A remote monitoring and analysis method and system for centrifuge data based on cloud computing

Abstract

The present invention relates to the field of data monitoring and analysis technology, including a remote monitoring and analysis method and system for centrifuge data based on cloud computers, including the following steps: collecting centrifuge operation index data, performing time tag synchronization, identifying and eliminating abnormal values ​​in the data, integrating data through data normalization processing, and generating synchronous monitoring data. In the present invention, by implementing centrifuge data frequency domain analysis on a cloud platform, the vibration signal spectrum and energy distribution are accurately identified, efficient marking and trend analysis of abnormal data are achieved, the meticulousness and response speed of abnormal detection are enhanced, and the monitoring of the centrifuge operation status is made more accurate. Combined with real-time trend analysis and historical fault data, not only the accuracy of fault prediction is improved, but also the classification of fault modes is made more scientific, the situation of false alarms and missed alarms is effectively reduced, the timeliness and pertinence of preventive maintenance measures are enhanced, and the operation efficiency and safety of the centrifuge are significantly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data monitoring and analysis, and in particular to a method and system for remote monitoring and analysis of centrifuge data based on a cloud computer. Background Art

[0002] The field of data monitoring and analysis technology focuses on collecting and analyzing data through various monitoring systems to optimize business operations and improve system performance. It collects data from various sources in real time by using advanced data collection tools such as sensors, network devices, and application programming interfaces. The analysis process involves using data mining, machine learning, and statistical methods to interpret data, identify trends, predict future behaviors, and discover potential anomalies or failures. It is widely used in areas such as resource optimization, cost control, security reinforcement, and improving service quality. In the information age, data monitoring and analysis provides enterprises with real-time decision support, which can respond to market changes and user needs more effectively.

[0003] Among them, the remote monitoring and analysis method of centrifuge data involves the use of remote technology to monitor the operating data of the centrifuge in real time, and analyze the data to optimize equipment performance and prevent failures. It can monitor its various operating parameters such as speed, temperature, vibration, etc. without physical contact with the equipment. It is especially useful for equipment located in difficult-to-access areas or equipment that requires high-cost maintenance. The main uses include improving the operating efficiency of the centrifuge, preventing equipment failures, reducing downtime, and extending the service life of the equipment, making equipment management more efficient, reducing operating costs, and improving production safety.

[0004] Although existing technologies are widely used in data monitoring and system performance optimization, conventional methods have certain limitations when processing real-time data from complex equipment such as centrifuges. Traditional monitoring systems usually use relatively basic data analysis methods, such as threshold judgment and simple statistical analysis, and fail to deeply explore the potential information in the data, resulting in the inability to accurately predict and quickly respond to subtle abnormal changes in the equipment, leading to the neglect of initial failures and the inability to effectively perform preventive maintenance. For example, failure to use frequency domain analysis may miss signal changes caused by minor vibration anomalies, and then fail to take timely measures to avoid greater equipment damage. In addition, existing technologies may take a long time to process and analyze large amounts of data, increase the burden of manual monitoring, and affect decision-making efficiency and the actual effectiveness of equipment management. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a remote monitoring and analysis method for centrifuge data based on cloud computers.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a remote monitoring and analysis method of centrifuge data based on cloud computer, comprising the following steps:

[0007] S1: Collect centrifuge operation index data, synchronize time tags, identify and remove outliers in the data, integrate data through data normalization, and generate synchronous monitoring data;

[0008] S2: Perform frequency domain analysis on the synchronous monitoring data through a cloud computer, identify the vibration signal spectrum of the centrifuge, analyze the signal peak and record the energy distribution, perform trend analysis on the data in combination with the energy distribution recording results, identify data points that exceed the normal range and mark them as abnormal data points by comparing with the preset performance baseline of the centrifuge, reveal potential operating abnormalities of the centrifuge, and generate status analysis results;

[0009] S3: Based on the state analysis result, cross-validate the abnormal data points with the historical fault records of the centrifuge, classify the abnormal data points using cluster analysis, identify the centrifuge fault mode according to the abnormal point classification results, sort each fault mode according to the frequency and severity, and generate anomaly detection records;

[0010] S4: Using the abnormality detection record, record the abnormality type and impact range of the centrifuge, send an alarm to the monitoring center through the cloud computing platform, and generate real-time remote monitoring alarm information.

[0011] As a further solution of the present invention, the step of acquiring the synchronous monitoring data is:

[0012] S111: Install sensors on key components of the centrifuge to collect operating data, record speed, temperature and vibration frequency, and form a raw data set;

[0013] S112: Based on the original data set, each data point is timestamped, and all outliers exceeding two standard deviations are removed to obtain a data set that has been purified through time synchronization;

[0014] S113: normalizing the data set that has undergone time synchronization purification, using the formula:

[0015] ;

[0016] Calculate the normalized value for each data point , get the synchronous monitoring data, where is the original value of the data point, represents all data points in the entire dataset, Represents the minimum value in the data set. Represents the maximum value in the data set.

