normalized operation end-to-end change tracking helps to depict the state change trajectory of the bubble water machine in long-term normal operation. This change trajectory can help the system understand the typical behavior of the equipment in different time periods and different working environments. As a benchmark for "normal" behavior, by tracking the change trajectory of the normalized operation state, the deviation between the current state of the equipment and the historical normal trajectory can be compared in real time. If the equipment state deviates abnormally, potential faults or performance degradation can be quickly identified to achieve early warning. By correlating the normalized operation state change trajectory with multi-stage time series data, the deep-level relationship between the various stages can be revealed, and then how the various states interact with each other and affect the performance of the overall equipment can be understood. Multivariate regression modeling can extract the most relevant features of the operation state change from multiple factors. Create a more accurate state change prediction model. Through this modeling method, we can comprehensively evaluate the impact of various variables on equipment operation and improve the accuracy of fault prediction. Through state association modeling of time-series change characteristics, we can effectively identify the causal relationship between different working stages. For example, anomalies that occur during the gas injection stage of the bubble water machine may affect the pressure or temperature changes in the water source preparation stage, thereby causing failure of the entire system. Therefore, by establishing a time-series change network, these potential associations can be identified in advance. By performing global dynamic portrait evolution on the operating state time-series change network, the global operating state of the bubble water machine can be continuously tracked and updated. The global portrait not only represents the current operating status of the equipment, but also includes its long-term change trends and potential abnormal patterns. The evolution of the global dynamic portrait enables the system to compare the current operating status of the equipment with the historical status in real time, so as to determine whether the equipment is in a normal state. If the global portrait of the equipment deviates from the normal state, the system can automatically issue an early warning to help the operation and maintenance personnel intervene in time. By constructing a global operation portrait, the system can fully understand the performance of the sparkling water machine throughout its life cycle from start-up to stop, and identify the key indicators of each stage, which helps to accurately predict the health status of the equipment at different working stages, optimize resource allocation, and improve the overall reliability of the equipment.

[0014] Preferably, step S3 specifically comprises the following steps: Step S31: monitoring the vibration audio signal of the bubble water machine in real time based on the sensor; Step S32: performing multi-time-frequency decomposition on the operating vibration audio signal of the bubble water machine to generate operating vibration frequency domain data; Step S33: performing dynamic vibration frequency calculation on the operation vibration frequency domain data to generate dynamic operation vibration frequency characteristics; Step S34: performing time series fluctuation amplitude recognition on the dynamic operation vibration frequency characteristics to obtain the operation vibration amplitude characteristics; Step S35: performing an operation state evolution of the bubble water machine on the operation vibration amplitude characteristics, thereby obtaining an operation state vibration evolution characteristic.

[0015] The present invention utilizes sensors to collect the vibration audio signals of the bubble water machine in real time, and can realize dynamic monitoring of the equipment status. These vibration signals contain key information generated by the equipment during operation, such as whether the machine runs smoothly and whether there is abnormal vibration. The vibration audio signals can capture small changes in the mechanical parts inside the equipment. For example, early faults such as bearing wear and loose gears usually manifest as abnormal vibration signals. By monitoring these signals in real time, problems can be discovered early. By performing multi-time-frequency decomposition on the vibration signal, the characteristics of the frequency change over time in the signal can be extracted. This method can identify the vibration frequency changes of the equipment in different time periods and reveal more detailed status information. Vibration signals are usually nonlinear and complex. Direct processing may make it difficult to extract effective information. Multi-time-frequency decomposition converts vibration signals into frequency domain data that is easy to analyze, and can effectively identify potential fault characteristics. For example, vibration enhancement in a certain frequency band may indicate wear or failure of a certain component. Multi-time-frequency decomposition enables the real-time monitoring system to capture and analyze the vibration state changes of the bubble water machine in real time, providing immediate basis for dynamic fault prediction. By calculating the dynamic vibration frequency of the vibration frequency domain data, the frequency fluctuation trend of the equipment during operation can be revealed. Different parts of the equipment will have specific vibration frequencies during operation. Any abnormality of any part will cause a change in frequency. Certain mechanical failures, such as gear wear and bearing failure, often cause changes in the vibration frequency of the equipment. Through dynamic vibration frequency The frequency characteristics can detect problems in advance when these frequencies deviate from the normal range, thereby issuing an early warning to prevent further deterioration of the fault. The dynamic vibration frequency can reflect subtle changes in the operation of the bubble water machine and provide important information about the health of the equipment. For example, when the equipment is operating normally, the vibration frequency will remain stable, while irregular frequency changes may be a harbinger of failure. The timing fluctuation amplitude recognition of the dynamic vibration frequency characteristics helps to identify changes in the vibration amplitude during the operation of the equipment. Increased or unstable vibration amplitude is usually a warning sign of equipment failure. For example, bearing wear, loose parts, etc. may cause increased vibration amplitude. Changes in vibration amplitude usually reflect whether the operating status of the equipment is within the normal range. The analysis of vibration amplitude characteristics can determine whether the equipment has an abnormal state, such as unstable operation or excessive load, so that maintenance measures can be taken in advance. The analysis of the operating state evolution of the vibration amplitude characteristics can reveal the vibration evolution law of the sparkling water machine under different operating conditions. As the equipment goes from startup to operation and then to shutdown, the vibration amplitude will change with time, load and other factors. By analyzing these changes, we can more accurately understand the health status of the equipment. Equipment failure is usually not a one-time process, but a gradual evolution process. By analyzing the operating state vibration evolution characteristics, the gradual development of potential faults can be tracked. For example, abnormal evolution of vibration frequency and amplitude can indicate the gradual damage of internal components of the equipment, providing guidance for early maintenance.

[0016] Preferably, the specific steps of step S4 are: Step S41: identifying the current operation stage of the real-time working monitoring image based on the historical operation log of the bubble water machine, thereby obtaining the current working stage of the bubble water machine; Step S42: performing visual semantic recognition on the real-time working monitoring image and marking a plurality of water flow bubble nodes; Step S43: Calculating the bubble distribution density of multiple water flow bubble nodes to obtain water flow bubble distribution density characteristics; Step S44 performs bubble shape and size analysis on multiple water bubble nodes to generate bubble shape and size data; Step S45: mining the comprehensive bubble situation of the water bubble distribution density characteristics and the bubble shape and size data, thereby generating bubble situation characteristics; Step S46: Performing a staged carbonization state analysis on the operating state vibration evolution characteristics and the bubble state characteristics, so as to obtain the carbonization state characteristics of the current stage.

[0017] The present invention can accurately identify the working stage (such as water source preparation stage, gas injection stage, etc.) of the current equipment by analyzing the real-time working monitoring image based on the historical operation log of the bubble water machine. This stage identification can help the system accurately understand the working status and task progress of the equipment, and provide necessary context information for subsequent fault prediction and status evaluation. By identifying the current working stage in real time, the system can dynamically adjust the monitoring and fault prediction strategies. For example, the behavior and failure mode of the bubble water machine may be different in different working stages. Identifying the current stage helps to optimize monitoring according to the characteristics of different stages. Correctly identifying the current stage can reduce misjudgment and make fault diagnosis more accurate, especially in complex operating environments, to avoid confusion between different The data errors or analysis errors caused by the stage can be analyzed by using visual semantic recognition technology to analyze the real-time working monitoring images, and the bubble nodes in the water flow can be marked. These bubble nodes are the key indicators of the working status of the bubble water machine, which can reflect the efficiency and quality of water flow and gas injection. Image recognition technology makes the entire monitoring process highly automated without manual intervention. By marking the bubble nodes, the distribution of bubbles can be analyzed more accurately, so as to determine whether the bubble injection is uniform, whether there is blockage or uneven flow, etc. By visually identifying and marking the bubble nodes, the system can monitor the situation of bubbles in the water flow in real time. For example, if the bubble density is abnormal, it may be a warning signal of gas injection system failure. The system can automatically issue an alarm and calculate the water flow. The distribution density of bubbles in the flow can quantify the uniformity and stability of the bubble flow pattern. Changes in the bubble distribution density can reflect changes in the operating status of the bubble water machine, such as uneven bubble generation, uneven airflow distribution, or missing bubbles in certain areas. Abnormal changes in the bubble distribution density are usually an indication of equipment problems. For example, uneven bubble density may mean that there is a problem with the gas injection device, or that the water flow is blocked, resulting in uneven bubble generation. By calculating and monitoring this feature in real time, the system can detect these problems in a timely manner. The size and shape of the bubbles are one of the important parameters for the normal operation of the bubble water machine. By analyzing the bubble morphology, the bubble size data can be obtained to determine whether the gas injection is normal. If the shape and size of the bubbles are uneven, It may indicate that the equipment has failed or the bubble generation is uneven. The bubble shape and size analysis can help understand the quality of bubble injection. Under normal circumstances, the bubbles should maintain a certain size distribution. If the size is abnormal, it may indicate insufficient gas flow, blockage of the injection system, etc. Through the analysis of bubble shape and size data, potential problems in the bubble generation process can be accurately judged. For example, if the size of the bubble is too small, it may mean that the gas pressure is too low or the flow is unstable. The system can diagnose and repair these problems. Combined with the bubble distribution density and bubble shape and size data, comprehensive situation mining can be carried out to fully understand the overall operating status of the bubble water machine. The bubble situation characteristics integrate multiple dimensions such as the number, distribution, and shape of bubbles.It can provide more accurate and comprehensive equipment health information. The mining of bubble status characteristics helps to reveal potential equipment failures from multiple dimensions. Through comprehensive analysis of bubble behavior, it can be determined whether the equipment is in an abnormal working state. For example, a sudden change in bubble status may indicate that the airflow of the bubble water machine is unstable or the equipment is damaged.

[0018] Preferably, the specific steps of step S5 are: Step S51: performing local characteristic time series change analysis on the carbonization state characteristics of the current stage, thereby generating local carbonization state change characteristics; Step S52: mining the time series variation law of the local carbonization state variation characteristics to generate a stage-by-stage time series variation law; Step S53: simulating the evolution of the bubble state in all stages according to the phased time series variation law to obtain simulation data of the evolution of the bubble state in all stages; Step S54: Based on the global portrait of normalized operation, multi-period rolling fault prediction is performed on the simulation data of bubble state evolution in all stages to generate fault prediction data for multiple time periods.

