Sintering machine air leakage detection and alarm system based on sensing technology

Abstract

The invention relates to the technical field of sintering detection, in particular to a sintering machine air leakage detection and alarm system based on a sensing technology, which comprises sensor groups which are respectively arranged at different parts of a sintering machine and are used for collecting data including oxygen content, pressure and temperature of the sintering machine and transmitting the data to a data processing module and a data storage module; the data processing module is used for carrying out reliability evaluation, interactive verification and correction fusion processing on the data acquired by the sensor group so as to judge the air leakage condition of the sintering machine; after the data is processed by the data processing module, the alarm module is used for sending out an alarm signal when it is judged that the air leakage condition exceeds a preset threshold value; the data storage module is used for storing the original data acquired by the sensor group, the intermediate data in the data processing process of the data processing module, the final fusion data and the alarm record of the alarm module; the problem that the air leakage rate is difficult to directly and accurately calculate due to the fact that changes of temperature and dynamic pressure are possibly interfered by various factors is solved.

Description

Technical Field

[0001] This invention relates to the field of sintering detection technology, and in particular to a sintering machine air leakage detection and alarm system based on sensor technology. Background Technology

[0002] In steel production and related industries, sintering machines play a vital role. Their operating efficiency and stability directly affect the benefits of the entire production process. However, air leakage in sintering machines has always been one of the key factors restricting their efficient operation.

[0003] Traditional methods for detecting air leakage in sintering machines mainly include manual inspection and partially automated methods. Manual inspection relies on operators checking various parts of the sintering machine at specific time intervals, assessing air leakage through observation, experience, and some simple tools. This method has many drawbacks: manual inspection is extremely inefficient, and due to the large and complex structure of the sintering machine, it requires a lot of manpower and time, and it cannot achieve real-time continuous monitoring.

[0004] With technological advancements, several automated detection methods have emerged. Among them, temperature and dynamic pressure data-based detection technologies are relatively mature and have seen some application. These methods acquire relevant data by installing temperature and dynamic pressure sensors on the sintering machine's air box. However, relying solely on this data makes it difficult to directly and accurately calculate the leakage rate. This is because changes in temperature and dynamic pressure can be influenced by various factors, not just leakage. For example, changes in the materials inside the sintering machine and fluctuations in combustion conditions can alter temperature and dynamic pressure, leading to significant errors in leakage assessments based on this data.

[0005] Chinese patent CN118258562A discloses a method for detecting the air leakage rate of a sintering machine, comprising the following steps: temporarily stopping the sintering machine and cleaning the grate bars at the machine head; placing a high-temperature resistant detection tube and then turning on the sintering machine, with the detection tube rotating forward with the trolley; moving the high-temperature resistant detection tube to the igniter of the sintering machine, connecting the detection pipeline, and starting the detection program in advance; when the trolley with the high-temperature resistant detection tube installed moves to the igniter, inserting a rubber hose into the thin tube of the high-temperature resistant detection tube; sealing the gap between the high-temperature resistant detection tube and the trolley's pressure pin hole with clay; after the detector enters normal detection mode, it moves with the trolley, measuring the oxygen content as it travels from the head to the tail of the sintering machine through each air box, and recording the oxygen content detection data; using a portable high-precision oxygen detector to detect the oxygen content in various parts of the sintering machine, and calculating the air leakage rate of the sintering machine based on the formula for calculating the air leakage rate using the changes in oxygen in the flue gas of the sintering machine; however, this patent does not analyze the correlation between sensors or the reliability of the data.