[0017] As a further solution of the present invention, the step of recording the energy distribution is:

[0018] S211: receiving the synchronous monitoring data of the centrifuge through the cloud computing platform, converting the data into frequency domain using Fourier transform, and generating preliminary vibration signal spectrum results;

[0019] S212: Based on the preliminary vibration signal spectrum result, digital filtering is performed to remove noise, key frequency components of the vibration signal are extracted, and peak detection is performed to obtain a signal peak result;

[0020] S213: Using the signal peak value result, adopt the formula:

[0021] ;

[0022] Calculate the average energy distribution value , and obtain the energy distribution record, where Represents the total number of frequency points, Represents the absolute value of the amplitude at the i-th frequency point.

[0023] As a further solution of the present invention, the step of obtaining the state analysis result is:

[0024] S221: Based on the energy distribution record, perform time series analysis, model the change trend of energy data over time, analyze the autocorrelation and partial autocorrelation characteristics of the data, and generate energy trend analysis results;

[0025] S222: Compare the energy trend analysis result with the preset performance baseline of the centrifuge, identify the peak value beyond the normal range, capture the abnormal operation and performance degradation of the centrifuge, and calibrate the data points that deviate from the baseline by more than two standard deviations by setting a threshold to obtain the abnormal energy point identification result;

[0026] S223: Analyze the abnormal energy point identification result, analyze each abnormal point, analyze the correlation with the centrifuge operating conditions and historical maintenance records, and generate a state analysis result.

[0027] As a further solution of the present invention, the classification step of the abnormal data points is:

[0028] S311: According to the historical fault records of the centrifuge, the time, type and severity of all faults are counted, and a fault record database is established;

[0029] S312: extracting abnormal data points from the state analysis results, cross-validating with the fault record database, and obtaining matching abnormal data sets by analyzing the matching degree between the data points and the historical fault patterns;

[0030] S313: Perform cluster analysis on the matched abnormal data set, group the abnormal data according to the features, each group represents a potential fault cause, and generate abnormal data point classification results.

[0031] As a further solution of the present invention, the steps of obtaining the abnormality detection record are:

[0032] S321: According to the classification result of the abnormal data point, the classification result is compared with the known fault mode database, the fault type corresponding to each category is determined by feature matching, the fault mode represented by each classification is identified, and a fault mode identification list is generated;

[0033] S322: Analyze each failure mode according to the failure mode identification list, sort them according to the occurrence frequency and the severity of the operational impact and loss caused, and generate a sorted failure mode list;

[0034] S323: Based on the sorted fault mode list, actual information of each fault is recorded, including identification time, fault cause and maintenance measures, and an abnormality detection record is comprehensively generated.

[0035] As a further solution of the present invention, the step of obtaining the real-time remote monitoring alarm information is:

[0036] S411: Analyze the anomaly detection record, distinguish different anomaly types, and quantify the impact range of each anomaly, determine the impact range of the anomaly type, and generate anomaly classification and impact analysis results;

[0037] S412: Transmitting the anomaly classification and impact analysis results to the cloud computing platform through a security protocol, performing data quality verification, confirming the integrity of the received data, and generating verified transmission data;

[0038] S413: Based on the verified transmission data, according to the abnormality type and impact scope, record key real-time data and response strategies, automatically send them to the monitoring center, complete real-time alarm push, and output real-time remote monitoring alarm information.

[0039] A remote monitoring and analysis system for centrifuge data based on cloud computers, comprising:

[0040] The data acquisition module installs sensors on the key components of the centrifuge to collect operating data, timestamps each data point, removes outliers, performs normalization, calculates the normalized value of each data point, and obtains synchronous monitoring data;

[0041] The frequency domain analysis module receives the synchronous monitoring data of the centrifuge through the cloud computing platform, uses Fourier transform to perform frequency domain conversion on the data, extracts the key frequency components of the vibration signal and performs peak detection, calculates the average energy distribution value, and obtains the energy distribution record;

[0042] The trend modeling module performs time series analysis based on the energy distribution record, models the trend of energy data changing over time, compares it with the preset performance baseline of the centrifuge, analyzes the correlation with the centrifuge operating conditions and historical maintenance records, and generates a state analysis result;

[0043] The fault detection module establishes a fault record database based on the historical fault records of the centrifuge, extracts abnormal data points from the state analysis results, cross-validates with the fault record database, groups the abnormal data according to the features, and generates abnormal data point classification results;

[0044] The pattern comparison module compares the classification results of the abnormal data points with the known fault pattern database, analyzes each fault pattern, sorts them according to the frequency of occurrence and the severity of the operational impact and loss caused, records the actual information of each fault, and comprehensively generates anomaly detection records;

[0045] The real-time alarm module analyzes the anomaly detection records, determines the impact scope of each anomaly type, transmits them to the cloud computing platform through a security protocol, performs data quality verification, records key real-time data and response strategies, automatically sends them to the monitoring center, and outputs real-time remote monitoring alarm information.