[0019] The present invention can accurately capture the changes in the carbonization phenomenon in the local area by analyzing the local time-series changes of the carbonization state characteristics in the current stage. This local analysis can help identify the different carbonization degrees of various parts of the bubble water machine, and promptly discover whether certain areas are over-carbonized, thereby avoiding the overall performance degradation of the equipment. The local change characteristics of the carbonization state can reveal operating problems in certain parts of the equipment, especially under high load or special environments of the bubble water machine, the carbonization of local components may appear earlier than the carbonization of the overall system. By analyzing these local changes, potential problems can be discovered in advance, thereby preventing the expansion of faults. The local carbonization state change characteristics can provide maintenance personnel with more detailed data. It helps decision makers understand which areas or components are at greater risk, thereby achieving accurate maintenance and repair, improving maintenance efficiency, and reducing repair costs. By mining the time-series change rules of local carbonization state change characteristics, the evolution of carbonization over time can be revealed. The identification of this rule helps to understand the natural change trend of the carbonization state in different working stages or time periods. Mastering the time-series change rules of the carbonization state can help the system identify potential faults in advance. For example, when the carbonization change begins to deviate from the normal rule, it may indicate that a component of the sparkling water machine has an abnormality. The system can make early warnings and take corrective measures in time. The evolution simulation of the bubble state in all stages can fully understand the changes in the bubble state of the bubble water machine during the entire operation process. This simulation is not limited to the current stage, but covers the entire life cycle of the equipment or the state changes in different working stages, thereby providing a rich data basis for future fault prediction. By simulating the evolution of the bubble state in each stage, the potential problems of the bubble injection system can be understood more accurately. For example, the simulation results can reveal changes in bubble distribution, gas injection efficiency, etc., and timely discover fault signs such as uneven bubble injection or unstable flow. By performing multi-period rolling predictions on the simulation data of the bubble state evolution in all stages, the system can provide fault predictions at different time points. Fault prediction data. This rolling prediction can calculate potential faults in the next few time periods based on the current status changes of the equipment, thereby providing time warning for equipment maintenance. Fault prediction data for multiple time periods can help operation and maintenance personnel to arrange maintenance plans reasonably. The system can predict possible equipment failures in the future and notify relevant personnel in advance for maintenance to avoid production interruptions or equipment damage when a failure occurs. Through multi-time period fault rolling prediction, the bubble water machine can be kept in the optimal working state, and potential problems can be identified and repaired in time, which can not only improve the operating efficiency of the equipment, but also extend the service life of the equipment, reduce downtime, and improve the overall stability and reliability of the equipment.

[0020] Preferably, the specific steps of step S6 are: Step S61: Calculate the failure probability of each time period for the failure prediction data of multiple time periods, and generate the failure probability of each time period; Step S62: extracting the maximum failure probability time period based on the failure probability of each time period; Step S63: locating and tracing the fault based on the fault probability of each time period, thereby generating a fault tracing point; Step S64: making a dynamic fault diagnosis decision on the fault source tracing point, thereby obtaining a dynamic fault diagnosis strategy; Step S65: Optimize cloud-edge collaborative fault management for the dynamic fault diagnosis strategy and build an intelligent fault prediction optimization model.

[0021] The present invention calculates the failure probability of the failure prediction data time period by time period, and can generate a clear failure probability for the equipment status in each time period. This probabilistic analysis method can quantify the failure risk of the equipment in each future time period, and help operation and maintenance personnel better understand and foresee possible problems with the equipment. After calculating the failure probability, the system can dynamically monitor the equipment in real time and issue early warnings in time periods with a high probability of failure. This helps to take corresponding measures in advance to avoid major impacts of equipment failures on production or operations. The failure probability calculation provides data support for resource allocation. In time periods with a high probability of failure, the system can suggest increasing the monitoring frequency of operation and maintenance personnel or mobilizing more resources for inspection to prevent failures in advance. By extracting the time period with the highest probability of failure, the system can accurately identify the time period when the equipment is most likely to fail during operation. After identifying these high-risk time periods, operation and maintenance personnel can concentrate their energy and resources on more intensive monitoring and detection of these key time periods. By focusing on the time periods with the highest probability of failure, the effectiveness of fault warning can be significantly improved, and corresponding measures can be taken in a timely manner, such as adjusting equipment operating parameters, increasing inspection frequency, etc., which can reduce the probability of equipment failure during high-risk time periods. Analysis based on failure probability can help the system identify the source point of the fault, that is, the specific equipment component or system function where the fault occurs. This source tracing analysis can ensure the accuracy of fault diagnosis and avoid misdiagnosis and missed diagnosis. Fault location is not only about identifying The time period when the fault occurred can also be traced back to the root cause of the fault. Through fault tracing analysis, it can help identify the specific components or system problems that caused the equipment to fail within a specific period of time, and provide a clear direction for subsequent maintenance work. Based on the fault tracing point, the system can make dynamic fault diagnosis decisions. This dynamic decision is based on the real-time data and historical failure modes of the current equipment, and can adapt to different fault manifestations of the equipment in different states, providing more accurate diagnosis results. Through dynamic fault diagnosis strategies, the system can provide targeted maintenance plans. For example, according to different fault types and severity, the system can recommend different maintenance plans, including parts replacement, equipment adjustment or other repair measures, which can improve the efficiency of fault repair. Efficiency and ensure the accuracy of maintenance. Through the cloud-edge collaborative management model, the system can distribute data processing and decision optimization between the cloud and edge devices. Edge computing can quickly process local device data and make real-time responses, while the cloud is responsible for large-scale data analysis and long-term trend prediction. The collaborative work of the two can improve the system response speed and decision-making efficiency. The powerful computing power and big data analysis of the cloud can deeply mine the historical operation data of the equipment and discover potential failure modes, while the edge device can make an immediate response based on the current real-time data. For example, when the system identifies that a certain equipment component is at risk of failure, the edge device can immediately take local measures, such as adjusting working parameters or notifying operation and maintenance personnel. Based on the cloud-edge collaborative model,The system can continuously optimize fault prediction and management strategies. Through continuous learning and optimization, the system can gradually improve the accuracy of prediction and the intelligence level of decision-making. For example, as the equipment operation data accumulates, the system can predict the occurrence cycle of equipment failures and arrange maintenance in advance, thereby reducing the probability of failures and equipment downtime.

[0022] In this specification, a remote monitoring and fault prediction system for a sparkling water machine is provided, which is used to execute the remote monitoring and fault prediction method for a sparkling water machine as described above, including: The time series change module is used to obtain the historical operation logs and real-time working monitoring images of the bubble water machine; perform multi-stage operation status feature analysis based on the historical operation logs of the bubble water machine, and generate multi-stage operation status time series change characteristic data; The normalization portrait module is used to track the end-to-end changes of normalized operation based on the multi-stage operation status time series change characteristic data, and to perform global dynamic portrait evolution to build a global portrait of normalized operation; A vibration evolution module is used to monitor the vibration audio signal of the bubble water machine in real time based on a sensor; perform multi-time-frequency decomposition on the vibration audio signal of the bubble water machine and the bubble water machine operation state evolution, so as to obtain the vibration evolution characteristics of the operation state; The carbonization state module is used to identify the current operation stage of the real-time working monitoring image and perform a staged carbonization state analysis based on the vibration evolution characteristics of the operation state, so as to obtain the carbonization state characteristics of the current stage; The multi-period fault prediction module is used to simulate the evolution of the bubble state in all stages based on the global portrait of normalized operation and to perform multi-period fault rolling prediction to generate fault prediction data for multiple periods. The fault diagnosis decision module is used to calculate the fault probability of each time period for the fault prediction data of multiple time periods, and then make dynamic fault diagnosis decisions to build an intelligent fault prediction optimization model.

[0023] By combining real-time and historical data, the present invention can comprehensively obtain the operating status of the bubble water machine at different working stages, provide basic data support for subsequent analysis, and analyze the operating data at different stages to identify the key characteristics of each stage (such as pressure, flow, temperature, etc.), providing a comprehensive understanding of the operating status of the equipment, which helps to determine the health status of the equipment at each stage, as well as the potential failure risks at each stage. Based on the staged feature analysis, the time series change data of the bubble water machine operation process can be generated. This data helps to identify the trends and fluctuations of the equipment in operation, and warn of possible abnormal conditions in advance. By constructing a global portrait of normalized operation, the system can continuously track the operating changes of the equipment and establish normal operation of the equipment. The standard model of the equipment is used for operation. Any deviation from this standard can serve as an early warning of potential failures. The evolution of the global portrait can reflect the changes in the global status of the equipment during long-term operation. The system can continuously update the global portrait based on real-time data, provide continuous equipment health status monitoring, and identify long-term trends and short-term abnormal fluctuations in equipment operation through end-to-end change tracking. It can identify the phenomenon of equipment gradually deviating from the normal state in advance, which helps to detect potential failures early. Through real-time monitoring of vibration audio signals by sensors, it is possible to obtain changes in the operating status of the bubble water machine in a timely manner, especially the vibration status of key components, which helps to respond quickly when the equipment is abnormal. Multi-time-frequency decomposition can deeply analyze the frequency and amplitude of vibration signals and identify the equipment in Vibration characteristics under different operating conditions. This refined analysis helps to identify abnormal modes in operation, especially vibration signals related to faults. By analyzing the vibration evolution characteristics, signs of potential faults can be identified in advance, such as wear, looseness, bearing failure and other problems. This provides strong data support for fault diagnosis and maintenance. According to real-time images and vibration data, the carbonization state of the sparkling water machine at different working stages can be analyzed. This staged analysis can provide a combination of vibration signals and image recognition to more accurately identify carbonization phenomena and their impact on equipment, which helps to locate the root cause of the problem and provide guidance for repair. Based on the refined analysis of the carbonization state, maintenance personnel can deal with carbonization problems at specific stages in a targeted manner. , avoid unnecessary costs caused by comprehensive inspections and improve maintenance efficiency. By simulating the evolution of the bubble state of the bubble water machine in different working stages, the performance of the equipment under different loads and environmental conditions can be fully understood, which makes the equipment failure prediction more accurate and covers the entire operation cycle. Through rolling predictions for multiple time periods in the future, the health status of the equipment in the future can be accurately inferred based on the current status. In this way, the system can foresee the possibility of future failures in advance and help operation and maintenance personnel take preventive measures in advance to avoid sudden failures. Multi-period failure prediction provides a long-term early warning model, so that equipment maintenance is not limited to the current stage, but maintenance work can be planned and arranged in advance to minimize equipment downtime and maintenance costs.By calculating the probability of fault prediction data for each time period, the system can provide different fault probabilities for different time points, which makes fault prediction not only static, but also dynamically changing, which helps to accurately identify risks. By combining the time period fault probability, the system can perform intelligent fault diagnosis according to the current operating status of the equipment. The ability of dynamic decision-making can help the system optimize the fault diagnosis strategy according to real-time data. By building an intelligent fault prediction optimization model, the system can not only make accurate fault predictions, but also continuously optimize the prediction strategy according to historical data and real-time data. As the amount of data increases, the prediction accuracy of the system will continue to improve, thereby better supporting equipment management and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic flow chart of the steps of a method for remote monitoring and fault prediction of a sparkling water machine according to the present invention; Figure 2 Detailed implementation flow chart of step S1; Figure 3 Detailed implementation flow chart of step S2; Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0025] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0026] The present application example provides a method and system for remote monitoring and fault prediction of a bubble water machine. The execution subject of the method and system for remote monitoring and fault prediction of a bubble water machine includes but is not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio and image management system, an information management system, and a cloud data management system.

