A method and apparatus for assessing stability of a combustion process

Abstract

The application discloses a kind of stability evaluation method and device of combustion process, it is related to combustion safety monitoring technical field.Extract the flame static characteristic and flame dynamic characteristic of each time respectively;Flame static characteristic and flame dynamic characteristic, and state data and flue gas emission monitoring data are all regarded as characteristics, for any kind of characteristics, calculate the Z score of corresponding characteristics under each time, linear instability index and threshold instability index;According to the Z score of each kind of characteristics under each time, linear instability index and threshold instability index, determine the comprehensive instability index of each kind of characteristics under each time;According to the comprehensive instability index of each kind of characteristics under each time, the data of all times are clustered, and stability clustering center is obtained;According to the distance between the data of each time and stability clustering center, determine the stability evaluation result of the combustion process of incinerator under each time.The method can accurately evaluate the stability of combustion process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of combustion safety monitoring, and in particular to a combustion process stability evaluation method and device. BACKGROUND

[0002] Incinerators are widely used for the treatment of sludge, garbage and industrial waste, and their stable operation is crucial for pollutant emission control and energy utilization efficiency. Existing incinerator control systems mainly rely on Distributed Control System (DCS) for parameter adjustment, and Continuous Emission Monitoring System (CEMS) for emission index monitoring. Although such systems can collect real-time data such as temperature, oxygen content, and fuel flow, they lack in-depth analysis of the combustion state and cannot effectively identify unstable factors in the combustion process.

[0003] Currently, some research attempts to analyze the combustion state using flame images, but most of them are based on static flame features (such as brightness and area) for judgment, ignoring the influence of dynamic flame change process on combustion process stability.

[0004] Therefore, there is a lack of a method that can accurately evaluate the stability of the combustion process in the prior art. SUMMARY

[0005] Therefore, it is necessary to provide a combustion process stability evaluation method and device that can accurately evaluate the stability of the combustion process.

[0006] The present application adopts the following technical solutions:

[0007] The present application provides a combustion process stability evaluation method, comprising:

[0008] Obtaining data collected at multiple consecutive time points, each time point data including flame images of the incinerator in the combustion process, as well as state data and flue gas emission monitoring data;

[0009] According to the flame images collected at each time point, the flame static features and flame dynamic features at each time point are extracted respectively;

[0010] The flame static features and flame dynamic features, as well as the state data and flue gas emission monitoring data, are used as features, and for any feature, the Z-score, linear instability index and threshold instability index of the corresponding feature at each time point are calculated;

[0011] According to the Z-score, linear instability index and threshold instability index of each feature at each time point, the comprehensive instability index of each feature at each time point is determined;

[0012] According to the comprehensive instability index of each feature at each time, the data of all times is clustered to obtain a stability clustering center;

[0013] According to the distance between the data of each time and the stability clustering center, the stability evaluation result of the combustion process of the incinerator at each time is determined.

[0014] Optionally, the flame static features include average brightness and brightness standard deviation; the flame dynamic features include average pulsation frequency, pulsation area rate, average speed and average direction; according to the flame images collected at each time, the flame static features and the flame dynamic features of each time are extracted respectively, including:

[0015] For the flame image collected at any time, the average brightness and the brightness standard deviation of the flame image are calculated;

[0016] According to the flame image and the preset frame flame image collected before the flame image, the average pulsation frequency of the flame image is calculated;

[0017] According to the current flame image and the flame images of the adjacent two frames before and after the current flame image, the pulsation area rate of the flame image is calculated;

[0018] According to the optical flow method, the speed of each pixel point in the flame image is calculated, and the average value of the speeds of all pixel points in the flame image is taken as the average speed of the flame image;

[0019] The direction of the speed of each pixel point in the flame image is calculated, and the average value of the directions of the speeds of all pixel points in the flame image is taken as the average direction of the flame image.

[0020] Optionally, the linear instability index of each feature at each time is calculated, including:

[0021] For any feature, the feature sequence of the current time is obtained according to the sliding window method;

[0022] The feature sequence is linearly fitted by linear regression to determine the slope of linear fitting and the goodness of fit;

[0023] The absolute value of the product of the slope of linear fitting and the goodness of fit is determined as the linear instability index of the feature.

