An industrial waste gas purification equipment fault data analysis method and system

Abstract

The application provides an industrial waste gas purification equipment fault data analysis method and system, and relates to the field of industrial waste gas purification equipment fault data analysis. In the periodic reversing operation of the industrial waste gas purification equipment, the operation data of each regenerative bed layer are acquired, the quantitative indexes corresponding to the operation data of each regenerative bed layer are calculated based on the operation data, and the potential fault is identified by combining the relative deviation indexes calculated among the multiple regenerative bed layers and the time sequence dynamic analysis of the relative deviation behaviors of the single regenerative bed layer. The method has the advantage that the early and subtle potential physical fault of the equipment can be captured more sensitively.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of industrial waste gas purification equipment fault data analysis, in particular to an industrial waste gas purification equipment fault data analysis method and system. BACKGROUND

[0002] In the industrial production process, industrial waste gas purification equipment is a key facility to ensure environmental compliance and production safety. For example, a regenerative thermal oxidizer (RTO) usually contains multiple regenerative bed layers connected in parallel, which preheat, oxidize, and recover heat from the waste gas through periodic flow switching. In order to monitor the running state of the equipment, a large number of sensors are usually arranged to collect real-time data such as combustion chamber temperature, regenerative bed layer temperature, pressure, waste gas composition (such as VOCs concentration), and valve state. The operating personnel mainly rely on the real-time measurement values displayed on the central control system interface and the preset fixed alarm thresholds to determine whether the equipment is running normally.

[0003] However, in the actual operating environment, this traditional monitoring method has significant limitations. Industrial waste gas purification equipment is operated in harsh conditions such as high temperature and corrosion for a long time, and equipment components such as regenerators, valves, and burners will gradually wear out or deteriorate, which may cause early and hidden faults. For example, the regenerator bed layer may be slightly broken and clogged due to long-term use, causing the air flow resistance through the bed layer to slowly increase, or the temperature response characteristics of the bed layer to change slightly during periodic switching operations. For example, the sealing element of the switching valve may have a slight internal leak, causing a temporary bypass of pollutants during valve switching or in certain combined states. The wear of the burner nozzle or the performance degradation of the ignition system may cause incomplete combustion, resulting in flame signal fluctuations or slight increases in pollutant concentration under certain operating conditions. These early faults often do not immediately cause any single measurement value to exceed the fixed alarm threshold. On the contrary, they may manifest as subtle changes in the data, relative performance differences between other regenerative bed layers in the switching sub-stage, or slow abnormal evolution of certain parameters over time. These subtle fault signals are easily masked by normal production condition fluctuations (such as changes in waste gas flow and concentration) and sensor drift, noise, or occasional errors that may exist, making it difficult for operating personnel to discover them in a timely manner by directly observing real-time data or simple historical trend comparisons.

[0004] In addition, different types of faults have different performance patterns at the data level, which may involve different combinations of measurement parameters, different time scales, such as transient spikes or slow trends, and different evolution characteristics. Traditional simple rules or statistical models based on a small number of known fault patterns are difficult to effectively cover all types of faults that may occur in the device, especially for complex faults or new fault patterns. At the same time, the quality problem of sensor data directly affects the reliability of data analysis. If the original data is not effectively preprocessed and quality evaluated, the analysis conclusion based on the flawed data may lead to false positives or false negatives. Finally, many faults are a gradual development process, and existing methods often fail to accurately capture the specific stage of the fault and its development trend, thus failing to provide effective support for formulating a forward-looking maintenance plan. SUMMARY

[0005] The present application provides an industrial waste gas purification equipment fault data analysis method and system, which has the advantage of being able to more sensitively capture early and subtle potential physical faults of the equipment.

[0006] In one aspect, the present application provides an industrial waste gas purification equipment fault data analysis method, comprising:

[0007] In the periodic reversing operation of the industrial waste gas purification equipment, the operating data of each regenerative bed layer is obtained, wherein the industrial waste gas purification equipment comprises a plurality of regenerative bed layers connected in parallel with each other, and the periodic reversing operation comprises a plurality of reversing sub-stages; the operating data is obtained in real time by a sensor group arranged in each regenerative bed layer, and the operating data includes temperature, pressure difference and VOCs concentration;

[0008] Quantitative indicators corresponding to the operating data of each regenerative bed layer are calculated, and the quantitative indicators include heat recovery efficiency, pressure drop change rate and VOCs purification efficiency;

[0009] In the same reversing sub-stage, for a first regenerative bed layer in the plurality of regenerative bed layers, a relative deviation indicator representing the operating performance difference of the first regenerative bed layer relative to a second regenerative bed layer is calculated based on the quantitative indicator of the first regenerative bed layer and the statistical correlation of the quantitative indicators corresponding to any second regenerative bed layer different from the first regenerative bed layer in the plurality of regenerative bed layers;

[0010] The relative deviation indicators of the first regenerative bed layer calculated in each reversing sub-stage are arranged in time sequence to form a time series in units of a complete cycle of the periodic reversing operation;

[0011] When the dynamic characteristics of the time series meet an abnormal condition, the first regenerative bed layer is identified as having a potential physical fault, and the abnormal condition indicates that the behavior pattern of the first regenerative bed layer deviates from the behavior pattern of the second regenerative bed layer.

[0012] Optionally, the plurality of sub-phase of the periodic switching operation is segmented by switching valve action signals, the plurality of sub-phase including an intake sub-phase, a purge sub-phase, a preheat sub-phase and an exhaust sub-phase.

[0013] Optionally, the step of calculating the quantified indicators corresponding to the operation data of each of the plurality of regenerative beds based on the operation data, the quantified indicators including thermal recovery efficiency, pressure drop change rate and VOCs purification efficiency, comprises:

[0014] acquiring working condition indicative data indicative of a current working condition of the industrial waste gas purification device, the working condition indicative data being extracted from the operation data, the working condition indicative data including at least an inlet waste gas flow indicative value or an inlet waste gas component concentration indicative value;

[0015] determining a set of quantified indicator calculation parameters for characterizing the current working condition based on the working condition indicative data, the set of quantified indicator calculation parameters including first weight coefficients or correction factors for calculating each quantified indicator;

[0016] calculating quantified indicators corresponding to the operation data of each of the plurality of regenerative beds according to the operation data of each of the plurality of regenerative beds and the set of quantified indicator calculation parameters, the quantified indicators including thermal recovery efficiency, pressure drop change rate and VOCs purification efficiency.

[0017] Optionally, the step of calculating, for a first regenerative bed of the plurality of regenerative beds, a relative deviation indicator indicative of a difference in performance between the first regenerative bed and a second regenerative bed of the plurality of regenerative beds based on a statistical correlation between the quantified indicators of the first regenerative bed and the quantified indicators corresponding to the second regenerative bed within the same sub-phase of the periodic switching operation, comprises:

[0018] for each quantified indicator of the first regenerative bed, determining an individual deviation degree of the quantified indicator of the first regenerative bed based on the quantified indicator of the first regenerative bed and the corresponding quantified indicator of the second regenerative bed;

[0019] calculating, based on the individual deviation degrees of the plurality of quantified indicators of the first regenerative bed, a relative deviation indicator indicative of a difference in performance between the first regenerative bed and the second regenerative bed.

[0020] Optionally, the step of calculating, for a first regenerative bed of the plurality of regenerative beds, a relative deviation indicator indicative of a difference in performance between the first regenerative bed and a second regenerative bed of the plurality of regenerative beds based on a statistical correlation between the quantified indicators of the first regenerative bed and the quantified indicators corresponding to the second regenerative bed within the same sub-phase of the periodic switching operation, comprises:

[0021] The individual deviation degree of the plurality of quantitative indexes of the first regenerative bed layer is subjected to numerical conversion processing to obtain a target individual deviation degree;

[0022] A second weight coefficient is determined, which is a preset parameter determined according to the operating state of the industrial waste gas purification equipment;

[0023] According to the second weight coefficient and the target individual deviation degree, a combination operation is performed to calculate the relative deviation index.

[0024] Optionally, when the dynamic characteristics of the time series meet the abnormal condition, the first regenerative bed layer is identified as having a potential physical fault, and the step of identifying the first regenerative bed layer as having a potential physical fault when the dynamic characteristics of the time series meet the abnormal condition includes:

[0025] The time series is subjected to dynamic characteristic analysis to extract statistical features, trend features, and correlation features;

[0026] When the abnormal condition is met, the first regenerative bed layer is identified as having a potential physical fault, wherein the abnormal condition is determined based on statistical outliers, trend mutations, and correlation abnormalities.