[0027] See also Figures 1 to 4 The present invention provides a method for remote monitoring and fault prediction of a bubble water machine, the method comprising the following steps: Step S1: Obtain the historical operation log and real-time working monitoring image of the bubble water machine; perform multi-stage operation state characteristic analysis based on the historical operation log of the bubble water machine to generate multi-stage operation state time series change characteristic data; Step S2: Based on the multi-stage operation status time series change characteristic data, end-to-end change tracking of normalized operation is performed, and global dynamic portrait evolution is performed to construct a global portrait of normalized operation; Step S3: monitoring the bubble water machine operation vibration audio signal in real time based on the sensor; performing multi-time-frequency decomposition on the bubble water machine operation vibration audio signal and the bubble water machine operation state evolution, thereby obtaining the operation state vibration evolution characteristics; Step S4: identifying the current operation stage of the real-time working monitoring image, and performing a staged carbonization state analysis based on the vibration evolution characteristics of the operation state, so as to obtain the carbonization state characteristics of the current stage; Step S5: Based on the global portrait of normalized operation, the bubble state evolution of the current carbonization state characteristics is simulated in all stages, and multi-period rolling fault prediction is performed to generate fault prediction data for multiple periods; Step S6: Calculate the fault probability of each time period for the fault prediction data of multiple time periods, make dynamic fault diagnosis decisions, and build an intelligent fault prediction optimization model.

[0028] The present invention can obtain comprehensive operation data of the equipment, including the behavior performance and working environment of the bubble water machine at different stages, through the collection of historical operation logs and real-time monitoring images. These data provide a basis for subsequent analysis and fault prediction. By performing multi-stage state characteristic analysis on historical data, characteristic changes of the bubble water machine at different working stages can be revealed, such as fluctuations over time of parameters such as pressure, temperature, and flow. This analysis helps to clarify the state boundary between normal operation and potential failure of the equipment, laying the foundation for subsequent state tracking and fault prediction. By normalizing the multi-stage operation data, the normal operating mode of the equipment can be better tracked, and the standard behavior of the equipment at different stages can be identified, which is very helpful for detecting deviations of the equipment from the normal state. It provides a basis for the state and timely identification of potential faults, and constructs a dynamic operation portrait, which can continuously reflect the behavioral changes of the bubble water machine in long-term operation. This not only helps to timely identify the dynamic changes in equipment operation, but also provides long-term trend reference data for future fault prediction. The sensor collects the vibration audio signal of the bubble water machine in real time, and the working state of the equipment can be monitored in real time. The vibration signal can reflect the mechanical state of the equipment, such as the operating conditions of components such as gears and bearings, and the possible source of equipment failure can be identified early. Time-frequency analysis of the vibration audio signal can more accurately capture the characteristics of different frequency bands in the signal, and then identify different working states. This is important for judging the changes in the operating state of the equipment and discovering potential faults in advance. Significance: By analyzing the real-time working monitoring images, the system can accurately identify the current operating stage of the bubble water machine. Combined with the vibration evolution characteristics, it can further identify whether there are physical or chemical changes such as carbonization, which helps to timely grasp the specific working status of the equipment, thereby improving the accuracy of monitoring. The carbonization state of the bubble water machine is usually related to the performance and life of the equipment. Through the staged carbonization state analysis, the loss of equipment can be warned in advance, so as to take necessary maintenance or replacement measures to prevent the occurrence of failures. By simulating the changes in the bubble state of the equipment at different operating stages, the health status of the equipment can be comprehensively evaluated. This simulation provides accurate input data for subsequent fault predictions and helps identify potential risks of equipment at different stages. By performing multi-period rolling fault prediction, the failure risk of equipment can be predicted at different time points. This method can provide more detailed fault warning information to ensure timely measures to avoid sudden failures. Fault probability calculation of fault prediction data for each period helps to quantify the future failure risk of equipment. This method can provide maintenance personnel with specific decision-making basis to help them formulate the best maintenance strategy. By establishing an intelligent fault prediction model, the system can continuously optimize the accuracy of fault prediction. Combined with historical data, real-time monitoring signals and dynamic analysis results, the model can continuously improve the accuracy of prediction and reduce false alarms and missed alarms. During the operation of the equipment, the fault diagnosis decision model can be adjusted in real time to adapt to changes in the status of the equipment.Provide flexible fault diagnosis and decision support, which makes fault prediction and diagnosis more accurate, dynamic and intelligent.

[0029] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of a method for remote monitoring and fault prediction of a bubble water machine of the present invention. In this example, the steps of the method for remote monitoring and fault prediction of a bubble water machine include: Step S1: Obtain the historical operation log and real-time working monitoring image of the bubble water machine; perform multi-stage operation state characteristic analysis based on the historical operation log of the bubble water machine to generate multi-stage operation state time series change characteristic data; In this embodiment, historical operation logs are obtained from the monitoring system of the bubble water machine. These logs usually record various working states of the equipment, including timestamps, operating parameters (such as water flow, bubble generation frequency), alarm information and other related data. At the same time, real-time working monitoring images are collected, which may include bubble generation process, visualization images of the equipment operating state, etc. These images usually come from cameras installed on the equipment, or data visualization collected by sensors. The collected historical operation logs are formatted to ensure data consistency and integrity, which may involve cleaning the data, removing redundant information and erroneous data, ensuring the accuracy of subsequent analysis, and labeling and classifying the real-time working monitoring images for subsequent analysis. Visual analysis and feature extraction, extract historical operation log data through the device interface (such as API or database connection), use SQL query or data extraction tools (such as Python's pandas library) to read the data, ensure that the extracted data includes key operating parameters and status information, and convert it into a structured format (such as CSV or Excel) for subsequent analysis. Get the current working image through the real-time monitoring system. You can set a scheduled capture or trigger the capture based on the device status. Convert the image format to ensure that it is in an analyzable format (such as JPEG, PNG), and store it in a specified directory. Integrate the extracted historical operation log with the real-time monitoring image to form a complete image with a timestamp. For the entire data set, each log record should be associated with the corresponding image data for subsequent analysis. The integrated data should be checked to ensure the accuracy and completeness of the data. The core of the multi-stage operation status feature analysis is to identify and extract the key features of the bubble water machine in different working stages. These features may include operating time, bubble generation rate, pressure change, vibration frequency, etc., determine the characteristic indicators of each stage, and set a calculation method for each feature. For example, the bubble generation rate can be calculated by the number of bubbles generated per unit time. According to the timestamp and status information in the historical operation log, the operation process is divided into multiple stages, such as preparation stage, operation stage, maintenance stage, etc., and stage division standards are formulated to ensure that each stage The division is scientific and reasonable, which can reflect the characteristic changes of the equipment under different working conditions. The data of each working stage is statistically analyzed to extract key features. For example, pandas is used to group and aggregate data, and the average, maximum and minimum values ​​of each stage are calculated. Data visualization tools (such as matplotlib or seaborn) are introduced to show the characteristic changes of each stage, which is convenient for intuitive understanding and analysis. The extracted features are analyzed in multiple dimensions to identify the changes in the operating status of different stages. For example, the bubble generation rate and vibration frequency of different stages are compared to find out the potential abnormal patterns. Statistical analysis methods (such as variance analysis or regression analysis) are applied to test the characteristic differences and correlations between different stages.

[0030] Step S2: Based on the multi-stage operation status time series change characteristic data, end-to-end change tracking of normalized operation is performed, and global dynamic portrait evolution is performed to construct a global portrait of normalized operation; In this embodiment, before starting, ensure that the extraction of the multi-stage operating state time series change characteristic data in the previous step has been completed. The data should include various operating parameters and their time series information of different working stages. These data are integrated into a unified data frame (DataFrame) for subsequent analysis and processing. By analyzing the multi-stage operating state time series change characteristic data, the normalized operation changes of the bubble water machine are tracked, and possible fluctuations and anomalies in the operation process are identified. The acquired multi-stage data are standardized to eliminate the influence of different dimensions and units on the analysis results. Z-score standardization or Min-Max standardization method can be used to convert the data to the same scale, clean the data, remove missing values ​​and outliers to ensure the accuracy and reliability of the data, select a suitable change detection method, such as the Cumulative Sum Control Chart (CUSUM) or the Coefficient of Variation method based on statistics, to identify changes in the operating state, set a threshold for change detection, so as to alarm or mark when the data change exceeds a preset range, analyze the standardized data point by time, apply the selected change detection method, identify the change point of each stage, record the timestamp of the change point and its corresponding characteristic value, and build a dynamic portrait for the subsequent construction. Provide a basis, build a global dynamic portrait of the sparkling water machine based on the results of change tracking, so as to visualize the operating status of the machine at different stages and time points, design a framework for the global dynamic portrait, including the required features, time dimensions, and change trends, ensure that the framework can reflect the operating status and important parameters of the sparkling water machine at different stages, select appropriate data visualization tools (such as matplotlib, plotly, or seaborn) to build dynamic portraits, and use heat maps, line graphs, or scatter graphs to show the changes in various features. Determine the layout and style of the chart to clearly show the features and change trends at different stages. Use the selected visualization tools to visualize the results of change detection. The system can map the time series data onto a chart, identify the characteristic values ​​and change points of each stage, and display the changes in continuous time periods through animation or interactive charts to help users intuitively understand the operating status of the equipment. The system can output the constructed global portrait of normalized operation, including the characteristic changes, change points and influencing factors of each stage, generate a comprehensive report, summarize the results of change tracking analysis and the global dynamic portrait, provide support for subsequent fault prediction, equipment maintenance and management decisions, set up a monitoring mechanism, and regularly update the dynamic portrait to ensure that it reflects the latest operating status. According to real-time data feedback, the construction method of the dynamic portrait can be optimized to improve its accuracy and practicality.