[0024] Optionally, the threshold instability index of each feature at each time is calculated, including:

[0025] For any feature, when the value of the feature is greater than or equal to the lower limit of the threshold and less than or equal to the upper limit of the threshold, the threshold instability index of the feature is determined as 0;

[0026] In a case where the value corresponding to the feature is greater than the upper limit of the feature threshold value or less than the lower limit of the feature threshold value, a threshold instability index of the feature is determined according to the value of the feature, the upper limit of the feature threshold value, and the lower limit of the feature threshold value.

[0027] Optionally, in a case where the value corresponding to the feature is greater than the upper limit of the feature threshold value, the calculation formula of the threshold instability index of the feature is:

[0028] ;

[0029] wherein, represents the threshold instability index, represents the value of the feature, represents the upper limit of the feature threshold value, represents the lower limit of the feature threshold value.

[0030] In a case where the value corresponding to the feature is less than the lower limit of the feature threshold value, the calculation formula of the threshold instability index of the feature is:

[0031] .

[0032] Optionally, a comprehensive instability index of each feature at each time point is determined according to the Z-score, the linear instability index, and the threshold instability index of each feature at each time point, including:

[0033] For any feature at each time point, a result of adding the Z-score, the linear instability index, and the threshold instability index of the feature is determined as the comprehensive instability index of the feature.

[0034] Optionally, a stability clustering center is obtained by clustering data of all time points according to the comprehensive instability index of each feature at each time point, including:

[0035] According to the comprehensive instability index of each feature at each time point, data of all time points are clustered by using a plurality of different clustering algorithms to obtain clustering results of each clustering algorithm; the clustering results include stability data clusters and instability data clusters in a two-dimensional plane.

[0036] A best clustering algorithm is determined according to the clustering results of each clustering algorithm.

[0037] A stability clustering center is determined as the clustering center of the stability data cluster of the best clustering algorithm.

[0038] Optionally, a stability evaluation result of the combustion process of the incinerator at each time point is determined according to the distance between the data at each time point and the stability clustering center, including:

[0039] determine the incineration system stability index of each moment according to the distance between the data of each moment and the stability clustering center and a preset distance factor;

[0040] For any moment, if the incineration system stability index is within a first preset range, the stability evaluation result is determined to be that the combustion process of the incinerator is stable.

[0041] If the incineration system stability index is within a second preset range, the stability evaluation result is determined to be that the combustion process of the incinerator is relatively stable.

[0042] If the incineration system stability index is within a third preset range, the stability evaluation result is determined to be that the combustion process of the incinerator is unstable. All values within the first preset range are greater than all values within the second preset range, and all values within the second preset range are greater than all values within the third preset range.

[0043] Optionally, the calculation formula of the incineration system stability index is:

[0044]

[0045] wherein, represents the incineration system stability index, represents the distance between the data point and the stability clustering center, represents the distance factor.

[0046] The present application provides a stability evaluation device for a combustion process, comprising:

[0047] an acquisition module for acquiring data collected at a plurality of consecutive moments, the data of each moment comprising a flame image of the incinerator in the combustion process, as well as state data and flue gas emission monitoring data;

[0048] an extraction module for extracting flame static features and flame dynamic features of each moment respectively according to the flame image collected at each moment;

[0049] a calculation module for taking the flame static features and the flame dynamic features, as well as the state data and the flue gas emission monitoring data, as features, and calculating the Z-score, the linear instability index and the threshold instability index of the corresponding features at each moment for each kind of feature;

[0050] a comprehensive module for determining the comprehensive instability index of each kind of feature at each moment according to the Z-score, the linear instability index and the threshold instability index of each kind of feature at each moment;

[0051] a clustering module for clustering the data of all moments according to the comprehensive instability index of each kind of feature at each moment to obtain a stability clustering center; ​

[0052] A determination module is configured to determine the stability evaluation result of the combustion process of the incinerator according to the distance between the data at each time point and the stability clustering center.

[0053] The present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the stability evaluation method of the combustion process.

[0054] The present application provides a computer device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the stability evaluation method of the combustion process when executing the program.