[0006] Chinese Patent CN109490001B discloses a method for detecting the air leakage rate of an iron ore sintering machine, comprising the following steps: first, fabricating a detection device and placing a row of detection devices on the sintering material layer; during sintering, recording the wind speed measured by each detection device, dividing the sintering material layer into a grid, and using the wind speeds recorded by each detection device from beginning to end as the wind speeds at each grid point on the longitudinal boundary line where the detection device is located, calculating the effective air volume QY passing through the material layer per unit time; calculating the total amount of flue gas QZ in the main flue per unit time; calculating the amount of water vapor generated QE per unit time; converting QY, QZ, and QE into quantities under standard conditions, and calculating the air leakage rate QL of the sintering system per unit time, and calculating the air leakage rate K of the sintering system per unit time. This invention improves the accuracy of sintering air leakage rate calculation by optimizing the measurement and calculation method of effective air volume in sintering. It is highly operable, quick and convenient, and does not affect sintering production, thus providing a theoretical basis for reducing energy consumption and increasing output in sintering production. However, this patent does not analyze the correlation between sensors or the reliability of data. Summary of the Invention

[0007] This invention provides a sintering machine air leakage detection and alarm system based on sensor technology, which solves the problem that it is difficult to directly and accurately calculate the air leakage rate because changes in temperature and dynamic pressure may be affected by various factors.

[0008] To achieve the above objectives, the present invention employs the following technical solution:

[0009] A sintering machine air leakage detection and alarm system based on sensor technology includes:

[0010] The sensor array is installed in different parts of the sintering machine to collect data on the sintering machine, including oxygen content, pressure and temperature, and transmit it to the data processing module and the data storage module.

[0011] The data processing module performs reliability assessment, interactive verification, and correction fusion processing on the data collected by the sensor group to determine the air leakage of the sintering machine;

[0012] The alarm module, after the data is processed by the data processing module, is used to issue an alarm signal when the air leakage exceeds a preset threshold.

[0013] The data storage module is used to store the raw data collected by the sensor group, the intermediate data during the data processing process of the data processing module, the final fused data, and the alarm records of the alarm module.

[0014] Furthermore, the sensor group includes an oxygen content sensor, a pressure sensor, and a temperature sensor;

[0015] The oxygen content sensor is installed in the flue of the sintering machine at a location that can stably detect changes in oxygen concentration, and is used to detect oxygen content data at different locations inside the sintering machine.

[0016] The pressure sensor is installed at the inlet and outlet of the sintering machine air box, and the measuring surface of the compression sensor is perpendicular to the airflow direction at the inlet and outlet, and is used to detect pressure data.

[0017] The temperature sensor is installed around the sealing area between the sintering machine trolley and the slide rail to detect temperature data.

[0018] Furthermore, the data processing module includes:

[0019] The correlation matrix unit constructs a sensor data correlation matrix, calculates the covariance of correlated sensor pairs to determine data consistency, and calculates information entropy to quantify data uncertainty.

[0020] The reliability assessment unit performs reliability assessments on the data collected by the sensor group based on Bayesian estimation.

[0021] The correction unit is constructed based on the likelihood function according to the sensor data distribution model and takes into account the spatial correlation between sensor data for correction.

[0022] Furthermore, the correlation matrix unit also includes a suspicious node marking subunit. If the covariance of the sensor group exceeds a preset threshold, and the posterior probability based on Bayesian estimation is less than the preset threshold, and its information entropy is greater than the preset information entropy difference threshold of adjacent nodes, then the sensor is marked as a suspicious node.

[0023] Furthermore, the correction unit also includes a weight determination subunit, which determines the weight of sensors that have been verified by the reliability assessment unit as non-suspicious and whose data is reliable, based on their posterior probability and data consistency index estimated by Bayes.

[0024] Furthermore, the alarm module includes:

[0025] The alarm signal triggering unit continuously receives analysis results data on the air leakage of the sintering machine from the data processing module. When the data shows that the air leakage rate of a certain area exceeds the preset alarm threshold, the alarm signal triggering mechanism is activated.

[0026] The alarm signal output unit emits audible and visual alarm signals according to the instructions of the alarm signal triggering unit.

[0027] Furthermore, the data storage module includes:

[0028] The raw data storage unit uses the sensor number and data acquisition time as indexes to store the raw data collected by the sensor completely in a dedicated raw database table;

[0029] The intermediate data storage unit associates the intermediate data generated during data processing, including data reliability assessment results and covariance calculation values, with the corresponding original data and stores them in the intermediate database table.