[0031] Step S3: monitoring the bubble water machine operation vibration audio signal in real time based on the sensor; performing multi-time-frequency decomposition on the bubble water machine operation vibration audio signal and the bubble water machine operation state evolution, thereby obtaining the operation state vibration evolution characteristics; In this embodiment, sensors need to be installed at key locations of the bubble water machine to monitor the operating vibration and audio signals of the device in real time. These sensors may include accelerometers, microphones, or vibration sensors, which can capture the vibration and sound characteristics of the bubble water machine in different working states. Ensure that the sensors can capture the vibration audio signals at an appropriate sampling rate (usually at the kilohertz level), store the signal data in the data acquisition system, and select a suitable data format (such as a WAV file or a CSV file) for subsequent analysis. Select and configure suitable sensors based on the operating characteristics of the bubble water machine. The accelerometer is used to capture mechanical vibrations, and the microphone is used to capture audio signals. , ensure that the sensor is located in a position where representative signal data can be obtained, establish a data acquisition system, set the sampling frequency and data storage method through programming (such as using Python or MATLAB), ensure that the system can record the vibration and audio signals output by the sensor in real time, adopt a circular cache mechanism, and regularly store the collected data in a local database or cloud storage to ensure the integrity and reliability of the data, set up an automatic backup mechanism to prevent data loss, and pre-process the signal before performing multi-time-frequency decomposition on the audio signal, including denoising, normalization, and filtering. Frequency domain filters (such as low-pass filters) can be used to remove high-frequency noise and improve the signal. Clarity, select appropriate time-frequency analysis methods, such as wavelet transform (WaveletTransform) or short-time Fourier transform (STFT), which can effectively combine time and frequency information to reveal the time-frequency characteristics of the signal. Use the selected time-frequency analysis method to decompose the preprocessed audio signal. For example, use wavelet transform, select appropriate wavelet basis and decomposition layers, decompose the signal into components of different frequency bands, and extract key features from the decomposed signal, such as vibration frequency, amplitude, energy spectral density, etc. These features can reflect the operation of the bubble water machine. The running status and health status are analyzed by statistical analysis to calculate the mean, variance and other statistics of the features to identify potential abnormal patterns. Combined with the extracted features, a data set of running vibration evolution features is constructed. Time series analysis can be used to observe the changing trend of features over time and identify key change points. Thresholds are set to monitor feature changes in real time. If the feature value exceeds the preset range, an alarm mechanism is triggered to respond to potential faults in a timely manner. The analysis results of the running vibration evolution features are output, including feature change graphs, statistical reports, etc. Visual charts are generated to show the changes in vibration status in different time periods, helping technicians quickly understand the running health status of the equipment.

[0032] Step S4: identifying the current operation stage of the real-time working monitoring image, and performing a staged carbonization state analysis based on the vibration evolution characteristics of the operation state, so as to obtain the carbonization state characteristics of the current stage; In this embodiment, before starting, ensure that real-time working monitoring images and operating state vibration evolution characteristic data have been obtained. These data will be used to identify the current operating stage of the sparkling water machine and perform carbonation state analysis. The real-time working monitoring images are processed to identify the current operating state of the sparkling water machine. Computer vision technology, especially deep learning models (such as YOLO or MaskR-CNN), can be used for image classification. The real-time monitoring images are preprocessed, including image scaling, normalization, and noise removal to ensure the image quality input to the model. These steps are implemented using image processing libraries such as OpenCV to ensure that the image size and format match the training model. A suitable deep learning model is selected for image classification. If a trained model already exists, it can be used directly; otherwise, it is necessary to prepare labeled training data for model training. When training the model, use appropriate loss functions and optimizers to ensure that the model can effectively identify different operating stages. The preprocessed real-time image is input into the trained model, and the model will output the prediction results of the current operating stage, such as the preparation stage, During the operation phase or maintenance phase, a threshold is set based on the prediction probability of the model to determine which phase the current state belongs to. By analyzing the vibration evolution characteristics of the current operation phase, the carbonation state characteristics of the sparkling water machine can be identified, which will help determine the operating health status and potential failure risks of the equipment. Features related to the carbonation state, such as vibration amplitude, frequency, energy spectrum, etc., are extracted from the previous vibration evolution feature analysis. These features can reflect the health status of the equipment at different operation phases. Combined with the characteristics of the current operation phase, a comprehensive feature set is constructed for subsequent analysis. Appropriate carbonation state analysis methods, such as support vector machines (SVMs), decision trees, or neural networks are selected to model the extracted features. The existing carbonation state data is used for model training. During the model training process, cross-validation is used to evaluate the performance of the model to ensure that it can accurately predict the carbonation state at different stages. The extracted features are input into the trained carbonation state analysis model to obtain the current carbonation state characteristics. A threshold is set to determine whether the equipment is in a normal or abnormal state based on the carbonization state characteristics output by the model.

[0033] Step S5: Based on the global portrait of normalized operation, the bubble state evolution of the current carbonization state characteristics is simulated in all stages, and multi-period rolling fault prediction is performed to generate fault prediction data for multiple periods; In this embodiment, before starting, ensure that the global portrait of normal operation, the carbonation state characteristic data of the current stage, and the historical operation data of the bubble water machine have been obtained. These data will be used to simulate the evolution of the bubble state and predict faults. Based on the carbonation state characteristics of the current stage and the global portrait of normal operation, simulate the evolution of the bubble state of the bubble water machine at different operation stages. This will help identify potential problems and optimize the operation of the equipment. Select a suitable simulation model, such as a system dynamics model or an agent model, to describe the operating state and bubble behavior of the bubble water machine. Define the input parameters of the model, including the carbonation state characteristics of the current stage, operating conditions (such as temperature, pressure, etc.) and historical data. Set the simulation parameters according to the global portrait of normal operation to ensure that the model can accurately reflect the operating characteristics and environmental conditions of the equipment. This may include setting parameters such as bubble generation rate, bubble diameter, and bubble stability. Run the simulation model to generate bubble state evolution data at different stages. Monte Carlo simulation or numerical simulation methods can be used to consider different scenarios and uncertainties. The bubble state evolution data set is formed by recording the bubble state changes at each time step and analyzing the bubble behavior under different operating conditions. The bubble state evolution simulation data is integrated with the past fault history data to extract the features related to the occurrence of faults. These features may include bubble state parameters, operating environment parameters and historical fault records. Statistical analysis methods (such as correlation analysis) are used to determine the features that are highly correlated with the probability of fault occurrence. A suitable machine learning model (such as random forest, support vector machine or long short-term memory network LSTM) is selected to train the extracted features. The model should be able to learn fault modes and make predictions based on historical data. During the training process, the performance of the model is evaluated using methods such as cross-validation to ensure its accuracy and generalization ability. The current bubble state features and historical data are input into the trained model for rolling prediction. The prediction window is set, such as predicting the probability of failure in the next hour, day or week. The fault prediction results for each time period are recorded, including the predicted probability of failure, possible types of failures and their impacts.

[0034] Step S6: Calculate the fault probability of each time period for the fault prediction data of multiple time periods, make dynamic fault diagnosis decisions, and build an intelligent fault prediction optimization model.

[0035] In this embodiment, the collected fault prediction data is cleaned to ensure the integrity and accuracy of the data, which includes removing duplicate data, filling missing values ​​and eliminating outliers, and organizing the data into a format suitable for analysis. It may be necessary to convert the data into a time series format so that it can be analyzed time period by time, and select appropriate statistical methods to calculate the fault probability. Commonly used methods include frequency method (based on the occurrence frequency of historical fault data), Bayesian inference method (combining historical data and prior knowledge) and probability predicted by machine learning models (such as classification models). The fault probability of each time period is calculated, and programming tools (such as Python's pandas library) are used to calculate the fault probability time period by time. For each time period, the number of faults is counted and the corresponding probability is calculated. The fault probability and its related features of each time period are recorded in a data frame for subsequent analysis. , according to historical data and expert experience, set the threshold of fault probability. If the fault probability in a certain period exceeds the set value, the fault diagnosis program is triggered. The threshold can be dynamically adjusted using a data-driven method, such as adaptive adjustment based on changes in real-time data. Select a suitable fault diagnosis model, such as a decision tree, random forest, or rule-based expert system. The model needs to be able to judge and recommend fault types based on fault probability and other relevant features. Use historical fault data for model training to ensure that the model can effectively identify different fault types and states. After obtaining the fault probability of each period, input it into the fault diagnosis decision model. The model will output the fault diagnosis results, including the fault type, possible causes, and recommended measures. If the model determines that there is a fault risk, it will generate an alarm message and provide recommended treatment measures to the maintenance personnel.

[0036] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Step S11: Obtaining the historical operation log and real-time working monitoring image of the bubble water machine; Step S12: performing a multi-stage operation status characteristic analysis based on the historical operation log of the bubble water machine, thereby obtaining the operation status characteristics of each working stage; Step S13: performing logical correlation analysis of upper and lower stage operations based on the historical operation log of the bubble water machine to generate logical rules for operation between stages; Step S14: Analyze the operation status time series changes of the operation status characteristics of each working stage according to the operation logic rules between stages, so as to generate multi-stage operation status time series change characteristic data.