[0046] Compared with the prior art, the advantages and positive effects of the present invention are:

[0047] In the present invention, by implementing the data frequency domain analysis of the centrifuge on the cloud platform, the spectrum and energy distribution of the vibration signal are accurately identified, the efficient marking and trend analysis of abnormal data are realized, the meticulousness and response speed of abnormal detection are enhanced, and the monitoring of the operation status of the centrifuge is made more accurate. Combined with real-time trend analysis and historical fault data, not only the accuracy of fault prediction is improved, but also the classification of fault modes is made more scientific, the situation of false alarms and missed alarms is effectively reduced, the timeliness and pertinence of preventive maintenance measures are enhanced, and the operation efficiency and safety of the centrifuge are significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a main step flow chart of the present invention;

[0049] Figure 2 This is a flowchart of acquiring synchronous monitoring data of the present invention;

[0050] Figure 3 It is a flow chart for recording energy distribution of the present invention;

[0051] Figure 4 A flowchart for obtaining the status analysis results of the present invention;

[0052] Figure 5 is a classification flow chart of abnormal data points of the present invention;

[0053] Figure 6 This is a flowchart for obtaining abnormal detection records of the present invention;

[0054] Figure 7 This is a flow chart for obtaining real-time remote monitoring alarm information of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0056] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0057] See also Figure 1 , a remote monitoring and analysis method for centrifuge data based on cloud computers, comprising the following steps:

[0058] S1: Collect centrifuge operation index data, synchronize time tags, identify and remove outliers in the data, integrate data through data normalization, and generate synchronous monitoring data;

[0059] S2: Perform frequency domain analysis on the synchronous monitoring data through cloud computers, identify the centrifuge vibration signal spectrum, analyze the signal peak and record the energy distribution, perform trend analysis on the data based on the energy distribution record results, and identify data points that exceed the normal range and mark them as abnormal data points by comparing with the preset performance baseline of the centrifuge, revealing the potential operation abnormalities of the centrifuge and generating status analysis results;

[0060] S3: Based on the state analysis results, cross-validate the abnormal data points with the historical centrifuge failure records, use cluster analysis to classify the abnormal data points, identify the centrifuge failure mode based on the abnormal point classification results, sort each failure mode according to the frequency and severity, and generate anomaly detection records;

[0061] S4: Utilize the anomaly detection record to record the anomaly type and impact range of the centrifuge, send alarms to the monitoring center through the cloud computing platform, and generate real-time remote monitoring alarm information.

[0062] Synchronous monitoring data includes data timestamps and normalized indicators; status analysis results include abnormal data points, vibration signal analysis results, energy distribution records and trend analysis reports; anomaly detection records include fault pattern recognition results, fault frequency and fault severity ranking records; real-time remote monitoring alarm information includes alarm type, impact range and warning level.

[0063] See also Figure 2 , the steps to obtain synchronous monitoring data are:

[0064] S111: Install sensors on key components of the centrifuge to collect operating data, record speed, temperature and vibration frequency, and form a raw data set;

[0065] Various sensors are installed on the key components of the centrifuge. The sensors can capture various data during the operation of the machine, such as rotation speed, temperature and vibration frequency. The data acquisition function of each sensor is realized through a built-in microprocessor, which can monitor and record data in real time without human intervention; these sensors are connected to a central data processing unit through the network. The unit is equipped with a high-speed data transmission interface to ensure the real-time and accuracy of data transmission; in addition, each data point is equipped with a timestamp, and the generation of the timestamp is completed by synchronizing the clock to the global unified time server, which can ensure that the recorded data has an accurate time stamp, providing a reliable time basis for subsequent data analysis.

[0066] S112: Based on the original data set, each data point is timestamped, and all outliers exceeding two standard deviations are removed to obtain a data set that has been purified through time synchronization;

[0067] The original data set will first be stored on a dedicated data server, and the integrity and accuracy of each data point will be checked according to preset rules; during the data cleaning process, data points that obviously do not meet physical limitations will be detected and eliminated, such as those that exceed the upper and lower limits of equipment operating parameters; at the same time, each data point in the data set is timestamped to ensure that the time stamps of all data points are consistent; in addition, through statistical analysis methods, such as calculating the mean and standard deviation of data points, abnormal data points that deviate from the normal range by more than two standard deviations can be automatically identified and eliminated, which can effectively identify abnormal points from large amounts of data, thereby providing accurate and reliable data support for data analysis and decision-making.