[0030] The recording unit has a separate database table for storing fused data and alarm records.

[0031] Furthermore, the data processing module also includes a machine learning algorithm module, which uses a neural network algorithm to take the data collected by the multi-sensor group as input and the air leakage type as output, and optimizes the weights and thresholds of the neural network through training.

[0032] Furthermore, it also includes a sensor self-calibration function module, which automatically adjusts the sensor's calibration parameters based on data reliability assessment results and changes in environmental parameters, ensuring that the sensor is always in optimal working condition.

[0033] Furthermore, it also includes a data compression algorithm module to reduce data transmission volume and transmission time.

[0034] Compared with the prior art, the beneficial effects of the present invention are:

[0035] 1) This invention achieves air leakage detection through the rational layout of multiple types of sensors in key parts of the sintering machine. Combined with data reliability assessment based on Bayesian estimation, multi-sensor data interaction verification mechanism, and data correction and fusion technology, it effectively eliminates the influence of environmental interference and sensor error factors, enabling more accurate judgment of the location and extent of air leakage, timely issuance of alarm signals, and prompting relevant personnel to take measures to repair the leak. This increases the amount of air pumped through the sintering layer, increases the constant volume of sintered ore, saves a lot of electricity, improves productivity, eliminates noise, makes the working environment more environmentally friendly, increases stability, and improves overall production efficiency.

[0036] 2) In the data processing process, this invention performs rigorous data screening and reliability assessment from the data acquisition source. During the data acquisition stage, the data collected by the sensors is initially verified to exclude obviously abnormal data. The data reliability assessment based on Bayesian estimation comprehensively considers the historical performance of the sensors and the spatial correlation of the data to accurately determine the reliability of the data. The multi-sensor data interaction verification mechanism further verifies the data comprehensively through covariance analysis and information entropy calculation, which effectively improves the data quality, ensures that the data participating in the air leakage analysis is true and reliable, reduces the risk of misjudgment caused by erroneous data, enhances the reliability of the entire detection and alarm system, and provides a solid data foundation for the long-term stable operation of the sintering machine.

[0037] 3) The data storage and analysis function provides strong support for the optimization of sintering machines. By storing data in a multi-dimensional classification and using data mining techniques such as cluster analysis and regression analysis, the relationship between air leakage data and sintering machine operating parameters can be studied in depth. This can accurately identify problems in the structural design, process operation and sealing performance of the sintering machine, provide a scientific basis for targeted optimization and improvement, improve resource utilization, reduce production costs and promote the sustainable development of sintering production. Attached Figure Description

[0038] Figure 1 This is an architecture diagram of the air leakage detection and alarm system in this invention.

[0039] Figure 2 This is a schematic diagram of the data processing module described in this invention.

[0040] Figure 3 This is a schematic diagram of the alarm module described in this invention.

[0041] Figure 4 This is a schematic diagram of the data storage module described in this invention.

[0042] Figure 5 This is a schematic diagram of the arrangement of the sensor described in this invention. Detailed Implementation

[0043] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings:

[0044] See Figure 1-5 The present invention provides a sintering machine air leakage detection and alarm system based on sensor technology, comprising:

[0045] The sensor array is installed in different parts of the sintering machine to collect data on the sintering machine, including oxygen content, pressure and temperature, and transmit it to the data processing module and the data storage module.

[0046] The data processing module performs reliability assessment, interactive verification, and correction fusion processing on the data collected by the sensor group to determine the air leakage of the sintering machine;

[0047] The alarm module, after the data is processed by the data processing module, is used to issue an alarm signal when the air leakage exceeds a preset threshold.

[0048] The data storage module is used to store the raw data collected by the sensor group, the intermediate data during the data processing process of the data processing module, the final fused data, and the alarm records of the alarm module.