[0037] In this embodiment, the sources of the historical operation log and real-time monitoring image of the bubble water machine are confirmed. The historical operation log usually records the working status, operating time, fault information, etc. of the equipment, while the real-time monitoring image provides visual information of the current operation of the equipment. Use appropriate tools and methods to collect historical operation logs, which may include database queries, API calls, or direct exports from the storage medium of the equipment to ensure the integrity and accuracy of the log data. The real-time working monitoring image can be obtained through a camera or monitoring system to ensure that the image is clear and can reflect the working status of the bubble water machine. The image data should be updated regularly to facilitate subsequent analysis. The collected historical operation logs and real-time monitoring images are organized into a unified format. The log data should include timestamps, working status parameters, etc., and the image data needs to be archived and sorted by time. Mark for subsequent analysis, define the operating status characteristics of the sparkling water machine, which may include operating time, power consumption, temperature, pressure, etc. These characteristics can reflect the performance of the device in different working stages, and segment the historical operation log according to the working stage of the device. Common working stages may include startup, stable operation, failure and maintenance, etc. According to the working mode and historical data of the device, determine the time range of each stage, use data analysis tools (such as Python's pandas library) to analyze the historical operation log, read the log data, and divide the different operating stages according to the set thresholds or conditions. For each stage, calculate the relevant operating status characteristics, such as average power, maximum temperature and The operating status characteristic data of each working stage is output, and a visual chart is generated to show the performance characteristics of different stages. This helps to identify the key performance indicators (KPI) of each stage and clarify the logical relationship between different working stages, such as the transition conditions from startup to stable operation, or the trigger factors from stable operation to failure. The operating status characteristic data of each stage are integrated, and the logical association between the stages is analyzed. Association rule learning (such as Apriori algorithm) can be used to identify the relationship between the stages. The logical association analysis is implemented using data analysis tools, and the characteristic data of each stage is converted into a format suitable for algorithm processing. The association rule algorithm is applied to identify the logical rules between the stages. Laws, such as "if the power is higher than a certain threshold during the startup phase, it may enter the stable operation phase", output the generated inter-phase operation logic rules, and display the logical relationship in the form of charts, which will help understand the working mode of the equipment at different stages and its transition conditions, determine the goals of the time series change analysis of the operating status, analyze the changing trends of the characteristics of the equipment at different stages over time, organize the operating status characteristics of each working stage into time series, and ensure that the timestamps of the data are consistent for time series analysis. Use time series analysis tools (such as statsmodels or pandas) to analyze the operating status characteristics of each stage, extract key features for each stage, and generate time series data.Use time series analysis methods (such as moving average, ARIMA model) to analyze the time-varying trend of state characteristics, output the generated multi-stage operation state time series change characteristic data, and generate visual charts to show the changes of state characteristics in different stages over time, which will help identify the dynamic characteristics and potential problems of equipment operation.

[0038] In this embodiment, the specific steps of step S14 are: Identify multiple regular working stages of the bubble water machine based on the historical operation log of the bubble water machine; the multiple regular working stages of the bubble water machine include a water source preparation stage, a gas injection stage, a pressure loading stage, and a storage and water discharge stage; Monitor the working status parameters of each subsystem during the water source preparation stage and extract multiple subsystem status monitoring parameters, wherein the status monitoring parameters include temperature, pressure and water flow; The fluctuation range of multiple subsystem status monitoring parameters is calculated to obtain the status fluctuation range characteristics of the water source preparation stage; During the gas injection stage, the CO2 injection pressure in the gasification chamber is counted and the injection pressure parameters are extracted; Performing pressure variation analysis on injection pressure parameters to generate injection pressure variation data; The water flow rate is calculated during the gas injection stage to obtain the water flow rate characteristics during the gas injection stage; Based on the injection pressure change data and the water flow characteristics during the gas injection stage, dynamic balance evolution is performed to generate dynamic balance characteristics during the gas injection stage; Calculate the real-time pressure of the compression pump based on the pressure loading stage; Fitting the time-series pressure change of the real-time pressure of the compression pump to obtain the time-series pressure change curve in the pressure loading stage; Based on the storage and discharge stages, the transient flow change analysis of the water outlet valve is carried out to obtain the transient flow change characteristics of the water outlet valve; Calculate the water source temperature during the storage and water discharge stages, and extract the water outlet temperature of the bubble machine; Perform regular water outlet state detection on the transient flow change characteristics of the water outlet valve and the water outlet temperature of the bubble machine, thereby generating regular water outlet state characteristics; Based on the state fluctuation range characteristics of the water source preparation stage, the water flow characteristics of the gas injection stage, the time-series pressure change curve of the pressure loading stage and the conventional water discharge state characteristics, the normalized operation state characteristics of each stage are analyzed to obtain the operation state characteristics of each working stage.

[0039] In this embodiment, first, multiple conventional working stages of the bubble water machine are clarified, including the water source preparation stage, the gas injection stage, the pressure loading stage, and the storage and water discharge stage. These stages are key links in the operation of the bubble water machine, each with different operating characteristics and monitoring requirements. Confirm the historical operation log of the bubble water machine to ensure that it contains the timestamp, operating status, and related monitoring parameters of each stage. These data will be used for subsequent stage analysis. Use data analysis tools (such as Python's pandas library) to analyze the historical operation log and extract the start and end time of each working stage. Read the operation log data and identify the identifiers of each working stage. According to the timestamp and state changes, the operating time period of each stage is divided. Output the identified working stage data, including the start time, end time, and duration of each stage. This result will provide a basis for subsequent state monitoring and feature extraction. In the water source preparation stage, monitor the working state parameters of each subsystem, including temperature, pressure, and water flow. These parameters are crucial for evaluating the working efficiency of the system. Collect monitoring data in the water source preparation stage to ensure the integrity and accuracy of the data. Use sensors and monitoring equipment to record relevant parameters and organize the data into a unified format. Use data analysis tools to calculate the fluctuation range of state monitoring parameters of multiple subsystems. Calculate the maximum, minimum and average values ​​of each parameter to obtain its fluctuation range. Generate data visualization charts to show the fluctuation of parameters. Output the state fluctuation range characteristics of the water source preparation stage, including the fluctuation range data and visualization results of each monitoring parameter. This will help identify the stability and performance of the system at this stage. In the gas injection stage, extract the carbon dioxide injection pressure data of the gasification chamber to evaluate the effectiveness of gas injection. Perform statistical analysis on the injection pressure parameters to generate injection pressure change data. These data will be used for subsequent dynamic balance analysis. Use data analysis tools (such as Python's numpy or pandas) to process the pressure data. Count the injection pressures at different time points and calculate the average, maximum and minimum values. Generate time series data of pressure changes and analyze their changing trends. Output the injection pressure change data of the gas injection stage and generate corresponding visualization charts to show the changes in pressure over time. This will provide data support for the subsequent dynamic balance evolution. In the gas injection stage, calculate the water flow characteristics to evaluate the mixing effect of gas and water. Collect water flow monitoring data during the gas injection phase to ensure data integrity and accuracy. Use data analysis tools to calculate water flow characteristics.

[0040] Count the water flow rate in each time period, and calculate its average value and fluctuation range. Generate a time series graph of water flow change to show its dynamic characteristics. Output the water flow characteristic data of the gas injection stage and generate a visualization graph to analyze the effect of gas injection. Based on the injection pressure change data and the water flow characteristics of the gas injection stage, perform dynamic equilibrium evolution analysis to evaluate the interaction between gas and water flow. Combine the injection pressure change data with the water flow characteristics to analyze the relationship between them. Use statistical analysis tools (such as regression analysis) to quantify the dynamic balance of gas injection. Evaluate the correlation between injection pressure and water flow. Generate dynamic equilibrium characteristic data combined with dynamic changes. Output the dynamic equilibrium characteristic data of the gas injection stage and generate a visualization graph to show the dynamic relationship between gas and water flow. During the pressure loading stage, monitor the real-time pressure of the compression pump to evaluate the working status of the system. Collect real-time pressure monitoring data to ensure the accuracy and completeness of the data. Use data analysis tools to monitor and analyze the real-time pressure. Record the pressure values ​​of the compression pump at different time points and calculate their average, maximum and minimum values. Generate a time series graph of pressure change to show the pressure change trend over time. During the storage and water discharge stage, analyze the transient flow changes of the water discharge valve to evaluate the water discharge efficiency. Collect the flow monitoring data of the water discharge valve to ensure the integrity and accuracy of the data. Use data analysis tools to analyze the transient flow changes of the water discharge valve. Count the changes in flow at different time points and calculate its transient flow characteristics. Generate a time series diagram of flow changes to show the dynamic changes of water discharge efficiency. Output the flow change characteristic data of the storage and water discharge stage, and generate a visual chart to facilitate the analysis of water discharge performance. During the storage and water discharge stage, extract the water outlet temperature of the bubble machine to evaluate the water outlet quality. Collect the monitoring data of the water outlet temperature to ensure its accuracy. Use data analysis tools to calculate the water outlet temperature data. Count the water outlet temperature in different time periods, calculate its average value and fluctuation range. Generate a time series diagram of the water outlet temperature change to show the dynamic characteristics of the water outlet quality. Perform regular water outlet state detection based on the transient flow characteristics and water outlet temperature of the water outlet valve to evaluate the stability and quality of the water outlet state. Output the water outlet temperature data and regular water outlet state characteristics, and generate a visual chart to facilitate the analysis of water outlet quality. Integrate the state fluctuation range characteristics of the water source preparation stage, the water flow characteristics of the gas injection stage, the time-series pressure change curve of the pressure loading stage, and the conventional water output state characteristics. Analyze the normalized operating state characteristics of each stage to evaluate the overall performance of the bubble water machine. Use data analysis tools to conduct a comprehensive analysis of the characteristics of each stage. Combine the characteristic data of each stage to establish a comprehensive state model. Calculate the key performance indicators (KPI) of each stage to evaluate the operating efficiency of the system. Output the operating state characteristic data of each working stage and generate a visual chart to show the overall performance characteristics of the bubble water machine.

[0041] In this embodiment, refer to Figure 3, is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Step S21: performing full-process operation dependency analysis based on the historical operation log of the bubble water machine to obtain full-process operation dependency data; Step S22: performing end-to-end change tracking of normalized operation based on the full-process operation dependency data, thereby generating a normalized operation state change trajectory; Step S23: performing state-related multivariate regression modeling on the multi-stage operation state time series change characteristic data based on the normalized operation state change trajectory to obtain an operation state time series change network; Step S24: Evolve the global dynamic profile of the operation status time series change network and construct a global profile of normalized operation.