[0068] S113: Normalize the data set after time synchronization purification using the formula:

[0069] ;

[0070] Calculate the normalized value for each data point , get the synchronous monitoring data, where is the original value of the data point, represents all data points in the entire dataset, Represents the minimum value in the data set. Represents the maximum value in the data set;

[0071] Collect the centrifuge operating temperature data for one day, with the lowest temperature being The maximum temperature is At a certain moment, the measured temperature .

[0072] The calculation process is as follows:

[0073] Determine the value of the original data point, .

[0074] Find the maximum and minimum values ​​from a data set, and .

[0075] Substituting the values ​​into the formula:

[0076] ;

[0077] Normalized value It shows that within the observation range, the measured temperature is exactly in the middle of the lowest and highest temperatures. The normalized data value represents the position of the data point relative to the entire temperature range. It is dimensionless and can effectively reflect the relative position of the machine's operating status in the overall temperature range. It helps to quickly identify whether there are abnormal fluctuations or trends in temperature, which is of great significance for preventing centrifuge failures and optimizing operating parameters.

[0078] See also Figure 3 , the steps for recording the energy distribution are:

[0079] S211: receiving synchronous monitoring data of the centrifuge through the cloud computing platform, converting the data into frequency domain using Fourier transform, and generating preliminary vibration signal spectrum results;

[0080] The real-time monitoring data of the centrifuge is synchronously received through the cloud computing platform. Fourier transform is a mathematical tool that converts time domain signals into frequency domain signals. By converting time series data into frequency components, it can reveal the periodic characteristics and frequency distribution hidden in the data. The real-time monitoring data is transmitted to the data processing unit of the cloud platform through the network interface. The received data is processed by the Fourier transform module. Frequency domain analysis can not only identify the main vibration frequency, but also detect possible abnormal vibrations through spectrum analysis. The identification of frequency components is completed through digital signal processing algorithms, including the fast Fourier transform (FFT) algorithm. The FFT algorithm decomposes complex vibration signals into multiple simple sine waves. The frequency, amplitude and phase of the wave constitute the spectrum of the vibration signal, showing the relationship between frequency and vibration intensity.

[0081] S212: Based on the preliminary vibration signal spectrum result, digital filtering is performed to remove noise, key frequency components of the vibration signal are extracted, and peak detection is performed to obtain a signal peak result;

[0082] According to the preliminary vibration signal spectrum results, digital filtering technology is applied to remove unnecessary frequency components or noise and retain useful signal information. Predefined filter designs, such as low-pass, high-pass, band-pass and band-stop filters, are used to operate the signal according to the set frequency threshold. The extraction of the main frequency components of the vibration signal is achieved by designing a suitable band-pass filter. The design of the filter takes into account the frequency characteristics of the signal and the noise level that needs to be suppressed. Through the filtering process, only signals within a specific frequency range can pass, thereby accurately extracting the main vibration frequency representing the health status of the machine. Peak detection is to identify the maximum amplitude point in the filtered signal. The peak point represents the part with the most concentrated energy in the vibration signal. The accurate detection of the peak depends on the digital signal processing algorithm, which can ensure stability and accuracy under various operating conditions. Through further analysis of the peak, the dynamic response of the centrifuge under different working conditions can be understood in detail.

[0083] S213: Using the signal peak result, the formula is:

[0084] ;

[0085] Calculate the average energy distribution value , and obtain the energy distribution record, where Represents the total number of frequency points, Represents the absolute value of the amplitude at the i-th frequency point;

[0086] There are five frequency points in the data obtained from the centrifuge with amplitudes of 10, 20, 30, 40, and 50. First, take the absolute value of each amplitude and square it:

[0087] ;

[0088] Add the values ​​to get the sum:

[0089] ;

[0090] Find the square root of the sum:

[0091] ;

[0092] Divide this value by the total number of frequency points , the calculated average energy distribution result is:

[0093] ;

[0094] The results show that the average energy distribution of the centrifuge during this particular monitoring period was 14.832 units. A higher value indicates that the vibration of some parts of the centrifuge during operation is beyond the normal range, indicating a potential mechanical failure or the need for further maintenance inspection. By comparing the energy distribution data at different time periods or under different operating conditions, the health status of the centrifuge can be effectively monitored and predicted, thereby achieving preventive maintenance and reducing the risk of unexpected downtime.