[0049] The specific steps for detecting and preventing air leakage are as follows:

[0050] 1. Data collection;

[0051] Based on the complex structure and internal physicochemical processes of the sintering machine, various types of pressure sensors are installed at the inlet and outlet of the air box. The measuring surface of the pressure sensors must be perpendicular to the airflow direction to capture subtle changes in airflow pressure. These pressure sensors employ a piezoresistive sensing principle; the core piezoresistive resistor changes resistance under pressure, and this change is converted into a voltage signal via a Wheatstone bridge circuit, thus enabling pressure data acquisition. The pressure sensors installed at the inlet and outlet of the sintering machine's air box are primarily used for real-time monitoring of airflow pressure changes. Since the air box should maintain a stable negative pressure state during normal operation, any air leakage will cause rapid fluctuations in local negative pressure. By arranging the measuring surface of the pressure sensors perpendicular to the airflow direction, this system can efficiently capture these abnormal fluctuations and quickly determine if air leakage exists. Traditional systems rely on changes in fan current to indirectly infer air leakage, which suffers from delays or insufficient accuracy. This system, however, directly monitors negative pressure changes, reacting rapidly and providing immediate warnings, thus improving the immediacy and accuracy of air leakage identification.

[0052] Inside the flue, the oxygen content sensor operates based on electrochemical principles. Its probe consists of a permeable membrane that allows only oxygen molecules to pass through and electrodes. Oxygen reacts chemically with the electrodes to generate an electrical signal, the intensity of which is proportional to the oxygen content. The sensor is installed in an area within the flue where airflow is stable and represents the overall trend of oxygen content changes. The oxygen content sensor provides a basis for judging air leakage from a chemical composition perspective; it compensates for the "detection blind spot" that pressure sensors may have when negative pressure changes slowly or when there is a small-scale leak, enhancing the system's ability to identify minor or chronic air leaks. Furthermore, it can further confirm the impact of air leakage on the process atmosphere (such as reducing atmosphere or combustion conditions), providing data support for system optimization. The oxygen content sensor is installed in a location within the sintering machine flue where airflow is relatively stable and can represent the overall oxygen concentration changes of the system. Its main function is to monitor changes in the oxygen content in the air. When air leakage occurs, external air enters the sintering system, and the oxygen concentration usually increases, especially in high-temperature operating environments, where this change is particularly sensitive around the sealing areas of the trolley and slide.

[0053] The temperature sensor employs a thermocouple, utilizing the Seebeck effect—the thermoelectric potential generated when two metal wires of different materials are connected at different temperatures. The temperature is calculated by measuring this thermoelectric potential and referring to a thermocouple calibration table. The temperature sensor is installed close to the sealing area to sensitively detect temperature fluctuations caused by air leakage. The temperature sensor is primarily deployed around the sealing areas of the sintering machine trolley and slide rails. These locations are often high-risk areas for air leakage and are also the most difficult to place sensors in traditional technologies. The entry of cold air caused by leakage can lead to significant local temperature drops or fluctuations. The temperature sensor can detect these abnormal temperature changes to determine if the seal has failed, thus further confirming the air leakage. Traditional systems rarely consider the role of temperature changes in leak detection, while this system utilizes temperature sensors to improve the ability to detect areas that are difficult to monitor directly. Especially in scenarios where negative pressure and oxygen concentration changes are not significant, temperature anomalies can become a key criterion, effectively improving the system's spatial coverage and leak location accuracy.

[0054] The aforementioned sensors collect data once every 0.5 seconds. The data is transmitted to the data acquisition unit through a combination of wired and wireless communication links. In the wireless communication link, a 5G module is selected and equipped with a high-gain antenna to enhance signal strength, ensure the stability and efficiency of data transmission, and effectively prevent data loss and transmission delay.

[0055] 2. Data reliability assessment based on Bayesian estimation;

[0056] The collected data were evaluated for reliability based on Bayesian estimation. Under the stable operating conditions of the sintering machine with no obvious air leakage, data for a period of 10 days was continuously collected as a sample to construct a data distribution model.