[0042] In this embodiment, the historical operation log of the sparkling water machine is collected to ensure data integrity. The log should contain status information, timestamps, operating parameters, etc. of each working stage. These data need to be integrated into a format suitable for analysis to clarify the dependencies between the various working stages. For example, whether the water source preparation stage is successful affects the start of the gas injection stage, and whether the pressure loading stage depends on the completion of gas injection. These relationships will become the basis for the full-process operation dependency analysis. Data analysis tools (such as Python's pandas and networkx libraries) are used for dependency analysis. The specific steps are as follows: Convert the historical operation log into a graph structure (nodes represent working stages) The dependency relationship between each stage is identified by traversing the log records, and the dependency graph is constructed. The full-process operation dependency data is output, including the dependency graph and description of each stage. The dependency relationship can be displayed using graphical tools for subsequent analysis and understanding. The key indicators of the change of normalized operation status are determined, including the running time and performance parameters of each stage, so as to track the change process from the beginning to the end, integrate the full-process operation dependency data, and combine it with the actual operation data to form a complete data set for normalized operation. Use data analysis and visualization tools (such as matplotlib and seaborn) to track changes. According to the dependency diagram, the state changes in each stage are analyzed, the end-to-end change trajectory is recorded, and a change tracking diagram is generated to show the state changes of the entire process from water source preparation to water discharge. The normalized operation state change trajectory data is output, and a visual chart is generated to facilitate the analysis of state changes and their causes in different stages. This provides clear time series data for subsequent analysis. Through the state change trajectory data, a multivariate regression model is constructed to analyze the correlation between the states, identify the key factors affecting system performance, and extract key features from the normalized operation state change trajectory, such as the operating time, flow, pressure and other parameters of each stage, to form a data set suitable for modeling. Statistical analysis tools (such as Python's statsmodels or scikit-learn) are used for multivariate regression modeling. The extracted features are used as independent variables, and system performance indicators (such as water quality, energy consumption, etc.) are used as dependent variables. Regression analysis is performed to evaluate the fitting effect of the model, and statistical indicators such as R² value and F test are used for Model verification, output the data of the operating status time-series change network, including the coefficients and statistical significance of the regression model, and generate visual charts to show the correlation between each state. This will provide an important basis for understanding the dynamic characteristics of system operation. By analyzing the operating status time-series change network, a global portrait of normalized operation is constructed to facilitate the evaluation and monitoring of the overall system performance. The output results of the multivariate regression model are integrated to identify the key factors and operating modes of system performance. Visualization tools (such as Gephi or networkx) are used to perform dynamic evolution analysis on the operating status time-series change network. The regression model results are converted into a dynamic network structure. Nodes represent different operating states, and edges represent the strength of association between them. Through dynamic network analysis, key nodes and paths are identified to show the changing trend of the operating state. The constructed global portrait of normalized operation is output, including the visualization results of the dynamic network, showing the relationship between different states and their evolution. This provides an intuitive reference for subsequent decision support.

[0043] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Step S31: monitoring the vibration audio signal of the bubble water machine in real time based on the sensor; Step S32: performing multi-time-frequency decomposition on the operating vibration audio signal of the bubble water machine to generate operating vibration frequency domain data; Step S33: performing dynamic vibration frequency calculation on the operation vibration frequency domain data to generate dynamic operation vibration frequency characteristics; Step S34: performing time series fluctuation amplitude recognition on the dynamic operation vibration frequency characteristics to obtain the operation vibration amplitude characteristics; Step S35: performing an operation state evolution of the bubble water machine on the operation vibration amplitude characteristics, thereby obtaining an operation state vibration evolution characteristic.

[0044] In this embodiment, a suitable vibration sensor (such as an accelerometer or a microphone) is selected and installed at key locations of the bubble water machine, such as the pump, vaporization chamber, and valve, to ensure accurate capture of the vibration audio signal of the device during operation, and to build a real-time monitoring system, including the connection between the sensor and the data acquisition device, to ensure that the sensor can transmit the vibration audio signal to the data acquisition system in real time, and to use a suitable sampling frequency (such as 10kHz) to capture subtle changes in the vibration signal. At various stages of the bubble water machine operation (such as water source preparation, gas injection, etc.), the vibration audio signal is continuously collected and stored. The data should include timestamps, vibration intensity and other information for subsequent analysis. The collected audio signals are pre-processed, including De-noising and filtering to eliminate environmental interference and ensure the accuracy of the signal. Digital signal processing (DSP) technology can be used to process the signal, and appropriate time-frequency analysis methods such as short-time Fourier transform (STFT), wavelet transform or Hilbert-Huang transform (HHT) can be selected to perform multi-time-frequency decomposition of the vibration audio signal and extract richer frequency domain information. Data analysis tools (such as Python's scipy or numpy libraries) are used to perform time-frequency decomposition of the vibration audio signal. The specific steps are as follows: segment the original audio signal, apply the selected time-frequency analysis method to obtain the frequency components in each time window, generate a time-frequency diagram to show the changing characteristics of the signal in time and frequency, and output the running vibration frequency domain data. The data includes frequency components and amplitude information in different time periods. These data will provide a basis for subsequent analysis to identify the working status and potential faults of the equipment. By analyzing the frequency domain data, the dynamic vibration frequency characteristics of the bubble water machine during operation are calculated to identify the working status of the equipment. The data analysis tool is used to calculate the dynamic vibration frequency of the frequency domain data, and the frequency components are extracted from the results of multi-time-frequency decomposition. The main frequency in each time period is calculated, the change of frequency over time is recorded, and its stability and fluctuation are analyzed. The dynamic operation vibration frequency characteristics are extracted in combination with the frequency data. The mean, standard deviation and rate of change of frequency can be calculated. The generated dynamic operation vibration frequency characteristic data is output, and a visual chart is generated to show the dynamic operation vibration frequency characteristic data. The changing trend of vibration frequency over time will help identify the normal operating status of the equipment and its potential abnormalities. The dynamic operating vibration frequency characteristics are analyzed in time series to identify the changing characteristics of the vibration amplitude and evaluate the operating status of the equipment. The dynamic vibration frequency characteristics are calculated using statistical analysis methods to calculate the fluctuation amplitude. The dynamic frequency data is analyzed in time series to calculate the vibration amplitude in each time period (such as the difference between the maximum and minimum values), identify important fluctuation events, record the time points and amplitudes of abnormal fluctuations, output the operating vibration amplitude characteristic data, and generate a visual chart to show the changes in the vibration amplitude over time. This will provide an important basis for the health assessment of the equipment. Based on the operating vibration amplitude characteristics, the operating state evolution of the sparkling water machine is analyzed.In order to identify the differences in vibration characteristics under different working conditions, use data analysis tools to analyze the evolution of vibration amplitude characteristics, combine vibration amplitude characteristics with the working stage data of the equipment, analyze the changes in vibration amplitude under different stages, identify the characteristic patterns of each stage, evaluate the stability and reliability of the equipment under different working conditions, output the operating state vibration evolution characteristic data, and generate visual charts to show the vibration characteristics under different working conditions, which will help to judge the operating health of the sparkling water machine and provide a basis for maintenance.

[0045] In this embodiment, step S4 includes the following steps: Step S41: identifying the current operation stage of the real-time working monitoring image based on the historical operation log of the bubble water machine, thereby obtaining the current working stage of the bubble water machine; Step S42: performing visual semantic recognition on the real-time working monitoring image and marking a plurality of water flow bubble nodes; Step S43: Calculating the bubble distribution density of multiple water flow bubble nodes to obtain water flow bubble distribution density characteristics; Step S44 performs bubble shape and size analysis on multiple water bubble nodes to generate bubble shape and size data; Step S45: mining the comprehensive bubble situation of the water bubble distribution density characteristics and the bubble shape and size data, thereby generating bubble situation characteristics; Step S46: Performing a staged carbonization state analysis on the operating state vibration evolution characteristics and the bubble state characteristics, so as to obtain the carbonization state characteristics of the current stage.

[0046] In this embodiment, the historical operation logs of the bubble water machine are collected to ensure that the data contains the timestamps, status information and related parameters of each working stage. These data will be used to identify the current operation stage and determine the working stage of the bubble water machine, such as water source preparation, gas injection, pressure loading, storage and water discharge. The characteristics and identification of each stage will serve as the basis for subsequent analysis. The historical operation logs are analyzed using data analysis tools (such as Python's pandas and numpy libraries). The specific steps are as follows: sort the historical logs according to the timestamp, identify the last complete stage, and match it with the timestamp of the real-time monitoring image. Combined with the real-time data, the current working stage of the bubble water machine can be determined. Simple conditional judgment or machine Machine learning classification methods are used to collect real-time working monitoring images, ensure that the image quality is high enough to support subsequent visual semantic recognition, select appropriate computer vision and deep learning models (such as YOLO, MaskR-CNN, etc.) for image analysis to identify bubble nodes in the water flow, use deep learning frameworks (such as TensorFlow or PyTorch) to process real-time monitoring images, preprocess images (such as normalization, size adjustment), ensure that the input data meets the model requirements, use the trained model to infer the images, identify the bubble nodes in the water flow, and mark their locations, output the marked images, display the identified water flow bubble nodes and their location information, and generate statistics on the number and distribution of bubble nodes. This will provide a basis for subsequent density calculation and morphological analysis. Collect the marked bubble node data, including the location information of each node (such as coordinates), select appropriate density calculation methods, such as kernel density estimation (KDE) or grid-based statistical methods, to evaluate the distribution density of water flow bubbles, use data analysis tools (such as Python's scipy and scikit-learn libraries) for density calculation, input the coordinate data of the bubble node into the density estimation model, calculate the bubble density of each area, generate a bubble distribution density map, show the bubble density changes in different areas, output the water flow bubble distribution density characteristic data, and generate a visual chart to show the spatial distribution of bubble density, which will provide a basis for the subsequent bubble situation characteristics. The analysis provides important information, collects the data of the marked bubble nodes, ensures that the image information of each node can be used for morphological analysis, defines the key features of the bubble morphology, such as diameter, area, roundness, etc., which will be used for subsequent analysis, uses image processing tools (such as OpenCV) to perform morphological size analysis on the bubble nodes, extracts the contours of the marked bubbles, calculates the diameter and area of ​​each bubble, generates a bubble morphological size data set, summarizes the morphological features of each bubble, outputs the bubble morphological size data, including the key features of each bubble, and generates a visualization to show the morphological distribution of the bubbles, which will provide a basis for subsequent comprehensive situation analysis, and combines the distribution density characteristics and morphological size data of water flow bubbles to conduct a comprehensive situation analysis of bubbles.In order to identify the overall state of bubbles, the bubble distribution density characteristics are combined with the bubble shape and size data to form a comprehensive data set. The comprehensive data is mined using data mining tools and techniques (such as cluster analysis, principal component analysis, etc.). Clustering algorithms (such as K-means) are used to identify different state modes of bubbles, and principal component analysis is performed to extract the main factors affecting the bubble state. The bubble state characteristic data is output, and a visual chart is generated to show the comprehensive state of the bubbles. This will provide an important basis for the subsequent carbonation state analysis. Based on the operating state vibration evolution characteristics and bubble state characteristics, the carbonation state of the bubble water machine is analyzed in stages to evaluate the design. To monitor the health of the equipment, the vibration evolution characteristics of the operating state are combined with the bubble state characteristics to form a comprehensive data set for carbonization state analysis. Data analysis tools (such as Python's statsmodels or scikit-learn) are used to perform staged carbonization state analysis, identify the carbonization characteristics under each state, and analyze the relationship between it and the vibration and bubble state. Through regression analysis or classification models, the carbonization state characteristics of different working stages are evaluated, the carbonization state characteristic data of the current stage is output, and a visual chart is generated to show the relationship between the carbonization state and the equipment operation characteristics, which will provide an important basis for equipment maintenance and optimization.