[0095] See also Figure 4 , the steps to obtain the status analysis results are:

[0096] S221: Based on the energy distribution records, perform time series analysis, model the changing trend of energy data over time, analyze the autocorrelation and partial autocorrelation characteristics of the data, and generate energy trend analysis results;

[0097] The collected data is subjected to time series analysis, which identifies trends and periodicities from consecutive data points. The process begins by plotting the energy values ​​against time from the data set, using mathematical methods to evaluate the statistical dependencies between the data points. For example, by calculating the rate of change of consecutive points in the data, an increase or decrease in energy can be detected. Next, the presence and pattern of periodicity are revealed by measuring the correlation between points in the data that are separated by specific time intervals. In addition, this stage also includes evaluating seasonal fluctuations in the data, which is particularly important for predicting possible future behavior patterns of the equipment. A coherent analysis process lays the foundation for subsequent analysis by identifying long-term trends and short-term fluctuations in the data, thereby generating detailed analysis results of energy trends.

[0098] S222: Compare the energy trend analysis result with the preset performance baseline of the centrifuge, identify the peak value beyond the normal range, capture the abnormal operation and performance degradation of the centrifuge, and calibrate the data points that deviate from the baseline by more than two standard deviations by setting a threshold to obtain the abnormal energy point identification result;

[0099] The trend analysis results are further evaluated and compared with the preset performance baseline of the centrifuge. This comparison process involves specific numerical comparisons, where the preset baseline is usually determined by the standard operating parameters of the equipment, such as the average value of energy consumption. The baseline value is compared with the currently observed energy value, with special attention paid to those data points that are significantly higher or lower than expected. Data points that indicate abnormal equipment performance are marked by setting mathematical thresholds, such as the mean plus twice the standard deviation. This process not only relies on numerical calculations, but also includes visual inspection of the data, such as graphically displaying the relationship between data points and thresholds, so that it is possible to intuitively see which points deviate from the normal range, thereby obtaining the identification results of abnormal energy points.

[0100] S223: analyzing the abnormal energy point identification results, analyzing each abnormal point, analyzing the correlation with the centrifuge operating conditions and historical maintenance records, and generating a state analysis result;

[0101] Energy points marked as abnormal are analyzed in detail. This includes a review of the conditions behind each abnormal point, such as operating time, machine load, and maintenance history. Each abnormal point is recorded in detail, including the time it occurred and the corresponding machine operating status. Next, through comparative analysis, the pattern of abnormal points is found, such as whether they are concentrated under certain specific operating conditions. In addition, the relationship between these abnormal points and past maintenance activities will be explored to see if any untimely maintenance may have caused performance degradation. All of these analysis results will be summarized into a detailed report that not only points out the problem, but also provides improvement suggestions for each problem, thereby helping the maintenance team to develop a more effective maintenance strategy.

[0102] See also Figure 5 , the classification steps of abnormal data points are:

[0103] S311: According to the historical fault records of the centrifuge, the time, type and severity of all faults are counted, and a fault record database is established;

[0104] The historical fault records of the centrifuge are extracted from the maintenance database, which includes a variety of information such as the specific date of the fault, the type of fault, the faulty component and its severity. By sorting the data, a comprehensive fault record database can be built. The database construction process includes data cleaning (removing incomplete or erroneous records), data normalization (unifying data format) and data classification (classifying according to fault type), which can ensure the accuracy and efficiency of subsequent analysis. The database not only helps to quickly retrieve historical fault cases similar to the current abnormal data points, but also provides the necessary historical comparison for the classification of data points, and finally obtains a comprehensive database reflecting the similarity between current data points and historical faults.

[0105] S312: extracting abnormal data points from the state analysis results, cross-validating with the fault record database, and obtaining matching abnormal data sets by analyzing the matching degree between the data points and the historical fault patterns;

[0106] Using the established fault record database, the currently monitored abnormal data points are compared and analyzed in detail with the records therein. In this process, similarity analysis technology is used to perform a systematic matching test on the features of each abnormal data point. Specifically, the features of each abnormal point, such as vibration frequency, temperature deviation, etc., are extracted and then compared with the corresponding features of each record in the fault database. The similarity score is calculated to determine the possible fault types corresponding to these abnormal data points. This comparison and analysis process is completely based on the precise matching of data features, ensuring the high reliability and accuracy of the analysis, thereby effectively identifying abnormal data sets that match known fault patterns.

[0107] S313: performing cluster analysis on the matched abnormal data set, grouping the abnormal data according to the features, each group representing a potential fault cause, and generating abnormal data point classification results;

[0108] When performing cluster analysis on abnormal data points, the formula is used: , where each data point To cluster center The weighted distance is calculated by calculating its Euclidean distance There are three data points and two cluster centers The characteristics of the data points are , , The characteristics of the cluster centers are , .