[0057] Taking a pressure sensor as an example, let's assume that the sample data it collects is... The mean can be obtained through statistical analysis:

[0058]

[0059] variance:

[0060]

[0061] Thus, a normal distribution model is established. For prior probability The determination of this, in one possible implementation method, comprehensively considers multiple factors, if a certain pressure sensor in the past In this calibration, the average error is And it has only appeared once. All faults were repaired promptly, with an average repair time of [time missing]. Hourly, based on these historical calibration data, maintenance records, and operational stability data under similar operating conditions, prior probabilities are determined using specific empirical formulas or expert evaluation systems. During real-time monitoring, for new data points, the likelihood function value is calculated based on the established normal distribution model.

[0062]

[0063] When considering the spatial correlation between sensor data, Markov random field (MRF) theory is introduced. Specifically, for pressure sensors in adjacent bellows... and Let the distance between them be... The correlation coefficient of the pressure data obtained from historical data is: Construct the potential function:

[0064]

[0065] in, and The association weights are calculated based on the empirically determined weight coefficients and the potential function constructed above.

[0066]

[0067] By incorporating the association weights into the likelihood function calculation, we obtain the corrected likelihood function value:

[0068]

[0069] in, For the set of sensors that are adjacent to and associated with the sensor, combined with prior probabilities According to Bayes' theorem:

[0070]

[0071] Calculate the posterior probability If the posterior probability is greater than the preset threshold, the sensor data is determined to be reliable; otherwise, it is marked as data to be verified and enters the multi-sensor data interaction verification mechanism.

[0072] 3. Multi-sensor data interaction verification mechanism;

[0073] Construct a data association matrix and perform multi-sensor data interaction verification;

[0074] In one possible implementation, based on the structural design and physical process principles of the sintering machine, the potential relationships between various sensors are determined in detail, and a data correlation matrix is ​​constructed. If there is a physical connection between the pressure sensors of adjacent air boxes, and between the temperature sensors and oxygen content sensors in the same area, then the corresponding positions in the correlation matrix will be... Mark 1 if there is a correlation, otherwise mark 0. For correlated sensor pairs, calculate the covariance of their data:

[0075]

[0076] in, For the number of data samples, and Sensors and The mean of the data is calculated, and the information entropy of the sensor data is also calculated. Generally, for sensors... dataset First, count the frequency of each data point. Calculate its probability in the dataset:

[0077]

[0078] Then, based on the information entropy formula:

[0079]

[0080] To calculate information entropy, during the depth-first search verification process, starting from the core pressure sensor of the main air box, a stack structure is used to store the nodes to be visited. The sensor nodes are traversed according to the first-in-last-out principle. If the covariance of a certain sensor exceeds the threshold (e.g., if the covariance threshold is set to 10, the actual calculated value is 12), and the posterior probability based on Bayesian estimation is less than the threshold (e.g., if the threshold is 0.6, the actual posterior probability is 0.5), and the information entropy is larger than that of the adjacent nodes by a certain value (e.g., if the information entropy difference threshold is set to 0.4, the actual difference is 0.5), then the sensor is marked as a suspicious node.

[0081] 4. Data correction and fusion;

[0082] The data undergoes correction and fusion processing. For sensors that are not questionable and whose data is reliable, in some embodiments, the data is fused based on their posterior probability estimated by Bayes. and data consistency metrics Determine weights Specifically, data consistency metrics It can be obtained through covariance normalization. For example, if the covariance is... ,but The weight calculation formula is: Combining multi-sensor data into a data matrix ,in For the number of sensors, Calculate the covariance matrix for the number of data samples. The eigenvalue decomposition algorithm is used for eigenvalue decomposition; the eigenvalue decomposition can be expressed as:

[0083]

[0084] in, The eigenvector matrix, Given an eigenvalue diagonal matrix, select the first eigenvalues ​​based on their magnitude. If we select the first three feature values, the cumulative contribution rate is:

[0085]

[0086] The corresponding eigenvectors form the transformation matrix. The original data is dimensionality reduced to obtain a dimensionality-reduced data matrix. The weighted least squares method is used to fuse the dimensionality-reduced data. In one possible implementation, the dimensionality-reduced data vector collected by the sensors is fused. Weight vector Substituting into the weighted least squares formula:

[0087]

[0088] This is used to fuse vectors The calculation is performed, and finally, the PCA transformation matrix is ​​used. The merged data The data is then transformed back to the original data space to obtain the final fused data. This is used for subsequent air leakage analysis and judgment in the sintering machine. If the fused data shows that the air leakage rate in a certain area exceeds the set alarm threshold, the alarm unit will issue an audible and visual alarm signal and send alarm information to the remote monitoring center through the network communication module so that relevant personnel can take timely measures.