[0047] In this embodiment, step S5 includes the following steps: Step S51: performing local characteristic time series change analysis on the carbonization state characteristics of the current stage, thereby generating local carbonization state change characteristics; Step S52: mining the time series variation law of the local carbonization state variation characteristics to generate a stage-by-stage time series variation law; Step S53: simulating the evolution of the bubble state in all stages according to the phased time series variation law to obtain simulation data of the evolution of the bubble state in all stages; Step S54: Based on the global portrait of normalized operation, multi-period rolling fault prediction is performed on the simulation data of bubble state evolution in all stages to generate fault prediction data for multiple time periods.

[0048] In this embodiment, the carbonization state characteristic data of the current stage are collected, including vibration characteristics, bubble state characteristics, etc. These data should cover a sufficient time range for time series change analysis. Determine the local characteristics of the carbonization state, such as carbonization degree, bubble density, vibration frequency, etc. These characteristics will serve as the basis for analysis and can reflect subtle changes in the operating state. Use data analysis tools (such as Python's pandas and numpy libraries) to perform time series change analysis on local characteristics. The specific steps are as follows: Calculate the change of each feature in the time dimension and generate a corresponding time series data set. Analyze the change trend of each local feature through the sliding window method to determine its fluctuation range and change rate. Output local carbonization state change characteristic data, including a chart of the change of characteristic values ​​over time. This will provide basic data for subsequent time series change law mining and help identify potential abnormal trends. By analyzing the local carbonization state change characteristics, identify its time series change law to discover potential patterns and trends. Integrate the local feature change data to form a complete data set, including information such as change rate and fluctuation amplitude. Use time series analysis tools (such as statsmodels or scikit-learn) to mine time series change laws. Autocorrelation analysis (ACF) and partial autocorrelation analysis (PACF) are used to identify the temporal dependency between features. Seasonal decomposition or trend analysis methods are used to extract the main change patterns of local features and analyze their periodicity and trend. Data of phased temporal change patterns are output, including the change patterns and trend graphs of key features. This will provide an important regular basis for the subsequent simulation of bubble state evolution. According to the identified phased temporal change patterns, a bubble state evolution model is constructed to simulate the bubble state changes during the entire operation phase. The temporal change patterns of local features are integrated to define the input parameters and model structure of the bubble state evolution. Simulation tools and algorithms (such as system dynamics models or proxy models) are used to simulate the evolution of bubble states. According to the phased temporal change patterns, the initial conditions and parameters of the bubble state evolution are set. Run the simulation program to generate simulation data of the bubble state evolution in all stages and record the state changes at each time step. Output simulation data of the bubble state evolution in all stages and generate visualization charts to show the dynamic changes of the bubble state during the simulation process. This will provide data support for fault prediction. Use global portraits and bubble state evolution simulation data to perform multi-period fault rolling predictions and identify potential fault risks. Integrate normalized operation global portraits and bubble state evolution simulation data to prepare for fault prediction analysis. Use machine learning and prediction models (such as LSTM or random forest) for fault rolling prediction. Use bubble state evolution simulation data as input to train the prediction model and identify the relationship between features and faults. Implement rolling predictions to predict fault risks in multiple future periods based on historical data and current status.Output fault prediction data for multiple time periods and generate visual charts to show the change of fault risk over time and potential fault points. This will provide important reference for equipment maintenance and management.

[0049] In this embodiment, step S6 includes the following steps: Step S61: Calculate the failure probability of each time period for the failure prediction data of multiple time periods, and generate the failure probability of each time period; Step S62: extracting the maximum failure probability time period based on the failure probability of each time period; Step S63: locating and tracing the fault based on the fault probability of each time period, thereby generating a fault tracing point; Step S64: making a dynamic fault diagnosis decision on the fault source tracing point, thereby obtaining a dynamic fault diagnosis strategy; Step S65: Optimize cloud-edge collaborative fault management for the dynamic fault diagnosis strategy and build an intelligent fault prediction optimization model.

[0050] In this embodiment, the fault prediction data for each time period generated from the previous steps are collected, including bubble state evolution simulation data and related features. These data will serve as the basis for calculating the fault probability. A suitable statistical method is selected to calculate the fault probability, such as the frequency method based on historical data, the Bayesian inference method, or the probability predicted by the machine learning model. A data analysis tool (such as Python's scipy and pandas libraries) is used to calculate the fault probability for each time period. The state data of each time period is analyzed to calculate the frequency of fault occurrence. For example, the ratio of the number of actual faults occurring in the time period to the total number of observations is counted. If a machine learning model is used, the features of each time period are extracted and input into the trained model to output the fault probability of the time period. Failure probability, output the failure probability data of each time period, and generate a visual chart to show the changes in failure probability in different time periods. This will provide basic data for subsequent failure probability extraction and analysis. Determine the "maximum failure probability period" as the period with the highest failure probability among all time periods. This is crucial for subsequent failure analysis. Use data analysis tools (such as numpy and pandas) to analyze the failure probability data, identify the period with the highest failure probability through the max() function or other statistical methods, record the specific information of the period, including timestamp, failure probability value and related features, output the information of the maximum failure probability period, and generate a visual chart to show the failure probability characteristics of the period, as well as the comparison with other periods. By comparison, this will provide a specific goal for the subsequent fault tracing analysis. By analyzing the characteristics of the maximum fault probability period, the specific cause or location that may cause the fault is identified, and the data of the maximum fault probability period, including the relevant operating status, bubble status and vibration characteristics, etc., are integrated to facilitate fault tracing analysis. Fault tree analysis (FTA) or fishbone diagram and other methods are used in combination with machine learning models to locate and trace the fault, identify all features related to the maximum fault probability period, and analyze their contribution to the fault. Through model prediction and feature importance analysis, the fault tracing point is determined, such as a specific sensor failure, abnormal operating conditions, etc. The information of the fault tracing point is output, including possible causes of the fault and their impact, and a visual diagram is generated. , display the fault tracing path, which will provide a basis for subsequent dynamic fault diagnosis decisions. For the identified fault tracing points, formulate corresponding dynamic fault diagnosis strategies to quickly respond to and handle faults. Select appropriate decision models (such as rule-based decision systems or machine learning classification models) for dynamic fault diagnosis. Use decision analysis tools to perform dynamic fault diagnosis on fault tracing points. According to the characteristics of the fault tracing points, set fault diagnosis rules or conditions, such as threshold judgment, pattern recognition, etc. Run the model, diagnose the current state, output fault diagnosis results and recommended treatment measures, and combine dynamic fault diagnosis strategies with fault management systems through cloud-edge collaboration to improve the efficiency and intelligence level of fault management. Design a cloud-edge collaborative architecture.To ensure the high efficiency of real-time data transmission and processing, the edge device is responsible for data collection and preliminary analysis, while the cloud performs in-depth analysis and decision support. Cloud computing platforms and edge computing devices are used to optimize fault management, and dynamic fault diagnosis strategies are deployed to the cloud, and connected with the data collection system of edge devices to achieve real-time data flow and information sharing. Through machine learning and optimization algorithms, fault diagnosis strategies are continuously updated and improved to improve prediction accuracy and response speed. The constructed intelligent fault prediction optimization model is output, including cloud-edge collaborative workflows, fault management strategies and their performance indicators, and a visualization diagram is generated to show the overall process and effect of fault management, which will provide strong support for the intelligent maintenance and management of sparkling water machines.

[0051] In this embodiment, a remote monitoring and fault prediction system for a sparkling water machine is provided, which is used to execute the remote monitoring and fault prediction method for a sparkling water machine as described above, including: The time series change module is used to obtain the historical operation logs and real-time working monitoring images of the bubble water machine; perform multi-stage operation status feature analysis based on the historical operation logs of the bubble water machine, and generate multi-stage operation status time series change characteristic data; The normalization portrait module is used to track the end-to-end changes of normalized operation based on the multi-stage operation status time series change characteristic data, and to perform global dynamic portrait evolution to build a global portrait of normalized operation; A vibration evolution module is used to monitor the vibration audio signal of the bubble water machine in real time based on a sensor; perform multi-time-frequency decomposition on the vibration audio signal of the bubble water machine and the bubble water machine operation state evolution, so as to obtain the vibration evolution characteristics of the operation state; The carbonization state module is used to identify the current operation stage of the real-time working monitoring image and perform a staged carbonization state analysis based on the vibration evolution characteristics of the operation state, so as to obtain the carbonization state characteristics of the current stage; The multi-period fault prediction module is used to simulate the evolution of the bubble state in all stages based on the global portrait of normalized operation and to perform multi-period fault rolling prediction to generate fault prediction data for multiple periods. The fault diagnosis decision module is used to calculate the fault probability of each time period for the fault prediction data of multiple time periods, and then make dynamic fault diagnosis decisions to build an intelligent fault prediction optimization model.

[0052] By combining real-time and historical data, the present invention can comprehensively obtain the operating status of the bubble water machine at different working stages, provide basic data support for subsequent analysis, and analyze the operating data at different stages to identify the key characteristics of each stage (such as pressure, flow, temperature, etc.), providing a comprehensive understanding of the operating status of the equipment, which helps to determine the health status of the equipment at each stage, as well as the potential failure risks at each stage. Based on the staged feature analysis, the time series change data of the bubble water machine operation process can be generated. This data helps to identify the trends and fluctuations of the equipment in operation, and to warn of possible abnormal conditions in advance. By constructing a global portrait of normalized operation, the system can continuously track the operating changes of the equipment and establish normal equipment operation. The system can continuously update the global portrait based on real-time data, provide continuous equipment health status monitoring, and identify long-term trends and short-term abnormal fluctuations in equipment operation through end-to-end change tracking. It can identify the phenomenon of equipment gradually deviating from the normal state in advance, which is helpful for early detection of potential faults. Through real-time monitoring of vibration audio signals by sensors, it is possible to obtain timely changes in the operating status of the bubble water machine, especially the vibration status of key components, which is helpful for rapid response when equipment abnormalities occur. Multi-time-frequency decomposition can deeply analyze the frequency and amplitude of vibration signals and identify equipment failures. The vibration characteristics of the equipment under different operating conditions. This refined analysis helps to identify abnormal modes in operation, especially vibration signals related to faults. By analyzing the vibration evolution characteristics, signs of potential faults can be identified in advance, such as wear, looseness, bearing failure and other problems. This provides strong data support for fault diagnosis and maintenance. According to real-time images and vibration data, the carbonization state of the sparkling water machine at different working stages can be analyzed. This staged analysis can provide a combination of vibration signals and image recognition to more accurately identify carbonization phenomena and their impact on equipment, which helps to locate the root cause of the problem and provide guidance for repair. Based on the refined analysis of the carbonization state, maintenance personnel can carry out targeted carbonization problems at specific stages. Processing, avoiding unnecessary costs caused by comprehensive inspections, and improving maintenance efficiency. By simulating the evolution of the bubble state of the bubble water machine in different working stages, it is possible to fully grasp the performance of the equipment under different loads and environmental conditions, which makes the equipment failure prediction more accurate and covers the entire operation cycle. Through rolling predictions for multiple time periods in the future, it is possible to accurately infer the health status of the equipment in the future based on the current status. The system can foresee the possibility of future failures in advance and help operation and maintenance personnel take preventive measures in advance to avoid sudden failures. Multi-period failure prediction provides a long-term early warning model, so that equipment maintenance is not limited to the current stage, but maintenance work can be planned and arranged in advance, minimizing equipment downtime and maintenance costs.By calculating the probability of fault prediction data for each time period, the system can provide different fault probabilities for different time points, which makes fault prediction not only static, but also dynamically changing, which helps to accurately identify risks. By combining the time period fault probability, the system can perform intelligent fault diagnosis according to the current operating status of the equipment. The ability of dynamic decision-making can help the system optimize the fault diagnosis strategy according to real-time data. By building an intelligent fault prediction optimization model, the system can not only make accurate fault predictions, but also continuously optimize the prediction strategy according to historical data and real-time data. As the amount of data increases, the prediction accuracy of the system will continue to improve, thereby better supporting equipment management and maintenance.