[0109] Assumption Weight , , Then, calculate the data point To cluster center Distance:

[0110] ;

[0111] Substitute the formula into the calculation Contribution to cluster 1:

[0112] ;

[0113] Similarly, we can calculate To cluster center The distance and contribution of the data points, as well as the contribution of other data points to each cluster center. In this way, the cluster center is updated each iteration until the clustering result is stable. This process not only involves the distance calculation between the data point and the cluster center, but also takes into account the weight of the data point, so as to more accurately reflect the importance of the data point in the cluster, help understand the potential patterns and structures in the data set, and provide a basis for further failure mode analysis.

[0114] See also Figure 6 , the steps to obtain anomaly detection records are:

[0115] S321: According to the classification result of the abnormal data point, the classification result is compared with the known fault mode database, the fault type corresponding to each category is determined through feature matching, the fault mode represented by each category is identified, and a fault mode identification list is generated;

[0116] When identifying the centrifuge failure mode represented by each classification, the first step is data comparison analysis, which includes detailed matching of the clustered abnormal data points with a detailed database containing historical failure information. In specific operations, the feature set of each abnormal data point (for example, vibration frequency, temperature anomaly, pressure reading deviation) will be compared one by one with the fault features recorded in the database. The process manually selects highly relevant features for matching to enhance the accuracy of the match. For example, if an abnormal data point shows similar vibration frequency and temperature anomalies to a historical failure, this similarity will be quantified by calculating a similarity score. It ensures that each type of abnormal data can accurately correspond to a specific failure mode, providing a solid foundation for subsequent processing and analysis.

[0117] S322: Analyze each failure mode according to the failure mode identification list, sort them according to the occurrence frequency and the severity of the operation impact and loss caused, and generate a sorted failure mode list;

[0118] The process of analyzing and ranking failure modes involves a detailed assessment of the frequency and severity of failures. The process is completed through a comprehensive review of historical maintenance records and failure reports. The frequency of each failure mode is determined by counting the number of times the same failure has occurred in a certain period of time in the past, while the severity is scored based on the assessment report of the impact of the failure on production, usually with reference to a standardized severity rating system. Then, a simple weighting method is used to combine these two factors to calculate a comprehensive impact score for each failure mode. For example, the frequency of the failure may be given a higher weight, and if the failure causes a long production downtime, its severity is also correspondingly higher, which can ensure that those types of failures that are most likely to affect production stability are given priority.

[0119] S323: Based on the sorted fault mode list, record the actual information of each fault, including identification time, fault cause and maintenance measures, and comprehensively generate an abnormality detection record;

[0120] The resulting anomaly detection record includes comprehensive details of each failure mode. The generation of the record involves extracting and entering key information from the sorted failure mode list, automatically recording and updating the progress of the failure handling, and each record contains a number of key information such as the date of the failure, the predicted cause, the measures taken, and the prevention suggestions. In addition, in order to ensure the timeliness and accuracy of the failure handling measures, an automatic reminder function is also configured. When a failure record is created or updated, the relevant maintenance team will receive a notification. The systematic recording method not only improves the efficiency of failure response, but also provides data support for the continuous improvement of maintenance strategies and preventive measures.

[0121] See also Figure 7 ,The steps to obtain real-time remote monitoring alarm information are:

[0122] S411: Analyze the anomaly detection records, distinguish the different anomaly types, and quantify the impact range of each anomaly, determine the impact range of the anomaly type, and generate anomaly classification and impact analysis results;

[0123] Starting with the anomaly detection records, we first import the records into a dedicated data analysis platform. Each record contains detailed information about the centrifuge's operating parameters, such as temperature, pressure, speed, and vibration. Using data mining technology, we conduct an in-depth analysis of the records. The first step is to preprocess the data, including removing incomplete records and smoothing prominent noise data. Subsequently, a clustering algorithm is used to group the data to identify the pattern differences between normal operating data and abnormal data. Each abnormal event is labeled with a specific abnormality type and time of occurrence, which helps further time series analysis. By comparing the data changes before and after the abnormality occurs, the potential causes and scope of impact of the abnormality can be identified. Finally, all analysis results are summarized and classified. Each type of result details the abnormality type and predicted scope of impact, providing a basis for subsequent monitoring and maintenance decisions.

[0124] S412: Transmit the anomaly classification and impact analysis results to the cloud computing platform through a security protocol, perform data quality verification, confirm the integrity of the received data, and generate verified transmission data;

[0125] After the received anomaly classification and impact analysis data enters the cloud computing platform, the serial number is first registered to ensure that each batch of data can be traced back to a specific device and time point. Then, the data is parsed using a scripting language. The script can read the data and convert it into a format that the analysis software can recognize. Data cleaning will be performed during the conversion process, including removing null values, outliers, and duplicate records to ensure the data quality of subsequent analysis. After data cleaning, the preset analysis model is used to conduct in-depth analysis of the data. This model is based on historical data and can identify specific types of abnormal patterns, providing a scientific basis for formulating rapid response measures.