[0089] 5. Data storage and analysis;

[0090] This embodiment focuses on data storage and analysis. The system is equipped with a dedicated data storage unit, employing a large-capacity solid-state drive array for data storage. Data storage is categorized and stored according to multiple dimensions, including data type, acquisition time, and sensor location. Raw data is indexed by sensor number and acquisition time and stored in the raw database table for easy subsequent querying and tracing. Intermediate data, such as data reliability assessment results and covariance calculation values, are associated with the corresponding raw data and stored in the intermediate database table. Fusion data and alarm records are stored in separate database tables for quick retrieval of alarm information and corresponding fusion data details. Professional data mining and analysis software is used to conduct in-depth analysis of the stored data. For example, clustering algorithms are used to cluster leakage data under different operating conditions to analyze the characteristic differences of various leakage situations. Alternatively, regression analysis methods are employed to study the quantitative relationship between leakage rate and sintering machine operating parameters, including air volume, air pressure, and material layer thickness. This provides scientific data for sintering machine structural optimization, process adjustment, and sealing performance improvement, further enhancing the overall operating efficiency and stability of the sintering machine and reducing energy consumption and production costs.

[0091] As an extension of the technology, machine learning algorithms can be introduced into the system. By training the system with a large amount of historical data, it can automatically learn the sensor data characteristics under different leakage modes, thereby achieving intelligent identification and classification of leakage types. A neural network algorithm can be selected, using multi-sensor data as input and leakage type as output. Through training and optimization of the neural network's weights and thresholds, the system's ability to identify complex leakage situations can be improved. Simultaneously, the sensor's self-calibration function can be further optimized. During system operation, based on data reliability assessment results and changes in environmental parameters, the sensor's calibration parameters can be automatically adjusted to ensure the sensor is always in optimal working condition, improving the accuracy and stability of data acquisition. Furthermore, in terms of data transmission, more efficient data compression algorithms can be researched and developed to reduce data transmission volume and time, improve the system's real-time response capability, and meet the needs of large-scale data transmission and processing.

[0092] Working principle: This invention provides a sintering machine air leakage detection and alarm system based on sensor technology, as detailed below:

[0093] Various types of sensors are deployed in a targeted manner according to the physicochemical characteristics of different parts of the sintering machine. At the inlet and outlet of the air box, the pressure sensor uses the piezoresistive principle. When the airflow pressure acts on the pressure-sensitive resistor, it causes the resistance to change. The Wheatstone bridge converts this resistance change into a voltage signal. This voltage value corresponds to the magnitude of the airflow pressure. It is installed perpendicular to the airflow direction to ensure accurate sensing of pressure changes.

[0094] The oxygen content sensor in the flue is based on the principle of electrochemistry. The breathable membrane selectively allows oxygen to pass through, and the oxygen reacts with the electrodes to generate an electrical signal that is proportional to the oxygen content. When installed in a stable and representative area, it can accurately reflect the overall oxygen content.

[0095] Thermocouple temperature sensors around the sealed area between the trolley and the slide rely on the Seebeck effect. The temperature difference between the two ends of metal wires of different materials generates a thermoelectric potential. By measuring the thermoelectric potential and calculating the temperature value according to the calibration table, the sensors can quickly detect temperature fluctuations caused by air leakage near the sealed area. These sensors collect data at a frequency of once every 0.5 seconds and transmit it to the data acquisition unit via a communication link combining wired and 5G wireless (equipped with a high-gain antenna to enhance the signal), ensuring stable and efficient data transmission.