[0027] Optionally, when the abnormal condition is met, the first regenerative bed layer is identified as having a potential physical fault, wherein the abnormal condition is determined based on statistical outliers, trend mutations, and correlation abnormalities.

[0028] In each sub-stage of the periodic switching operation, stage identification information representing the type of the current sub-stage is obtained;

[0029] According to the stage identification information, statistical outlier determination parameters for statistical outlier analysis of the time series in the current sub-stage, trend mutation determination parameters for trend mutation analysis, and correlation anomaly determination parameters for correlation anomaly analysis are determined;

[0030] The statistical outlier determination parameters are used to determine the statistical outlier condition of the time series to obtain a statistical outlier determination output; the trend mutation determination parameters are used to determine the trend mutation condition of the time series to obtain a trend mutation determination output; and the correlation anomaly determination parameters are used to determine the correlation anomaly condition of the time series to obtain a correlation anomaly determination output;

[0031] If the statistical outlier determination output, the trend mutation determination output, and the correlation anomaly determination output collectively indicate an abnormal state, it is determined that the abnormal condition is met.

[0032] Optionally, the step of calculating, in the same sub-phase of switching, for a first regenerative bed layer of the plurality of regenerative bed layers, a relative deviation index representing a difference in performance between the first regenerative bed layer and a second regenerative bed layer of the plurality of regenerative bed layers based on a statistical correlation between the quantification index of the first regenerative bed layer and the quantification index corresponding to the second regenerative bed layer different from the first regenerative bed layer comprises:

[0033] In the same sub-phase of switching, obtaining the quantification index of the second regenerative bed layer for a first regenerative bed layer of the plurality of regenerative bed layers as an initial reference quantification index;

[0034] Judging the applicability of the initial reference quantification index as a reference to obtain an applicability judgment result;

[0035] Based on the applicability judgment result, adjusting the use of the initial reference quantification index to form a final reference quantification index, wherein if the applicability judgment result indicates that the initial reference quantification index of the second regenerative bed layer is not applicable as a reference, the degree of influence of the initial reference quantification index of the second regenerative bed layer on the calculation result of the relative deviation index is reduced;

[0036] For the first regenerative bed layer, based on the statistical correlation between the quantification index and the initial reference quantification index, a relative deviation index representing a difference in performance between the first regenerative bed layer and the second regenerative bed layer is calculated.

[0037] On the other hand, the present application provides an industrial waste gas purification equipment fault data analysis system, comprising:

[0038] A data acquisition module is configured to obtain operation data of each regenerative bed layer during periodic switching operation of the industrial waste gas purification equipment, wherein the industrial waste gas purification equipment comprises a plurality of regenerative bed layers connected in parallel, and the periodic switching operation comprises a plurality of sub-phases of switching; the operation data is obtained in real time by a sensor group arranged in each regenerative bed layer, and the operation data includes temperature, pressure difference and VOCs concentration;

[0039] An index calculation module is configured to calculate quantification indexes corresponding to the operation data of each regenerative bed layer, wherein the quantification indexes include heat recovery efficiency, pressure drop change rate and VOCs purification efficiency;

[0040] A relative deviation calculation module is configured to calculate, in the same sub-phase of switching, for a first regenerative bed layer of the plurality of regenerative bed layers, a relative deviation index representing a difference in performance between the first regenerative bed layer and a second regenerative bed layer of the plurality of regenerative bed layers based on a statistical correlation between the quantification index of the first regenerative bed layer and the quantification index corresponding to the second regenerative bed layer different from the first regenerative bed layer;

[0041] a time series tracking module, configured to arrange the relative deviation indicators calculated by the first regenerative bed layer at each sub-stage of each commutation in time sequence to form a time series, in units of a complete cycle of the periodic commutation operation;

[0042] a fault identification module, configured to identify the first regenerative bed layer as having a potential physical fault when the dynamic characteristics of the time series meet an abnormal condition, the abnormal condition indicating that the behavior pattern of the first regenerative bed layer deviates from the behavior pattern of the second regenerative bed layer.

[0043] The application provides an industrial waste gas purification equipment fault data analysis method and system, which identifies potential faults based on relative deviation indicators and time series dynamic characteristic analysis between regenerative bed layers, and has the advantage of being able to more sensitively capture early and subtle potential physical faults of the equipment. BRIEF DESCRIPTION OF DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the present application, the drawings needed in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0045] Figure 1 Fig. 1 exemplarily shows a schematic diagram of an industrial waste gas purification equipment fault data analysis method in the embodiment;

[0046] Figure 2 Fig. 2 exemplarily shows a module configuration block diagram of an industrial waste gas purification equipment fault data analysis system 100 in the embodiment.

[0047] Fig. 1 exemplarily shows a schematic diagram of an industrial waste gas purification equipment fault data analysis method in the embodiment; DETAILED DESCRIPTION

[0048] The technical solutions of the present application will be described in detail below with reference to the drawings in the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. The components of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0049] It should be noted that similar reference numerals and letters refer to like items in the several figures of the drawings, and that, as such, once an item is defined in one figure, it should not have to be further defined and explained in the subsequent figures. Also, in the description of the present application, the terms "first", "second", and so on are used merely to distinguish one item from another, and are not to be construed as indicating or implying relative importance.

[0050] Traditional industrial waste gas purification equipment, especially the system using multiple sets of regenerative thermal oxidizers (RTO), mainly relies on real-time measurement values and pre-set fixed alarm thresholds when monitoring the running state. This method has the problem of being difficult to effectively identify when early and hidden faults occur in the equipment. For example, slight blockage of the regenerator bed or slight internal leakage of the reversing valve, the initial data changes are weak, which are easily covered by normal production condition fluctuations, environmental influences, and sensor instability, making it difficult for users to discover problems through direct observation or simple comparison. If these subtle abnormal signals are not captured and analyzed in time, the fault may further develop, affecting the performance of the equipment.

[0051] For example, assume that a set of RTO equipment contains multiple parallel regenerator beds. When the regenerator in one of the beds starts to slightly powder and causes partial airflow passage obstruction, during the purging or preheating sub-phase of the periodic reversing operation of the RTO, the airflow resistance through that bed will slightly increase, causing the pressure difference sensor reading to show a slow upward trend, and the temperature rise rate of that bed may be slightly lower than that of other normal beds. These changes may be very small, superimposed on the normal data fluctuations caused by waste gas flow, concentration fluctuations, and environmental temperature changes, making it difficult to trigger a threshold alarm for a single bed pressure difference or temperature, and also difficult to identify abnormalities through simple historical data comparison.

[0052] If the above problems are not solved, early potential physical faults of industrial waste gas purification equipment will not be discovered in time. This will lead to the continuous development of the fault, which may cause the performance of the equipment to gradually decline, such as the reduction of VOCs purification efficiency, which cannot meet the emission standard requirements. In severe cases, it may cause damage to key components of the equipment, cause unexpected shutdown, cause production interruption and economic loss. At the same time, due to the failure to give early warning, maintenance can only be carried out after the fault occurs, increasing the urgency and cost of repair, and also failing to achieve predictive maintenance based on the state of the equipment.

[0053] In the face of the above problems, the first thought of the present application is to set more fine dynamic threshold for the key operating parameters of each regenerative bed or to carry out more complex single variable trend analysis. However, this method is still difficult to effectively distinguish the normal working condition fluctuation from the weak change caused by early failure, and cannot capture the correlation change between different parameters. Further, considering that the RTO device contains multiple parallel regenerative beds with similar structure and function, they should normally show similar behavior patterns. Therefore, the parallel structure can be used for mutual reference, and the difference in relative performance between one bed and the rest of the beds in the same operating stage is compared to identify abnormalities. At the same time, considering that the failure is a dynamic evolution process, the relative difference should also be tracked and analyzed over time. Based on this idea, the present application proposes a method for calculating the relative deviation index of each bed relative to other beds, and forming a time series in chronological order, and then analyzing the dynamic characteristics of the time series to determine whether there is a potential failure.