[0126] S413: Based on the verified transmission data, according to the abnormality type and impact scope, record key real-time data and response strategies, automatically send them to the monitoring center, complete real-time alarm push, and output real-time remote monitoring alarm information;

[0127] After the data processing is completed, the cloud platform instantly generates real-time alarms and is configured with a variety of response mechanisms. According to the severity and type of the anomaly, the appropriate alarm level will be selected and the corresponding alarm notification process will be activated. The alarm information will include the specific anomaly type, the predicted impact area, and the recommended response strategy. The information is securely transmitted to the main console of the monitoring center through the internal network. The operators of the monitoring center will immediately receive these alarms and evaluate the situation based on the information provided to determine whether manual intervention is required or to adjust the settings of the automatic response system, thereby ensuring a quick and effective response to abnormal events.

[0128] A remote monitoring and analysis system for centrifuge data based on cloud computers, comprising:

[0129] The data acquisition module installs sensors on the key components of the centrifuge to collect operating data, timestamps each data point, removes outliers, performs normalization, calculates the normalized value of each data point, and obtains synchronous monitoring data;

[0130] The frequency domain analysis module receives the synchronous monitoring data of the centrifuge through the cloud computing platform, uses Fourier transform to convert the data into frequency domain, extracts the key frequency components of the vibration signal and performs peak detection, calculates the average energy distribution value, and obtains the energy distribution record;

[0131] The trend modeling module performs time series analysis based on energy distribution records, models the trend of energy data over time, compares it with the preset performance baseline of the centrifuge, analyzes the correlation with the centrifuge operating conditions and historical maintenance records, and generates status analysis results;

[0132] The fault detection module establishes a fault record database based on the historical fault records of the centrifuge, extracts abnormal data points from the state analysis results, cross-validates with the fault record database, groups the abnormal data according to the features, and generates abnormal data point classification results;

[0133] The pattern comparison module compares the classification results of abnormal data points with the known fault pattern database, analyzes each fault pattern, sorts them according to the frequency of occurrence and the severity of the operational impact and loss caused, records the actual information of each fault, and comprehensively generates anomaly detection records;

[0134] The real-time alarm module analyzes the anomaly detection records, determines the impact range of each anomaly type, transmits them to the cloud computing platform through a security protocol, performs data quality verification, records key real-time data and response strategies, automatically sends them to the monitoring center, and outputs real-time remote monitoring alarm information.

[0135] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A remote monitoring and analysis method for centrifuge data based on cloud computing, characterized in that: The following steps are involved: S1: Collect centrifuge operation index data, synchronize time tags, identify and remove outliers in the data, integrate data through data normalization, and generate synchronous monitoring data; S2: Perform frequency domain analysis on the synchronous monitoring data through a cloud computer, identify the vibration signal spectrum of the centrifuge, analyze the signal peak and record the energy distribution, perform trend analysis on the data in combination with the energy distribution recording results, identify data points that exceed the normal range and mark them as abnormal data points by comparing with the preset performance baseline of the centrifuge, reveal potential operating abnormalities of the centrifuge, and generate status analysis results; S3: Based on the state analysis result, cross-validate the abnormal data points with the historical fault records of the centrifuge, classify the abnormal data points using cluster analysis, identify the centrifuge fault mode according to the abnormal point classification results, sort each fault mode according to the frequency and severity, and generate anomaly detection records; S4: using the abnormality detection record, recording the abnormality type and impact range of the centrifuge, sending an alarm to the monitoring center through the cloud computing platform, and generating real-time remote monitoring alarm information; The energy distribution is recorded as follows: S211: receiving the synchronous monitoring data of the centrifuge through the cloud computing platform, converting the data into frequency domain using Fourier transform, and generating preliminary vibration signal spectrum results; S212: Based on the preliminary vibration signal spectrum result, digital filtering is performed to remove noise, key frequency components of the vibration signal are extracted, and peak detection is performed to obtain a signal peak result; S213: Using the signal peak value result, adopt the formula: ； Calculate the average energy distribution value E and obtain the energy distribution record, where n represents the total number of frequency points. Represents the absolute value of the amplitude at the i-th frequency point.

2. The method for remote monitoring and analyzing centrifuge data based on cloud computing according to claim 1, characterized in that: The steps for acquiring the synchronous monitoring data are as follows: S111: Install sensors on key components of the centrifuge to collect operating data, record speed, temperature and vibration frequency, and form a raw data set; S112: Based on the original data set, each data point is timestamped, and all outliers exceeding two standard deviations are removed to obtain a data set that has been purified through time synchronization; S113: normalizing the data set that has undergone time synchronization purification, using the formula: ; Calculate the normalized value for each data point , get the synchronous monitoring data, where is the original value of the data point, represents all data points in the entire dataset, Represents the minimum value in the data set. Represents the maximum value in the data set.