[0096] Under normal, leak-free operating conditions, 10 days of data are collected to construct a data distribution model for each sensor. Taking the pressure sensor as an example, a normal distribution model is established by obtaining the mean and variance from the statistical sample data. The prior probability is determined by combining historical calibration errors, failure frequency and repair time, and stability data under similar operating conditions, and is determined by specific empirical formulas or expert evaluation. During real-time detection, the likelihood function value of new data points is calculated based on the normal distribution model. Considering the spatial correlation of data, the MRF theory is introduced. The potential function is constructed based on the distance between adjacent pressure sensors and the correlation coefficient of historical data to calculate the correlation weight. This weight is then incorporated into the likelihood function to obtain the corrected likelihood function value. The prior probability is combined with Bayes' theorem to calculate the posterior probability, which is used to judge the reliability of the data. Data exceeding a preset threshold is considered reliable; otherwise, an interactive verification mechanism is initiated.

[0097] A data association matrix is ​​constructed based on the structure and physical process of the sintering machine, and sensor pairs with physical associations are marked. The covariance of the associated sensor pairs is calculated by averaging the products of the differences between the data points and their respective means, which reflects the consistency of the changing trends between the data. At the same time, the information entropy is calculated by first calculating the probability of the data points based on their frequency, and then calculating it according to the information entropy formula to quantify the uncertainty of the data. In the depth-first search, starting from the core pressure sensor of the main air box, the nodes are traversed according to the first-in-last-out principle. When the covariance of a certain sensor exceeds the threshold, the posterior probability is lower than the threshold, and the information entropy is significantly different from that of the adjacent nodes, it is marked as a suspicious node. The reliability of the data is judged by comprehensively considering multiple factors.

[0098] For non-doubtful reliable sensors, weights are determined based on the consistency index after Bayesian posterior probability and covariance normalization. After multi-sensor data are combined into a matrix, the covariance matrix is ​​calculated and eigenvalue decomposition is performed. Eigenvalue decomposition yields an eigenvector matrix and an eigenvalue diagonal matrix. The eigenvectors corresponding to the top few eigenvalues ​​with the required cumulative contribution rate are selected to form a transformation matrix for data dimensionality reduction. The dimensionality-reduced data vector and weight vector are substituted into the weighted least squares formula to calculate the fusion vector. Finally, the PCA transformation matrix is ​​used to inversely transform back to the original data space to obtain the final fused data for air leakage analysis. If the air leakage rate exceeds the threshold, the alarm unit activates, triggering an audible and visual alarm and transmitting information to the remote monitoring center.

[0099] The data storage unit utilizes a large-capacity solid-state drive array for multi-dimensional classification and storage. Raw data is indexed by number and time in a raw table for easy traceability; intermediate data is linked to raw data in an intermediate table; and fused data and alarm records are stored in separate tables for quick alarm detail retrieval. Clustering analysis algorithms are used to cluster leakage data under different operating conditions to identify feature differences. Regression analysis studies the quantitative relationship between leakage rate and sintering machine operating parameters, providing a basis for equipment optimization and improving overall performance. In the extended technologies, machine learning algorithms (such as neural networks) are trained with a large amount of historical data, using multi-sensor data as input and leakage type as output. Optimized weight thresholds enable intelligent identification of leakage types. Sensor self-calibration functions automatically adjust parameters based on data reliability and environmental changes to maintain accuracy. Efficient data compression algorithms reduce transmission volume and time, enhancing the system's real-time response capabilities to meet the needs of large-scale data processing.

[0100] The above embodiments are implemented based on the technical solution of the present invention, providing detailed implementation methods and specific operation processes. However, the scope of protection of the present invention is not limited to the above embodiments. Unless otherwise specified, the methods used in the above embodiments are conventional methods.