[0054] As shown in Figure 1 An exemplary flowchart of a method for analyzing failure data of an industrial waste gas purification device in some embodiments is shown. The present application proposes a method for analyzing failure data of an industrial waste gas purification device, comprising:

[0055] S10, in the periodic reversing operation of the industrial waste gas purification device, the operating data of each regenerative bed is obtained, wherein the industrial waste gas purification device comprises multiple regenerative beds in parallel, and the periodic reversing operation comprises multiple reversing sub-stages; the operating data is obtained in real time by a sensor group arranged in each regenerative bed, and the operating data includes temperature, pressure difference and VOCs concentration;

[0056] S20, the quantitative index corresponding to the operating data of each regenerative bed is calculated, and the quantitative index includes heat recovery efficiency, pressure drop change rate and VOCs purification efficiency;

[0057] S30, in the same reversing sub-stage, for a first regenerative bed in the multiple regenerative beds, based on the statistical correlation between the quantitative index of the first regenerative bed and the quantitative index corresponding to any second regenerative bed different from the first regenerative bed in the multiple regenerative beds, a relative deviation index is calculated to represent the difference in operating performance of the first regenerative bed relative to the second regenerative bed;

[0058] S40, in the complete cycle of the periodic reversing operation, the relative deviation index of the first regenerative bed calculated in each reversing sub-stage is arranged in chronological order to form a time series;

[0059] S50, when the dynamic characteristic of the time sequence satisfies an abnormal condition, identifying that the first regenerative bed layer has a potential physical fault, the abnormal condition indicating that the behavior pattern of the first regenerative bed layer deviates from the behavior pattern of the second regenerative bed layer.

[0060] wherein, the periodic reversing operation refers to the process that the industrial waste gas purification equipment, such as the regenerative thermal incinerator RTO, switches the gas flow direction during operation, so that the waste gas is sequentially preheated, reacted and cooled in different regenerative bed layers. The multiple reversing sub-stages of the periodic reversing operation are divided by the reversing valve action signals, and the multiple reversing sub-stages include the gas inlet sub-stage, the purge sub-stage, the preheating sub-stage and the gas outlet sub-stage, which together realize the heat recovery and effective purification of VOCs. The reversing valve action signal refers to the signal sent or received by the valve (such as a poppet valve or a rotary valve) that controls the switching of the gas flow, indicating the change of its state. The gas inlet sub-stage refers to the period when the waste gas enters the regenerative bed layer for treatment. The purge sub-stage refers to the period when clean gas is used to remove residual waste gas in the regenerative bed layer before switching the gas flow direction. The preheating sub-stage refers to the period when the regenerative bed layer increases its temperature by introducing hot gas or other methods before entering the high-temperature oxidation state. The gas outlet sub-stage refers to the period when the treated clean gas is discharged from the equipment through the regenerative bed layer.

[0061] wherein, the operation data refers to the measurement values reflecting the current state of the equipment obtained in real time by the sensor groups arranged in each regenerative bed layer, including temperature, pressure difference and VOCs concentration, which are the basis for analyzing the performance and state of the equipment.

[0062] wherein, the quantitative indicators refer to the values calculated based on the operation data, which can represent the specific performance or state of the regenerative bed layer, including heat recovery efficiency, pressure drop change rate and VOCs purification efficiency, which convert the original measurement data into features with more physical meaning and diagnostic value.

[0063] wherein, the relative deviation indicator refers to the value calculated based on the statistical correlation between the quantitative indicator of the first regenerative bed layer and the quantitative indicator of any second regenerative bed layer different from the first regenerative bed layer in the same reversing sub-stage, which represents the difference between the performance of the first regenerative bed layer and the performance of the second regenerative bed layer in the current reversing sub-stage, and the purpose is to highlight individual abnormalities and suppress common interference through horizontal comparison between bed layers.

[0064] Specifically, the first regenerative bed layer refers to the bed layer selected as the analysis target among the multiple regenerative bed layers, and the second regenerative bed layer refers to the other regenerative bed layer different from the first regenerative bed layer, and the quantitative indicator of the second regenerative bed layer is used as a reference to provide baseline data for comparison.

[0065] The time series refers to a sequence formed by arranging the relative deviation indicators of each regenerative bed layer in each reversing sub-stage in time sequence, which records the dynamic evolution process of the relative deviation behavior of the bed layer over time, and its purpose is to provide a data basis for subsequent dynamic characteristic analysis.

[0066] The core innovation of the present application is that by combining the relative comparison between multiple regenerative bed layers (calculating the relative deviation indicator) and the dynamic analysis of the time series of the relative deviation behavior of a single regenerative bed layer, the abnormal evolution of the bed layer behavior pattern caused by early and subtle faults is captured, the accuracy and sensitivity of fault identification are improved, and the effect of early warning is achieved.

[0067] Specifically, the present application realizes the identification of potential physical faults of industrial waste gas purification equipment through a series of steps.

[0068] Firstly, when the equipment is operated periodically, the operation data of each regenerative bed layer are obtained in real time, which comprehensively reflects the physical state of the bed layer in different operation stages. Based on these raw operation data, quantitative indicators that can directly reflect the performance of the bed layer are calculated, such as heat recovery efficiency, pressure drop change rate and VOCs purification efficiency. These quantitative indicators are the refinement and transformation of the raw data, which are more convenient for subsequent analysis.

[0069] Secondly, for each regenerative bed layer, the relative deviation indicator of the bed layer relative to other bed layers is calculated by analyzing the statistical correlation between the quantitative indicator of the bed layer and the corresponding quantitative indicators of the remaining bed layers in the same reversing sub-stage. This step takes advantage of the characteristics of multiple bed layers in parallel, and through horizontal comparison, it can effectively distinguish the abnormality of individual bed layer from the overall working condition fluctuation or common error of sensors, and highlight the bed layer with real problems.

[0070] Subsequently, the relative deviation indicators of each bed layer calculated in different reversing sub-stages are arranged in time sequence to form their respective time series, with the complete cycle of periodic reversing operation as the unit. These time series record the dynamic trajectory of the relative deviation behavior of each bed layer over time.

[0071] Finally, the dynamic characteristics of these time series are analyzed, and when the dynamic characteristics meet the preset abnormal conditions, i.e., the behavior pattern of the bed layer continuously or significantly deviates from the behavior pattern of other bed layers, it is identified that the regenerative bed layer has potential physical faults. The whole process can extract the signals reflecting early and subtle faults from complex and variable data by combining the relative comparison between bed layers and the dynamic tracking of bed layer behavior over time, and realize the timely discovery of potential problems.

[0072] In a specific embodiment, it is assumed that the industrial exhaust gas purification device comprises three parallel regenerative bed layers A, B, C. During the periodic commutation operation of the device, the temperature, pressure difference and VOCs concentration data of bed layers A, B, C at each commutation sub-stage such as gas inlet, purging, preheating, gas outlet, etc. are collected in real time through the sensor groups arranged in each bed layer. Based on these data, the quantitative indicators of each bed layer at each commutation sub-stage are calculated, for example, the heat recovery efficiency, pressure drop change rate and VOCs purification efficiency of bed layer A at the gas inlet sub-stage are calculated. In the same gas inlet sub-stage, the statistical correlation of these quantitative indicators of bed layer A with the corresponding quantitative indicators of bed layers B and C is calculated, for example, the difference or ratio of the heat recovery efficiency of bed layer A with the average or median of the heat recovery efficiencies of bed layers B and C is calculated, to obtain the relative deviation indicator of bed layer A at the gas inlet sub-stage. Similar calculations are also performed for bed layers B and C at the gas inlet sub-stage and for all bed layers at other commutation sub-stages. Then, the relative deviation indicators of bed layer A calculated at each gas inlet sub-stage in a plurality of consecutive cycles are arranged in chronological order to form a time series of the relative deviation indicators of bed layer A at the gas inlet sub-stage. Dynamic characteristic analysis is performed on the time series, for example, statistical characteristics such as mean, variance, trend, periodicity, etc. are analyzed, as well as whether there is a mutation point or a change in correlation with the corresponding time series of other bed layers. If the analysis result (i.e. dynamic characteristic) meets the preset abnormal condition, for example, the mean of the time series is continuously higher than the normal range, or shows a significant upward trend, it is identified that bed layer A has a potential physical fault, for example, the regenerator is blocked or the heat exchange performance is decreased.

[0073] Through the above technical solution, the data noise and fluctuations generated by the industrial exhaust gas purification device under complex working conditions can be effectively dealt with. By introducing the relative deviation indicator, differential evaluation of the operating state of a single regenerative bed layer relative to other bed layers is realized, effectively filtering out common interference. Further, by constructing a time series of the relative deviation indicator and analyzing its dynamic characteristics, subtle changes and evolution trends from the inception to the development of the fault can be captured, thereby achieving early and sensitive identification of potential physical faults. This helps to take timely maintenance measures to avoid performance degradation, emission exceeding or unexpected shutdown of the device, and ensures stable operation of the device.