[0053] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0054] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for remote monitoring and fault prediction of a sparkling water machine, characterized in that: The following steps are involved: Step S1: Obtain the historical operation log and real-time working monitoring image of the bubble water machine; perform multi-stage operation state characteristic analysis based on the historical operation log of the bubble water machine to generate multi-stage operation state time series change characteristic data; Step S2: Based on the multi-stage operation status time series change characteristic data, end-to-end change tracking of normalized operation is performed, and global dynamic portrait evolution is performed to construct a global portrait of normalized operation; Step S3: monitoring the bubble water machine operation vibration audio signal in real time based on the sensor; performing multi-time-frequency decomposition on the bubble water machine operation vibration audio signal and the bubble water machine operation state evolution, thereby obtaining the operation state vibration evolution characteristics; Step S4: identifying the current operation stage of the real-time working monitoring image, and performing a staged carbonization state analysis based on the vibration evolution characteristics of the operation state, so as to obtain the carbonization state characteristics of the current stage; Step S5: Based on the global portrait of normalized operation, the bubble state evolution of the current carbonization state characteristics is simulated in all stages, and multi-period rolling fault prediction is performed to generate fault prediction data for multiple periods; Step S6: Calculate the fault probability of each time period for the fault prediction data of multiple time periods, make dynamic fault diagnosis decisions, and build an intelligent fault prediction optimization model.

2. The method for remote monitoring and fault prediction of a sparkling water machine according to claim 1, characterized in that: The specific steps of step S1 are: Step S11: Obtaining the historical operation log and real-time working monitoring image of the bubble water machine; Step S12: performing a multi-stage operation status characteristic analysis based on the historical operation log of the bubble water machine, thereby obtaining the operation status characteristics of each working stage; Step S13: performing logical correlation analysis of upper and lower stage operations based on the historical operation log of the bubble water machine to generate logical rules for operation between stages; Step S14: Analyze the operation status time series changes of the operation status characteristics of each working stage according to the operation logic rules between stages, so as to generate multi-stage operation status time series change characteristic data.

3. The method for remote monitoring and fault prediction of a sparkling water machine according to claim 2, characterized in that: The specific steps of step S14 are: Identify multiple regular working stages of the bubble water machine based on the historical operation log of the bubble water machine; the multiple regular working stages of the bubble water machine include a water source preparation stage, a gas injection stage, a pressure loading stage, and a storage and water discharge stage; Monitor the working status parameters of each subsystem during the water source preparation stage and extract multiple subsystem status monitoring parameters, wherein the status monitoring parameters include temperature, pressure and water flow; The fluctuation range of multiple subsystem status monitoring parameters is calculated to obtain the state fluctuation range characteristics of the water source preparation stage; During the gas injection stage, the CO2 injection pressure in the gasification chamber is counted and the injection pressure parameters are extracted; Performing pressure variation analysis on injection pressure parameters to generate injection pressure variation data; The water flow rate is calculated during the gas injection stage to obtain the water flow rate characteristics during the gas injection stage; Based on the injection pressure change data and the water flow characteristics during the gas injection stage, dynamic balance evolution is performed to generate dynamic balance characteristics during the gas injection stage; Calculate the real-time pressure of the compression pump based on the pressure loading stage; Fitting the time-series pressure change of the real-time pressure of the compression pump to obtain the time-series pressure change curve in the pressure loading stage; Based on the storage and discharge stages, the transient flow change analysis of the water outlet valve is carried out to obtain the transient flow change characteristics of the water outlet valve; Calculate the water source temperature during the storage and water discharge stages, and extract the water outlet temperature of the bubble machine; Perform regular water outlet state detection on the transient flow change characteristics of the water outlet valve and the water outlet temperature of the bubble machine, thereby generating regular water outlet state characteristics; Based on the state fluctuation range characteristics of the water source preparation stage, the water flow characteristics of the gas injection stage, the time-series pressure change curve of the pressure loading stage and the conventional water discharge state characteristics, the normalized operation state characteristics of each stage are analyzed to obtain the operation state characteristics of each working stage.

4. The method for remote monitoring and fault prediction of a sparkling water machine according to claim 1, characterized in that: The specific steps of step S2 are: Step S21: performing full-process operation dependency analysis based on the historical operation log of the bubble water machine to obtain full-process operation dependency data; Step S22: performing end-to-end change tracking of normalized operation based on the full-process operation dependency data, thereby generating a normalized operation state change trajectory; Step S23: performing state-related multivariate regression modeling on the multi-stage operation state time series change characteristic data based on the normalized operation state change trajectory to obtain an operation state time series change network; Step S24: Evolve the global dynamic profile of the operation status time series change network and construct a global profile of normalized operation.

5. The method for remote monitoring and fault prediction of a sparkling water machine according to claim 1, characterized in that: The specific steps of step S3 are: Step S31: monitoring the vibration audio signal of the bubble water machine in real time based on the sensor; Step S32: performing multi-time-frequency decomposition on the operating vibration audio signal of the bubble water machine to generate operating vibration frequency domain data; Step S33: performing dynamic vibration frequency calculation on the operation vibration frequency domain data to generate dynamic operation vibration frequency characteristics; Step S34: performing time series fluctuation amplitude recognition on the dynamic operation vibration frequency characteristics to obtain the operation vibration amplitude characteristics; Step S35: performing an operation state evolution of the bubble water machine on the operation vibration amplitude characteristics, thereby obtaining an operation state vibration evolution characteristic.

6. The method for remote monitoring and fault prediction of a sparkling water machine according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: identifying the current operation stage of the real-time working monitoring image based on the historical operation log of the bubble water machine, thereby obtaining the current working stage of the bubble water machine; Step S42: performing visual semantic recognition on the real-time working monitoring image and marking a plurality of water flow bubble nodes; Step S43: Calculating the bubble distribution density of multiple water flow bubble nodes to obtain water flow bubble distribution density characteristics; Step S44 performs bubble shape and size analysis on multiple water flow bubble nodes to generate bubble shape and size data; ： Step S45: performing comprehensive bubble situation mining on the water bubble distribution density characteristics and bubble shape and size data, thereby generating bubble situation characteristics; Step S46: Performing a staged carbonization state analysis on the operating state vibration evolution characteristics and the bubble state characteristics, thereby obtaining the carbonization state characteristics of the current stage.

7. The method for remote monitoring and fault prediction of a sparkling water machine according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: performing local characteristic time series change analysis on the carbonization state characteristics of the current stage, thereby generating local carbonization state change characteristics; Step S52: mining the time series variation law of the local carbonization state variation characteristics to generate a stage-by-stage time series variation law; Step S53: simulating the evolution of the bubble state in all stages according to the phased time series variation law to obtain simulation data of the evolution of the bubble state in all stages; Step S54: Based on the global portrait of normalized operation, multi-period rolling fault prediction is performed on the simulation data of bubble state evolution in all stages to generate fault prediction data for multiple time periods.

8. The method for remote monitoring and fault prediction of a sparkling water machine according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: Calculate the failure probability of each time period for the failure prediction data of multiple time periods, and generate the failure probability of each time period; Step S62: extracting the maximum failure probability time period based on the failure probability of each time period; Step S63: locating and tracing the fault based on the fault probability of each time period, thereby generating a fault tracing point; Step S64: making a dynamic fault diagnosis decision on the fault source tracing point, thereby obtaining a dynamic fault diagnosis strategy; Step S65: Optimize cloud-edge collaborative fault management for the dynamic fault diagnosis strategy and build an intelligent fault prediction optimization model.

9. A remote monitoring and fault prediction system for a sparkling water machine, characterized in that: The method for remote monitoring and fault prediction of a sparkling water machine according to claim 1 comprises: The time series change module is used to obtain the historical operation logs and real-time working monitoring images of the bubble water machine; perform multi-stage operation status feature analysis based on the historical operation logs of the bubble water machine, and generate multi-stage operation status time series change characteristic data; The normalization portrait module is used to track the end-to-end changes of normalized operation based on the multi-stage operation status time series change characteristic data, and to perform global dynamic portrait evolution to build a global portrait of normalized operation; A vibration evolution module is used to monitor the vibration audio signal of the bubble water machine in real time based on a sensor; perform multi-time-frequency decomposition on the vibration audio signal of the bubble water machine and the bubble water machine operation state evolution, so as to obtain the vibration evolution characteristics of the operation state; The carbonization state module is used to identify the current operation stage of the real-time working monitoring image and perform a staged carbonization state analysis based on the vibration evolution characteristics of the operation state, so as to obtain the carbonization state characteristics of the current stage; The multi-period fault prediction module is used to simulate the evolution of the bubble state in all stages based on the global portrait of normalized operation and to perform multi-period fault rolling prediction to generate fault prediction data for multiple periods. The fault diagnosis decision module is used to calculate the fault probability of each time period for the fault prediction data of multiple time periods, and then make dynamic fault diagnosis decisions to build an intelligent fault prediction optimization model.

Images