Claims

1. A sintering machine air leakage detection and alarm system based on sensing technology, characterized in that, The application relates to a sintering machine air leakage detection system. The system comprises: a sensor group, which is respectively installed at different positions of a sintering machine, collects data including oxygen content, pressure and temperature of the sintering machine and transmits the data to a data processing module and a data storage module; a data processing module, which performs reliability evaluation, interactive verification and correction fusion processing on the data collected by the sensor group to determine the air leakage condition of the sintering machine; an alarm module, which is used for sending an alarm signal when the air leakage condition exceeds a preset threshold after the data is processed by the data processing module; 2. The sintering machine air leakage detection and alarm system based on sensing technology according to claim 1, characterized in that, a data storage module, which is used for storing the original data collected by the sensor group, intermediate data in the data processing process of the data processing module, final fusion data and alarm records of the alarm module. The sensor group comprises oxygen content sensors, pressure sensors and temperature sensors. The oxygen content sensors are arranged at positions capable of stably detecting oxygen concentration changes in flues of the sintering machine and are used for detecting oxygen content data at different positions in the sintering machine. The pressure sensors are arranged at the inlet and outlet of a wind box of the sintering machine, the measurement surface of the pressure sensors is perpendicular to the airflow direction of the inlet and outlet, and the pressure sensors are used for detecting pressure data.

3. The sintering machine air leakage detection and alarm system based on sensing technology according to claim 1, characterized in that, The temperature sensors are arranged around the sealing area between a sintering machine trolley and a slide, and are used for detecting temperature data. The data processing module comprises: a correlation matrix unit, which constructs a sensor data correlation matrix, calculates the covariance of a sensor pair with correlation to determine data consistency and quantifies data uncertainty according to information entropy; a reliability evaluation unit, which performs reliability evaluation on the data collected by the sensor group based on Bayesian estimation; 4. The sintering machine air leakage detection and alarm system based on sensing technology according to claim 3, characterized in that, a correction unit, which constructs a sensor data distribution model based on a likelihood function and performs correction considering the spatial correlation between sensor data.

5. The sintering machine air leakage detection and alarm system based on sensing technology according to claim 3, characterized in that, The correlation matrix unit further comprises a suspicious node marking subunit, if the covariance of the sensor group exceeds a preset threshold, the posterior probability based on Bayesian estimation is less than a set threshold, and the information entropy of the sensor is greater than a set information entropy difference threshold of adjacent nodes, the sensor is marked as a suspicious node.

6. The sintering machine air leakage detection and alarm system based on sensing technology according to claim 1, characterized in that, The correction unit further comprises a weight determination subunit, for a sensor verified as non-suspect and reliable by the reliability evaluation unit, the weight of the sensor is determined according to the posterior probability based on Bayesian estimation and a data consistency index. The alarm module comprises: an alarm signal triggering unit, which continuously receives analysis result data about the air leakage condition of the sintering machine from the data processing module, and starts an alarm signal triggering mechanism when the data shows that the air leakage rate of a certain area exceeds a preset alarm threshold; 7. The sintering machine air leakage detection and alarm system based on sensing technology according to claim 1, characterized in that, an alarm signal output unit, which sends an audible and visual alarm signal according to the instruction of the alarm signal triggering unit. The data storage module comprises: an original data storage unit, which stores the original data collected by the sensors in a special original database table with the number of the sensors and the data collection time as indexes; an intermediate data storage unit, which stores the intermediate data generated in the data processing process, including data reliability evaluation results and covariance calculation values, in association with the corresponding original data in an intermediate database table; a recording unit, which separately establishes a database table for storing fusion data and alarm records.

8. The sintering machine air leakage detection and alarm system based on sensing technology according to claim 1, characterized in that, The data processing module further comprises a machine learning algorithm module, adopts a neural network algorithm, takes the data collected by the multi-sensor group as input and takes the air leakage type as output, and optimizes the weight and threshold of the neural network through training.

9. The sintering machine air leakage detection and alarm system based on sensing technology according to claim 3, characterized in that, It also includes a sensor self-calibration function module, which automatically adjusts the calibration parameters of the sensor according to the data reliability evaluation results and environmental parameter changes, to ensure that the sensor is always in the best working state.

10. The sintering machine air leakage detection and alarm system based on sensing technology according to claim 1, characterized in that, It also includes a data compression algorithm module for reducing data transmission volume and transmission time.

Images