[0074] In some embodiments, the present application further proposes that the step of calculating the quantitative indicators corresponding to the operating data for each regenerative bed layer based on the operating data includes:

[0075] Obtaining working condition indication data representing the current working condition of the industrial exhaust gas purification device, the working condition indication data being extracted from the operating data, and the working condition indication data at least including an exhaust gas inlet flow indication value or an exhaust gas inlet component concentration indication value;

[0076] The working condition indication data is used to determine a set of quantitative index calculation parameters for characterizing the current operation condition, and the set of quantitative index calculation parameters includes first weight coefficients or correction factors for calculating each quantitative index.

[0077] According to the operation data of each regenerative bed and the set of quantitative index calculation parameters, the quantitative index corresponding to the operation data of each regenerative bed is calculated, and the quantitative index includes heat recovery efficiency, pressure drop change rate and VOCs purification efficiency.

[0078] The working condition indication data is data extracted from the operation data and capable of reflecting the current operation state of the industrial waste gas purification equipment, and at least includes an inlet flow indication value or an inlet component concentration indication value of the waste gas, and specifically can be real-time acquisition of the total inlet flow data or the total VOCs concentration data of the waste gas from the sensor group. The purpose is to provide a basis for subsequent adjustment of the quantitative index calculation parameters.

[0079] The set of quantitative index calculation parameters refers to a set of parameters for adjusting the calculation method of the quantitative index, which includes first weight coefficients or correction factors for calculating each quantitative index, and specifically can be a set of preset numerical values or a set of numerical values calculated by a model. The purpose is to make the calculation of the quantitative index adapt to different operation conditions.

[0080] The first weight coefficient refers to a multiplication factor in the quantitative index calculation formula for adjusting the importance of different operation data items or intermediate calculation results, which can reflect the difference in the contribution of different data items to the quantitative index under a specific working condition. The correction factor refers to an addition or multiplication correction term based on the quantitative index calculation result, which is used to compensate or correct the systematic deviation that may occur under a specific working condition.

[0081] The present application dynamically adjusts the calculation parameters of the quantitative index according to the current operation condition of the industrial waste gas purification equipment, thereby improving the accuracy of the quantitative index calculation.

[0082] Specifically, first, the working condition indication data characterizing the current operation condition of the equipment is obtained, which directly reflects the load or processing difficulty of the equipment. Based on these working condition indication data, the system determines a set of quantitative index calculation parameters matched with the current working condition, which contains first weight coefficients or correction factors for adjusting the calculation of the quantitative index. These parameters can adjust the influence weight of the operation data in the calculation according to different working conditions, or correct the calculation result.

[0083] According to the real-time operation data of each regenerative bed and the dynamically determined parameter set, quantitative indexes such as heat recovery efficiency, pressure drop change rate and VOCs purification efficiency of each regenerative bed are calculated. Due to the calculation process of the quantitative indexes considering the specific operation condition of the equipment and adjusting the calculation parameters accordingly, the calculation results can more accurately reflect the real performance of the equipment under the working condition. The accurate quantitative indexes provide reliable input for the subsequent fault diagnosis steps, thereby improving the reliability of the entire fault diagnosis method.

[0084] In some preferred embodiments, specifically, the step of obtaining the working condition indicating data characterizing the current operation condition of the industrial waste gas purification equipment can be obtaining the waste gas inlet flow indicating value by reading the flow sensor on the waste gas inlet header, or obtaining the waste gas inlet component concentration indicating value by reading the VOCs concentration sensor on the waste gas inlet header. The step of determining the quantitative index calculation parameter set based on the working condition indicating data can be achieved by consulting a pre-set working condition parameter table, which records the first weight coefficient and the correction factor set corresponding to different waste gas inlet flow ranges or different waste gas inlet component concentration ranges. For example, when the waste gas inlet flow indicating value is in the low flow range, the parameter set corresponding to the low load working condition is selected; when it is in the high flow range, the parameter set corresponding to the high load working condition is selected. When calculating the quantitative indexes according to the operation data of each regenerative bed and the quantitative index calculation parameter set, for example, when calculating the heat recovery efficiency, a formula containing multiple temperature measurement values and flow measurement values can be used, and the first weight coefficient in the parameter set is applied to these measurement values or the combination calculation results, and the correction factor is used to adjust the final calculation result. The calculation of pressure drop change rate and VOCs purification efficiency also adopts a similar way, and is calculated according to the respective calculation formula and the corresponding parameters in the parameter set.

[0085] Through the above technical solutions, the calculation parameters of the quantitative indexes are dynamically adjusted according to the current operation condition of the industrial waste gas purification equipment, so that the calculated quantitative indexes can more accurately reflect the real performance state of the equipment under the working condition, thereby improving the accuracy of the quantitative indexes. The improvement of the accuracy of the quantitative indexes provides a more reliable data basis for the subsequent relative deviation index calculation, time series analysis and final fault diagnosis, thereby improving the reliability and accuracy of the entire fault diagnosis method.

[0086] In some embodiments, in the same reversing sub-stage, for a first regenerative bed in the plurality of regenerative beds, based on the quantitative index of the first regenerative bed, and the statistical correlation of the quantitative index corresponding to any second regenerative bed different from the first regenerative bed in the plurality of regenerative beds, the step of calculating the relative deviation index characterizing the performance difference of the first regenerative bed relative to the second regenerative bed includes:

[0087] For each quantitative index of the first regenerative bed layer, based on the quantitative index of the first regenerative bed layer and the corresponding quantitative index of the second regenerative bed layer, the individual deviation degree of the quantitative index of the first regenerative bed layer is determined;

[0088] Based on the individual deviation degrees of the multiple quantitative indexes of the first regenerative bed layer, a relative deviation index representing the difference in operation performance of the first regenerative bed layer relative to the second regenerative bed layer is calculated.

[0089] Wherein, determining the individual deviation degree of the quantitative index of the first regenerative bed layer means measuring the difference between the performance of the first regenerative bed layer in a certain specific quantitative index and the performance of other regenerative bed layers in the same batch in the quantitative index, which can be achieved by calculating the difference, ratio, percentage deviation of the value of the quantitative index of the regenerative bed layer and the average or median of the values of the quantitative index of other regenerative bed layers in the same period, or calculating the Z-score, percentile, etc. statistical quantity of the value in the distribution of the quantitative index of other bed layers, the purpose of which is to quantify the abnormality or deviation of each regenerative bed layer in a single dimension.

[0090] Wherein, based on the individual deviation degrees of the multiple quantitative indexes of each first regenerative bed layer, a relative deviation index representing the difference in operation performance of the regenerative bed layer relative to other second regenerative bed layers is calculated, which can be obtained by comprehensively processing the multiple individual deviation degrees calculated on different quantitative indexes of the regenerative bed layer to obtain a single or composite index that can reflect the overall operation state of the regenerative bed layer relative to other bed layers, which can integrate multi-dimensional information and more accurately evaluate the overall operation performance difference of the regenerative bed layer, thereby improving the accuracy of fault identification.

[0091] The present application first determines the individual deviation degree of each quantitative index of each first regenerative bed layer based on the quantitative index and the corresponding quantitative index of other regenerative bed layers in the same commutation sub-stage, which means that the performance differences of regenerative bed layers in heat recovery efficiency, pressure drop change rate, VOCs purification efficiency, etc. are quantified respectively.

[0092] Subsequently, based on the individual deviation degrees of the multiple quantitative indexes of the regenerative bed layer, a relative deviation index representing the difference in operation performance of the regenerative bed layer relative to other regenerative bed layers is calculated. This step-by-step processing method avoids the confusion or masking effect that may occur when directly correlating quantitative indexes of different properties. By calculating the individual deviation degree respectively, it can be clearly seen that the regenerative bed layer has abnormalities in which specific performance indicators.

[0093] Further, by synthesizing these individual deviation degrees, a more comprehensive and refined overall deviation indicator can be obtained. This approach can more effectively capture subtle and possibly inconsistent deviation signals produced by specific physical faults on different quantitative indicators. For example, a slightly clogged bed layer can exhibit a slight increase in pressure drop change rate and a slight decrease in heat recovery efficiency. By calculating the individual deviation degrees of these two indicators respectively and synthesizing them, this early fault indication can be more accurately identified than relying on a single indicator or simple synthesis. This refined and synthesized processing allows the calculated relative deviation indicator to more accurately reflect the true operating state of the regenerative bed layer, thereby improving the accuracy of subsequent fault identification.