3. The method for remote monitoring and analyzing centrifuge data based on cloud computing according to claim 1, characterized in that: The steps for obtaining the status analysis result are: S221: Based on the energy distribution record, perform time series analysis, model the change trend of energy data over time, analyze the autocorrelation and partial autocorrelation characteristics of the data, and generate energy trend analysis results; S222: Compare the energy trend analysis result with the preset performance baseline of the centrifuge, identify the peak value beyond the normal range, capture the abnormal operation and performance degradation of the centrifuge, and calibrate the data points that deviate from the baseline by more than two standard deviations by setting a threshold to obtain the abnormal energy point identification result; S223: Analyze the abnormal energy point identification result, analyze each abnormal point, analyze the correlation with the centrifuge operating conditions and historical maintenance records, and generate a state analysis result.

4. The method for remote monitoring and analyzing centrifuge data based on cloud computing according to claim 3 is characterized in that: The classification steps of the abnormal data points are: S311: According to the historical fault records of the centrifuge, the time, type and severity of all faults are counted, and a fault record database is established; S312: extracting abnormal data points from the state analysis results, cross-validating with the fault record database, and obtaining matching abnormal data sets by analyzing the matching degree between the data points and the historical fault patterns; S313: Perform cluster analysis on the matched abnormal data set, group the abnormal data according to the features, each group represents a potential fault cause, and generate abnormal data point classification results.

5. The method for remote monitoring and analyzing centrifuge data based on cloud computing according to claim 4 is characterized in that: The steps for obtaining the anomaly detection record are: S321: According to the classification result of the abnormal data point, the classification result is compared with the known fault mode database, the fault type corresponding to each category is determined by feature matching, the fault mode represented by each classification is identified, and a fault mode identification list is generated; S322: Analyze each failure mode according to the failure mode identification list, sort them according to the occurrence frequency and the severity of the operational impact and loss caused, and generate a sorted failure mode list; S323: Based on the sorted fault mode list, actual information of each fault is recorded, including identification time, fault cause and maintenance measures, and an abnormality detection record is comprehensively generated.

6. The method for remote monitoring and analyzing centrifuge data based on cloud computing according to claim 5, characterized in that: The steps for obtaining the real-time remote monitoring alarm information are as follows: S411: Analyze the anomaly detection record, distinguish different anomaly types, and quantify the impact range of each anomaly, determine the impact range of the anomaly type, and generate anomaly classification and impact analysis results; S412: Transmitting the anomaly classification and impact analysis results to the cloud computing platform through a security protocol, performing data quality verification, confirming the integrity of the received data, and generating verified transmission data; S413: Based on the verified transmission data, according to the abnormality type and impact scope, record key real-time data and response strategies, automatically send them to the monitoring center, complete real-time alarm push, and output real-time remote monitoring alarm information.

7. A remote monitoring and analysis system for centrifuge data based on cloud computing, characterized in that: The system is used to execute the cloud computer-based centrifuge data remote monitoring and analysis method according to any one of claims 1 to 6, comprising: The data acquisition module installs sensors on the key components of the centrifuge to collect operating data, timestamps each data point, removes outliers, performs normalization, calculates the normalized value of each data point, and obtains synchronous monitoring data; The frequency domain analysis module receives the synchronous monitoring data of the centrifuge through the cloud computing platform, uses Fourier transform to perform frequency domain conversion on the data, extracts the key frequency components of the vibration signal and performs peak detection, calculates the average energy distribution value, and obtains the energy distribution record; The trend modeling module performs time series analysis based on the energy distribution record, models the trend of energy data changing over time, compares it with the preset performance baseline of the centrifuge, analyzes the correlation with the centrifuge operating conditions and historical maintenance records, and generates a state analysis result; The fault detection module establishes a fault record database based on the historical fault records of the centrifuge, extracts abnormal data points from the state analysis results, cross-validates with the fault record database, groups the abnormal data according to the features, and generates abnormal data point classification results; The pattern comparison module compares the classification results of the abnormal data points with the known fault pattern database, analyzes each fault pattern, sorts them according to the frequency of occurrence and the severity of the operational impact and loss caused, records the actual information of each fault, and comprehensively generates anomaly detection records; The real-time alarm module analyzes the anomaly detection records, determines the impact scope of each anomaly type, transmits them to the cloud computing platform through a security protocol, performs data quality verification, records key real-time data and response strategies, automatically sends them to the monitoring center, and outputs real-time remote monitoring alarm information.

Images