[0094] In some embodiments, assuming that the industrial exhaust gas purification device comprises N regenerative bed layers, for a specific reversing sub-stage, for one of the regenerative bed layers, for example, labeled as bed layer i. First, obtain the values of various quantitative indicators of bed layer i in this reversing sub-stage, such as heat recovery efficiency, pressure drop change rate, and VOCs purification efficiency. At the same time, obtain the values of the quantitative indicators corresponding to any second regenerative bed layer in this reversing sub-stage. For the heat recovery efficiency quantitative indicator, calculate the difference between the heat recovery efficiency value of bed layer i and the statistical value (such as the mean or median) of the heat recovery efficiency values of the second regenerative bed, to determine the individual deviation degree of bed layer i in heat recovery efficiency. Similarly, calculate the individual deviation degrees of bed layer i in pressure drop change rate and VOCs purification efficiency, respectively. After obtaining the individual deviation degrees of bed layer i in heat recovery efficiency, pressure drop change rate, and VOCs purification efficiency, synthesize these individual deviation degrees, for example, by simply combining them in some form (such as summation, averaging, or more complex function operations), to calculate a relative deviation indicator representing the performance difference of bed layer i relative to other regenerative bed layers in this reversing sub-stage. The higher this relative deviation indicator value, the more likely it means that the performance of bed layer i in this reversing sub-stage is more different or abnormal compared to other bed layers.

[0095] By the above technical solutions, the performance differences of regenerative bed layers on different quantitative indicators are decomposed and quantified, and then synthesized for evaluation, which can more carefully capture the effects of different types of faults in different dimensions, avoiding the limitations of single indicators or simple synthesis. This allows the calculated relative deviation indicator to more accurately reflect the true operating state of the regenerative bed layer, improving the accuracy of identifying potential physical faults.

[0096] In some embodiments, based on the individual deviation degrees of the first regenerative bed layer in multiple quantitative indicators, the step of calculating a relative deviation indicator representing the performance difference of the regenerative bed layer relative to the second regenerative bed layer includes:

[0097] The individual deviation degree of the plurality of quantitative indicators of the first regenerative bed layer is subjected to numerical conversion processing to obtain a target individual deviation degree;

[0098] A second weight coefficient is determined, which is a preset parameter determined according to the operating state of the industrial waste gas purification equipment;

[0099] According to the second weight coefficient and the target individual deviation degree, a combination operation is performed to calculate the relative deviation indicator.

[0100] The individual deviation degree refers to the difference between the quantitative indicator of the first regenerative bed layer and the reference value (such as the average value, median value, or historical average value during normal operation of the corresponding quantitative indicator of the second regenerative bed layer) determined based on the statistical correlation between the quantitative indicator of the first regenerative bed layer and the corresponding quantitative indicator of any second regenerative bed layer. It can be quantified by difference, ratio, percentage deviation, or statistical distance (such as Z-score), and its purpose is to preliminarily represent the deviation of the individual regenerative bed layer in a certain quantitative indicator relative to the overall or normal state.

[0101] The original numerical value of the individual deviation degree is subjected to mathematical transformation to eliminate the incomparability between different quantitative indicators due to different dimensions, numerical ranges, or distribution characteristics. It can be achieved by standardization (such as Z-score standardization), normalization (such as Min-Max normalization), or logarithmic transformation. The purpose of numerical conversion processing is to map the individual deviation degree of different quantitative indicators to a unified numerical space, so that they have a fair weight allocation basis in subsequent combination operations.

[0102] The second weight coefficient is a numerical factor used to measure the relative importance of different quantitative indicators in calculating the relative deviation indicator. It can be a fixed value set by expert experience, a value obtained by statistical analysis of historical data, or a value dynamically calculated or adjusted according to the current operating conditions (such as waste gas flow, concentration, temperature, etc.) of the industrial waste gas purification equipment. Its purpose is to reflect the difference in the indication ability of different quantitative indicators to the overall operating state of the regenerative bed layer under different operating conditions.

[0103] The present application combines the target individual deviation degree of each quantitative indicator after numerical conversion processing with the corresponding second weight coefficient to obtain a single relative deviation indicator that can fully reflect the difference in the overall operating performance of the regenerative bed layer. It can fuse multiple dimensions of individual deviation information into a comprehensive indicator with physical meaning, which is convenient for subsequent analysis and judgment.

[0104] The relative deviation index is a value obtained through numerical conversion and combined operation, representing the overall deviation of the single regenerative bed layer in the current reversing sub-stage relative to other regenerative bed layers or the normal state. The purpose is to provide a unified and comparable value for subsequent time series analysis and fault identification.

[0105] The present application converts the individual deviation degrees of the various quantitative indicators of the first regenerative bed layer to obtain the individual deviation degrees after numerical conversion. This eliminates the differences in dimensions and numerical ranges between different quantitative indicators, making them comparable and avoiding the excessive influence of some indicators with large values on the final result. Meanwhile, the second weight coefficient is determined, which is a preset second weight coefficient or a second weight coefficient determined according to the operating state of the industrial waste gas purification equipment. This adjusts the importance of each quantitative indicator according to different operating states and more accurately reflects the real operating state of the regenerative bed layer under the current working condition. Based on this, combined operation is performed according to the determined second weight coefficient and the individual deviation degrees after numerical conversion to calculate the relative deviation index. This integrates the individual deviation degrees after numerical conversion and the second weight coefficient to obtain the final relative deviation index. This method of calculating the relative deviation index is an optimization based on calculating the relative deviation index based on the individual deviation degrees of the quantitative indicators. By introducing numerical conversion and dynamic weight, the calculated relative deviation index can more accurately reflect the real operating state difference of the regenerative bed layer. This more accurate relative deviation index can provide more reliable input for subsequent fault identification based on time series dynamic characteristics, thereby improving the accuracy and robustness of fault diagnosis and more effectively identifying the subtle data characteristics caused by early hidden faults from the numerous measurement information mixed with normal production operation disturbances and environmental influences.

[0106] In some preferred embodiments, specifically, individual deviation degrees of the three quantification indexes, i.e., heat recovery efficiency, pressure drop change rate and VOCs purification efficiency, are calculated for a certain regenerative bed. First, numerical conversion is performed on the three individual deviation degrees, for example, Z-score standardization method is adopted to convert each individual deviation degree into a deviation from the mean value of individual deviation degrees of all bed layers divided by the standard deviation, to obtain a standardized individual deviation degree. Then, the second weight coefficient is determined. This can be determined according to the current exhaust inlet flow indication value, for example, when the flow is high, a higher weight is given to the pressure drop change rate, and when the flow is low, a higher weight is given to the heat recovery efficiency. Assuming that the weights of the three indexes are 0.4, 0.3 and 0.3 respectively according to the current working condition. Finally, according to the determined second weight coefficient and the standardized individual deviation degree, a combination operation is performed, for example, a weighted average method is adopted to multiply the standardized individual deviation degree by the corresponding weight and sum up to obtain the relative deviation index of the regenerative bed in the current reversing sub-stage.

[0107] Through the above technical solution, the individual deviation degrees of multiple quantification indexes of each regenerative bed are subjected to numerical conversion processing, eliminating the dimensional and numerical range differences between different indexes, so that they are comparable. At the same time, the second weight coefficient is determined, which can dynamically adjust the importance of each quantification index according to the running state of the industrial waste gas purification equipment, more accurately reflecting the real running state of the regenerative bed under the current working condition. According to the determined second weight coefficient and the numerical conversion processed individual deviation degrees, a combination operation is performed, which can comprehensively consider the deviation of multiple indexes, to obtain a more accurate and more representative relative deviation index. This more accurate relative deviation index can more effectively identify the abnormal running state of the regenerative bed, improving the accuracy of fault diagnosis.

[0108] In some embodiments, when the dynamic characteristics of the time series satisfy the abnormal condition, the regenerative bed is identified as having a potential physical fault, wherein the abnormal condition is determined based on statistical outliers, trend mutations and correlation anomalies.

[0109] Wherein, the dynamic characteristics of the time series refer to the change law and statistical properties exhibited by the time series as it evolves over time, which can be extracted and characterized using statistical analysis, signal processing or machine learning techniques.

[0110] Wherein, the abnormal condition refers to the basis for determining that the running state of the regenerative bed deviates from the normal mode, which can be constructed based on pre-set rules, statistical models or machine learning models.

[0111] Statistical outlier refers to that the numerical distribution or statistical characteristics of the time series significantly deviate from its historical normal range or the corresponding characteristics of other reference sequences, which can be determined by Z-score, IQR, cluster analysis or distance-based anomaly detection method.

[0112] Trend mutation refers to that the long-term or short-term change direction, rate or fluctuation pattern of the time series changes significantly and relatively suddenly, which can be identified by sliding window analysis, difference analysis, regression analysis or model-based mutation point detection algorithm.

[0113] Correlation anomaly refers to that the statistical correlation, causality or synchronicity between the time series and other associated time series (such as corresponding time series of other regenerative beds in the same device) deviates significantly from the normal pattern, which can be determined by correlation coefficient analysis, Granger causality test, mutual information analysis or graph model-based correlation analysis method.

[0114] By the above technical solutions, the analysis results of the three dynamic characteristics of statistical outlier, trend mutation and correlation anomaly of the time series are comprehensively judged, avoiding the misjudgment caused by relying on a single dynamic characteristic analysis, improving the accuracy of potential physical fault recognition, and more reliably identifying the abnormality caused by internal changes of the device, thereby reducing unnecessary alarms and shutdowns and improving the stable operation level of the device.

[0115] Further, when the abnormal condition is met, the regenerative bed is identified as having a potential physical fault, wherein the step of determining the abnormal condition based on the statistical outlier, the trend mutation and the correlation anomaly includes:

[0116] In each reversing sub-stage of the periodic reversing operation, stage identification information representing the type of the current reversing sub-stage is obtained;

[0117] According to the stage identification information, statistical outlier determination parameters for performing statistical outlier analysis on the time series in the current reversing sub-stage, trend mutation determination parameters for performing trend mutation analysis, and correlation anomaly determination parameters for performing correlation anomaly analysis are determined;

[0118] The statistical outlier determination parameters are used to determine the statistical outlier of the time series, obtaining a statistical outlier determination output; the trend mutation determination parameters are used to determine the trend mutation of the time series, obtaining a trend mutation determination output; and the correlation anomaly determination parameters are used to determine the correlation anomaly of the time series, obtaining a correlation anomaly determination output;

[0119] If the statistical outlier determination output, the trend mutation determination output and the correlation anomaly determination output collectively indicate an abnormal state, it is determined that the abnormal condition is met.

[0120] wherein the phase identification information is indicative of the type of the current sub-phase of the periodic reversing operation, such as the intake sub-phase, purge sub-phase, warm-up sub-phase or exhaust sub-phase, by reading the reversing valve state signal or comparing the time stamp with a pre-set reversing timing table.

[0121] The statistical outlier determination parameter is a threshold value for defining the degree of deviation of the time series data point from the center of its statistical distribution within a certain sub-phase of the periodic reversing operation, which can be implemented by using the mean value, standard deviation, quantile calculated based on historical data or dynamic threshold determined based on machine learning model, and the purpose is to adjust the sensitivity of statistical outlier determination according to the characteristics of the current sub-phase of the reversing operation.

[0122] The trend mutation determination parameter is a threshold value for defining the magnitude or duration of the significant change in the trend of the time series data within a certain sub-phase of the periodic reversing operation, which can be implemented by using the change rate or slope threshold determined based on sliding window regression analysis, cumulative sum control chart (CUSUM) or exponential weighted moving average (EWMA) method, and the purpose is to adjust the sensitivity of trend mutation determination according to the characteristics of the current sub-phase of the reversing operation.

[0123] The correlation anomaly determination parameter is a threshold value for defining the degree of deviation of the correlation between the time series and other related time series from the normal pattern within a certain sub-phase of the periodic reversing operation, which can be implemented by using the correlation or model residual threshold determined based on correlation coefficient, mutual information, Granger causality or multivariate statistical model (such as principal component analysis PCA, independent component analysis ICA) method, and the purpose is to adjust the sensitivity of correlation anomaly determination according to the characteristics of the current sub-phase of the reversing operation.

[0124] In addition, the statistical outlier determination output is a determination result indicating whether the time series exhibits a statistical outlier state after applying the statistical outlier determination parameter. The trend mutation determination output is a determination result indicating whether the time series exhibits a trend mutation state after applying the trend mutation determination parameter. The correlation anomaly determination output is a determination result indicating whether the time series exhibits a correlation anomaly state after applying the correlation anomaly determination parameter.

[0125] The present application obtains phase identification information representing the type of the current sub-phase of the periodic reversing operation within each sub-phase of the periodic reversing operation, and determines statistical outlier determination parameters for statistical outlier analysis, trend mutation determination parameters for trend mutation analysis and correlation anomaly determination parameters for correlation anomaly analysis of the time series within the current sub-phase of the periodic reversing operation according to the phase identification information.

[0126] Due to different operating characteristics and data fluctuation patterns in different sub-phases, targeted determination parameters can more accurately capture abnormal signals specific to the phase, while avoiding misjudgment of normal fluctuations as abnormal. Subsequently, the determination parameters specific to the phase are used to determine statistical outliers, trend mutations and correlation anomalies in the time series, to obtain the corresponding determination output. Only when the statistical outlier determination output, trend mutation determination output and correlation anomaly determination output jointly indicate an abnormal state, is the abnormal condition determined to be met. This multi-dimensional, phase-adaptive joint determination mechanism can effectively improve the accuracy and reliability of fault identification, and reduce false positives and false negatives. In combination with the construction of the relative deviation index time series based on the multi-bed layer in the pre-procedure, and the extraction of statistical features, trend features and correlation features, the application further refines the abnormal determination process, making the judgment of the deviation of the behavior pattern of the regenerative bed layer from other bed layer behavior patterns more accurate and robust in different operating phases, thereby more effectively identifying potential physical faults.

[0127] In some specific embodiments, when the industrial waste gas purification device is periodically switched, the system monitors the opening and closing state of the switching valve in real time, and identifies the current sub-phase according to the change of the valve combination state, for example, when the inlet valve is opened, the exhaust valve is opened, the purge valve is closed, and the preheating valve is closed, it is identified as an inlet sub-phase. The system pre-stores a set of statistical outlier determination parameters, trend mutation determination parameters and correlation anomaly determination parameters optimized and adjusted for different switching sub-phases (inlet, purge, preheating, exhaust). For example, for the inlet sub-phase, due to the large fluctuation of VOCs concentration, the threshold for statistical outliers can be set relatively loosely; while for the purge sub-phase, due to the relatively stable temperature change, the sensitivity of the trend mutation determination can be set higher. When it is identified that the current is in the inlet sub-phase, the system loads and uses the statistical outlier determination parameters, trend mutation determination parameters and correlation anomaly determination parameters preset for the inlet sub-phase to analyze the relative deviation index time series of the regenerative bed layer. If the data points of the time series in the inlet sub-phase show deviation beyond the statistical outlier threshold of the inlet sub-phase, while its trend change exceeds the trend mutation threshold of the inlet sub-phase, and the correlation with other bed layers in the phase deviates beyond the correlation anomaly threshold of the inlet sub-phase, the statistical outlier determination output, trend mutation determination output and correlation anomaly determination output all indicate an abnormal state. At this time, the system determines that the abnormal condition of the regenerative bed layer is met, and identifies that there is a potential physical fault in the bed layer.

[0128] By the technical solution, the parameters of statistical outliers, trend mutation and correlation anomaly analysis for fault determination can be adaptively adjusted according to different sub-phases of periodic reversing operation of the industrial waste gas purification equipment. This makes the fault identification process fully consider the unique operation characteristics and data fluctuation rules of each stage, thereby effectively improving the accuracy and sensitivity of identifying potential physical faults under different working conditions and operation stages, reducing false positives and false negatives, and providing more reliable protection for stable operation of the equipment.

[0129] In some other embodiments of the present application, the step of calculating, in the same reversing sub-phase, a relative deviation index representing the performance difference of a first regenerative bed among the plurality of regenerative beds relative to a second regenerative bed different from the first regenerative bed, based on the statistical correlation between the quantitative index of the first regenerative bed and the quantitative index corresponding to the second regenerative bed, comprises:

[0130] In the same reversing sub-phase, the quantitative index of a second regenerative bed different from the first regenerative bed among the plurality of regenerative beds is obtained for the first regenerative bed, and is used as an initial reference quantitative index;

[0131] The suitability of the initial reference quantitative index as a reference is judged to obtain a suitability judgment result;

[0132] Based on the suitability judgment result, the use mode of the initial reference quantitative index is adjusted to form a final reference quantitative index. If the suitability judgment result indicates that the initial reference quantitative index of the second regenerative bed is not suitable as a reference, the influence degree of the initial reference quantitative index of the second regenerative bed on the calculation result of the relative deviation index is reduced.

[0133] For the first regenerative bed, a relative deviation index representing the performance difference of the first regenerative bed relative to the second regenerative bed is calculated based on the statistical correlation between the quantitative index of the first regenerative bed and the initial reference quantitative index.

[0134] In the same reversing sub-phase, for a first regenerative bed among the plurality of regenerative beds, first, the quantitative index of a second regenerative bed different from the first regenerative bed is obtained and used as an initial reference quantitative index. Then, the suitability of these initial reference quantitative indexes as references is judged to obtain a suitability judgment result. This judgment step is crucial as it identifies which reference data may have problems and is not suitable for direct comparison. Based on the suitability judgment result, the use mode of the initial reference quantitative index is adjusted to form a final reference quantitative index.

[0135] Specifically, if it is determined that the initial reference quantitative indicators of a certain second regenerative bed layer are not applicable as reference, the influence degree of the initial reference quantitative indicators of the second regenerative bed layer on the calculation result of the relative deviation indicator is reduced. This adjustment ensures that the reference data used for comparison is more reliable and representative.

[0136] Finally, for the first regenerative bed layer, based on the statistical correlation between its quantitative indicators and the initial reference quantitative indicators after the applicability adjustment, a relative deviation indicator representing the difference in operating performance of the first regenerative bed layer relative to the second regenerative bed layer is calculated. Since the reference data has been screened and optimized, the calculated relative deviation indicator can more accurately reflect the true operating state of the first regenerative bed layer, effectively distinguishing its own abnormality from the influence of the reference bed layer abnormality. By optimizing the selection and use of reference data, the application improves the calculation accuracy of the relative deviation indicator, thereby enhancing the accuracy and reliability of subsequent fault diagnosis based on the indicator.

[0137] The following is described by a specific example: Assume that an industrial waste gas purification device contains three regenerative bed layers, labeled as bed layer 1, bed layer 2, and bed layer 3. In a certain intake sub-stage, the relative deviation indicator of bed layer 1 needs to be calculated. At this time, bed layer 1 is the first regenerative bed layer, and bed layer 2 and bed layer 3 are the second regenerative bed layer. The quantitative indicators of bed layer 2 and bed layer 3 in this intake sub-stage, such as pressure drop change rate, are obtained as initial reference quantitative indicators. Then, the applicability of these initial reference quantitative indicators is determined. For example, it can be checked whether the pressure drop change rate of bed layer 2 is within a predetermined normal range, or whether it is significantly different from the pressure drop change rate of other bed layers (bed layer 3). If it is determined that the pressure drop change rate of bed layer 2 is abnormal, indicating that it is not applicable as reference, the influence degree of the pressure drop change rate data of bed layer 2 is reduced when calculating the relative deviation indicator of bed layer 1. This can be achieved by excluding bed layer 2 data from the reference calculation, using only the pressure drop change rate of bed layer 3 for averaging as reference, or assigning a lower weight to the data of bed layer 2. Finally, based on the pressure drop change rate of bed layer 1 itself and the adjusted reference data (e.g., the average pressure drop change rate of bed layer 3), the relative deviation indicator of the pressure drop change rate of bed layer 1 in this intake sub-stage is calculated.

[0138] Through the above technical solutions, when calculating the relative deviation indicator of the regenerative bed layer, the applicability of the reference quantitative indicators can be effectively evaluated and adjusted. This reduces the influence of non-representative or abnormal reference data on the calculation result, improving the accuracy and reliability of the relative deviation indicator. Therefore, the fault diagnosis based on the relative deviation indicator can more accurately identify the potential physical faults of the regenerative bed layer, avoid misjudgment or omission caused by abnormal reference data, and improve the accuracy of fault diagnosis.

[0139] By the above scheme, the following technical effects can be achieved: when calculating the relative deviation index of the heat storage bed layer, the applicability of the reference quantitative index can be effectively evaluated and adjusted. This reduces the influence of non-representative or abnormal reference data on the calculation results, and improves the accuracy and reliability of the relative deviation index. Therefore, the fault diagnosis based on the relative deviation index can more accurately identify the potential physical faults of the heat storage bed layer, avoid misjudgment or omission caused by abnormal reference data, and improve the accuracy of fault diagnosis.

[0140] On the other hand, as shown in Figure 2 An exemplary industrial waste gas purification equipment fault data analysis system 100 is shown. The present application provides an industrial waste gas purification equipment fault data analysis system, comprising:

[0141] The data acquisition module 10 is used to acquire the operation data of each heat storage bed layer in the periodic reversing operation of the industrial waste gas purification equipment, wherein the industrial waste gas purification equipment comprises a plurality of heat storage bed layers connected in parallel with each other, and the periodic reversing operation comprises a plurality of reversing sub-stages; the operation data is acquired in real time by a sensor group arranged in each heat storage bed layer, and the operation data includes temperature, pressure difference and VOCs concentration;

[0142] The index calculation module 20 is used to calculate the quantitative index corresponding to the operation data of each heat storage bed layer, and the quantitative index includes heat recovery efficiency, pressure drop change rate and VOCs purification efficiency;

[0143] The relative deviation calculation module 30 is used to calculate, in the same reversing sub-stage, a relative deviation index representing the difference in operation performance of a first heat storage bed layer in the plurality of heat storage bed layers relative to a second heat storage bed layer different from the first heat storage bed layer, based on the statistical correlation between the quantitative index of the first heat storage bed layer and the quantitative index corresponding to the second heat storage bed layer;

[0144] The time series tracking module 40 is used to arrange the relative deviation index of the first heat storage bed layer calculated in each reversing sub-stage in time sequence to form a time series, with the complete cycle of the periodic reversing operation as a unit;

[0145] The fault identification module 50 is used to identify that the first heat storage bed layer has a potential physical fault when the dynamic characteristics of the time series meet an abnormal condition, and the abnormal condition indicates that the behavior pattern of the first heat storage bed layer deviates from the behavior pattern of the second heat storage bed layer.

[0146] Wherein, the plurality of reversing sub-stages of the periodic reversing operation are divided by reversing valve action signals, and the plurality of reversing sub-stages include an air inlet sub-stage, a purge sub-stage, a preheating sub-stage and an exhaust sub-stage.

[0147] In some preferred embodiments, the industrial off-gas purification plant failure data analysis system can be implemented in a distributed architecture. The data acquisition module can be composed of distributed I / O modules arranged in the field, which are directly connected to the sensor groups of each regenerative bed and transmit the collected operational data to the central processing unit through an industrial network (e.g., Modbus TCP / IP protocol based on Ethernet). The central processing unit can be an industrial control computer or a high-performance programmable logic controller (PLC). The index calculation module, the relative deviation calculation module, the time series tracking module, and the failure identification module can be run as software applications on the central processing unit. These software modules can be developed in a programming language suitable for industrial data processing (e.g., Python, C++) and utilize corresponding data processing libraries and algorithm libraries to implement their functions. The collected raw operational data and the calculated quantitative indicators and relative deviation indicators can be stored in a local industrial database or a historical database for historical trend analysis and model training. The system can provide an operator workstation to display the real-time operational data of each bed, the time series graphs of the quantitative indicators and the relative deviation indicators, and the failure alarm information through human-machine interface (HMI) software, facilitating the operator to monitor the equipment status and troubleshoot failures. For example, the sensor groups can include thermocouples installed at the inlet and outlet of each bed and at different positions inside the bed for measuring temperature, pressure transmitters installed at the inlet and outlet of the bed for measuring pressure difference, and a VOCs online analyzer installed at the total inlet and outlet for measuring VOCs concentration. The abnormal condition can be set as when the relative deviation indicator of a certain bed continuously exceeds the threshold range determined based on the statistical distribution of normal beds for multiple cycles in a specific reversing sub-stage, and its time series presents a significant upward or downward trend, the system determines that the bed has a potential physical failure.

[0148] The above only describes the embodiments of the present application and is not used to limit the protection scope of the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for analyzing fault data of industrial waste gas purification equipment, characterized in that, include: During the periodic reversing operation of the industrial waste gas purification equipment, the operating data of each heat storage bed layer is acquired. The industrial waste gas purification equipment includes multiple heat storage beds connected in parallel. The periodic reversing operation includes multiple reversing sub-stages. The operating data is acquired in real time through a sensor group arranged in each heat storage bed layer. The operating data includes temperature, pressure difference and VOCs concentration. Calculate the quantitative indicators corresponding to the operating data of each heat storage bed layer. The quantitative indicators include heat recovery efficiency, pressure drop change rate and VOCs purification efficiency. Within the same commutation sub-stage, for the first heat storage bed among multiple heat storage beds, based on its quantitative index and the statistical correlation of the quantitative index corresponding to any second heat storage bed that is different from the first heat storage bed among multiple heat storage beds, a relative deviation index characterizing the difference in the operating performance of the first heat storage bed relative to the second heat storage bed is calculated. Using the complete cycle of the periodic reversing operation as a unit, the relative deviation indices calculated for each reversing sub-stage of the first thermal regenerator bed are arranged in chronological order to form a time series. When the dynamic characteristics of the time series meet the abnormal conditions, the first heat storage bed is identified as having a potential physical fault. The abnormal conditions indicate that the behavior pattern of the first heat storage bed deviates from the behavior pattern of the second heat storage bed.

2. The method for analyzing fault data of industrial waste gas purification equipment according to claim 1, characterized in that, The multiple commutation sub-stages of the periodic commutation operation are divided by the commutation valve action signal. The multiple commutation sub-stages include an intake sub-stage, a purging sub-stage, a preheating sub-stage, and an exhaust sub-stage.

3. The method for analyzing fault data of industrial waste gas purification equipment according to claim 2, characterized in that, The step of calculating the quantitative indicators corresponding to the operating data of each thermal regeneration bed layer, including heat recovery efficiency, pressure drop change rate, and VOCs purification efficiency, includes: Acquire operating condition indication data characterizing the current operating condition of the industrial waste gas purification equipment. The operating condition indication data is extracted from the operating data and includes at least the waste gas inlet flow rate indication value or the waste gas inlet component concentration indication value. Based on the operating condition indication data, a set of quantitative indicator calculation parameters is determined to characterize the current operating condition. The set of quantitative indicator calculation parameters includes a first weighting coefficient or correction factor for calculating each quantitative indicator. Based on the operating data of each heat storage bed and the set of quantitative index calculation parameters, the quantitative index corresponding to the operating data of each heat storage bed is calculated. The quantitative index includes heat recovery efficiency, pressure drop change rate and VOCs purification efficiency.

4. The method for analyzing fault data of industrial waste gas purification equipment according to claim 2, characterized in that, The step of calculating a relative deviation index characterizing the difference in operational performance between the first and second regenerative bed layers within the same commutation sub-stage, based on the statistical correlation between the first regenerative bed layer's quantitative index and the quantitative index corresponding to any second regenerative bed layer different from the first regenerative bed layer, includes: For each quantitative index of the first heat storage bed, based on the quantitative index of the first heat storage bed and the corresponding quantitative index of the second heat storage bed, the individual deviation degree of the quantitative index of the first heat storage bed is determined. Based on the individual deviations of various quantitative indicators of the first thermal regeneration bed, a relative deviation index characterizing the difference in operational performance between the first thermal regeneration bed and the second thermal regeneration bed is calculated.

5. The method for analyzing fault data of industrial waste gas purification equipment according to claim 4, characterized in that, The step of calculating the relative deviation index, which characterizes the difference in operational performance between the first thermal regeneration bed layer and the second thermal regeneration bed layer, based on the individual deviation degree of various quantitative indicators of the first thermal regeneration bed layer includes: The individual deviation degree of various quantitative indicators of the first heat storage bed is numerically converted to obtain the target individual deviation degree. A second weighting coefficient is determined, which is a preset parameter determined based on the operating status of the industrial waste gas purification equipment; The relative deviation index is calculated by combining the second weighting coefficient and the deviation degree of the target individual.

6. The method for analyzing fault data of industrial waste gas purification equipment according to claim 2, characterized in that, The step of identifying the first thermal storage bed as having a potential physical fault when the dynamic characteristics of the time series meet an abnormal condition, wherein the abnormal condition indicates that the behavior pattern of the first thermal storage bed deviates from the behavior pattern of the second thermal storage bed, includes: Dynamic characteristic analysis is performed on the time series to extract statistical features, trend features, and correlation features; When abnormal conditions are met, the first heat storage bed is identified as having a potential physical fault, wherein the abnormal conditions are determined based on a combination of statistical outliers, trend mutations, and correlation anomalies.

7. The method for analyzing fault data of industrial waste gas purification equipment according to claim 6, characterized in that, When an abnormal condition is met, the step of identifying the first thermal regeneration bed as having a potential physical fault includes: the abnormal condition being determined based on a combination of statistical outliers, trend mutations, and correlation anomalies. Within each commutation sub-stage of the periodic commutation operation, stage identification information characterizing the type of the current commutation sub-stage is obtained; Based on the stage identification information, statistical outlier determination parameters, trend change determination parameters, and association anomaly determination parameters are determined for statistical outlier analysis of the time series in the current commutation sub-stage, respectively. The statistical outlier determination parameter is used to determine the statistical outlier situation of the time series, and a statistical outlier determination output is obtained; the trend change determination parameter is used to determine the trend change situation of the time series, and a trend change determination output is obtained; and the correlation anomaly determination parameter is used to determine the correlation anomaly situation of the time series, and a correlation anomaly determination output is obtained. If the statistical outlier determination output, the trend mutation determination output, and the correlation anomaly determination output all indicate an abnormal state, then the abnormal condition is determined to be met.

8. The method for analyzing fault data of industrial waste gas purification equipment according to claim 2, characterized in that, The step of calculating a relative deviation index characterizing the difference in operational performance between the first and second regenerative bed layers within the same commutation sub-stage, based on the statistical correlation between the first regenerative bed layer's quantitative index and the quantitative index corresponding to any second regenerative bed layer different from the first regenerative bed layer, includes: Within the same commutation sub-stage, for the first of multiple regenerable bed layers, the quantitative index of the second regenerable bed layer is obtained and used as the initial reference quantitative index. The applicability of the initial reference quantification index as a reference is determined, and the applicability judgment result is obtained; Based on the applicability judgment result, the usage of the initial reference quantification index is adjusted to form the final reference quantification index. If the applicability judgment result indicates that the initial reference quantification index of the second thermal regeneration bed is not applicable as a reference, the influence of the initial reference quantification index of the second thermal regeneration bed on the calculation result of the relative deviation index is reduced. For the first thermal regeneration bed, based on the statistical correlation between its quantitative index and the initial reference quantitative index, a relative deviation index characterizing the difference in operating performance between the first thermal regeneration bed and the second thermal regeneration bed is calculated.

9. A fault data analysis system for industrial waste gas purification equipment, characterized in that, include: The data acquisition module is used to acquire the operating data of each heat storage bed layer during the periodic reversing operation of the industrial waste gas purification equipment. The industrial waste gas purification equipment includes multiple heat storage beds connected in parallel, and the periodic reversing operation includes multiple reversing sub-stages. The operating data is acquired in real time through a sensor group arranged in each heat storage bed layer. The operating data includes temperature, pressure difference, and VOCs concentration. The index calculation module is used to calculate the quantitative indexes corresponding to the operating data of each heat storage bed layer. The quantitative indexes include heat recovery efficiency, pressure drop change rate and VOCs purification efficiency. The relative deviation calculation module is used to calculate, within the same commutation sub-stage, a relative deviation index characterizing the difference in operating performance between the first thermal regeneration bed and the second thermal regeneration bed, based on the statistical correlation between the first thermal regeneration bed among multiple thermal regeneration bed layers and the quantitative index corresponding to any second thermal regeneration bed layer that is different from the first thermal regeneration bed layer. The time series tracking module is used to arrange the relative deviation indicators calculated by the first heat storage bed in each commutation sub-stage in chronological order, taking the complete cycle of the periodic commutation operation as the unit, to form a time series. The fault identification module is used to identify the first heat storage bed as having a potential physical fault when the dynamic characteristics of the time series meet abnormal conditions, wherein the abnormal conditions indicate that the behavior pattern of the first heat storage bed deviates from the behavior pattern of the second heat storage bed.

10. The industrial waste gas purification equipment fault data analysis system according to claim 9, characterized in that, The multiple commutation sub-stages of the periodic commutation operation are divided by the commutation valve action signal. The multiple commutation sub-stages include an intake sub-stage, a purging sub-stage, a preheating sub-stage, and an exhaust sub-stage.

Images