[0055] The above-mentioned at least one technical solution adopted by the present application can achieve the following beneficial effects:

[0056] In the present application, the state data, the flue gas emission monitoring data, the flame static features and the flame dynamic features extracted from the flame image are taken as features, and then the instability indexes of the features are calculated, including the Z-score, the linear instability index and the threshold instability index, which are equivalent to defining the mutation instability, the trend instability and the threshold instability, and then based on the three indexes, the comprehensive instability indexes of the corresponding features are obtained, the clustering analysis is performed on the comprehensive instability indexes of the features calculated, and the stability clustering center is accurately found, and on this basis, the stability evaluation result of the combustion process of the incinerator is determined based on the distance from the data to the stability clustering center, so that the accuracy of the stability evaluation of the combustion process is improved. BRIEF DESCRIPTION OF DRAWINGS

[0057] The accompanying drawings, which are included to provide a further understanding of the present application and constitute a part of this application, illustrate certain illustrative embodiments of the present application and together with the description serve to explain the present application. In the drawings:

[0058] Figure 1 A stability evaluation method of a combustion process provided by the present application is shown in the flowchart;

[0059] Figure 2 The clustering results of four clustering algorithms provided by the present application in a two-dimensional plane are shown in the diagrams, wherein, (a) is a diagram of the clustering result of KMeans clustering in a two-dimensional plane, (b) is a diagram of the clustering result of hierarchical clustering in a two-dimensional plane, (c) is a diagram of the clustering result of Gaussian mixture model in a two-dimensional plane, and (d) is a diagram of the clustering result of DBSCAN model in a two-dimensional plane;

[0060] Figure 3 A computer device for implementing the stability evaluation method of the combustion process provided by the present application is shown in the diagram. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0062] Existing methods mostly use fixed thresholds or empirical rules to identify combustion anomalies, lacking a unified standard for quantifying stability and making it difficult to adapt to complex changes in operating conditions.

[0063] Therefore, there is an urgent need for a stability assessment method based on the combustion process, which combines DCS (state data), CEMS (emission monitoring data), and flame image data to fully consider the dynamic changes in the combustion state, so as to achieve accurate quantification and real-time monitoring of the stability of the incineration system, thereby improving the operational stability and emission control level of the incinerator.

[0064] The stability assessment method for combustion processes provided in this invention can be implemented by a server located on a business platform, or by a device such as a desktop computer or laptop computer capable of executing the solution of this invention. The technical solutions provided by various embodiments of this invention are described in detail below with reference to the accompanying drawings.

[0065] Figure 1 This is a schematic diagram of a stability assessment method for a combustion process according to the present invention, which specifically includes the following steps:

[0066] S101, acquire data collected at multiple consecutive time points. The data at each time point includes flame images of the incinerator during the combustion process, as well as status data and flue gas emission monitoring data.

[0067] At the incinerator site, an industrial camera is fixed on the top of the incinerator to continuously capture a certain number of flame images. The images are clearly visible, and the data acquisition interval can be 1 minute. Specifically, the number of images acquired can be determined according to actual needs.

[0068] The status data is DCS data, and the flue gas emission monitoring data is CEMS data. The status data includes fuel feed rate, primary air volume, secondary air volume, auxiliary fuel flow rate, denitrification agent flow rate, desulfurization agent flow rate, oxygen content in the furnace, temperature of the dilute phase zone, and temperature of the dense phase zone. The flue gas emission monitoring data includes flue gas temperature, NOx emission concentration, SO2 emission concentration, CO emission concentration, and HCl emission concentration.

[0069] During the combustion process in the incinerator, the DCS and CEMS collect status data and flue gas emission monitoring data in real time. Therefore, after the flame image is collected, the status data and flue gas emission monitoring data at the corresponding time can be obtained from the database.

[0070] S102, based on the flame images acquired at each time moment, extract the static and dynamic features of the flame at each time moment.

[0071] Optionally, the static features of the flame include average brightness and brightness standard deviation; the dynamic features of the flame include average pulsation frequency, pulsation area ratio, average velocity, and average direction. Based on the flame image acquired at each moment, the static and dynamic features of the flame at each moment are extracted, including: for the flame image acquired at any moment, calculating the average brightness and brightness standard deviation of the flame image; calculating the average pulsation frequency of the flame image based on the flame image and the preset frame flame images acquired before this flame image; calculating the pulsation area ratio of the flame image based on the current flame image and the flame images of the two adjacent frames; calculating the velocity of each pixel in the flame image using the optical flow method, and taking the average of the velocities of all pixels in the flame image as the average velocity of the flame image; calculating the velocity direction of each pixel in the flame image, and taking the average of the velocity directions of all pixels in the flame image as the average direction of the flame image.

[0072] First, before acquiring the static and dynamic features of the flame image, the flame image is first processed by Gaussian filtering with a kernel size of 5. The filtered flame image is then converted into an HSV image, which corresponds to hue, saturation, and value.

[0073] Flame features are extracted from the V channel of the HSV image. The flame features include static flame features and dynamic flame features. That is, the static flame features and dynamic flame features extracted from the flame image are both extracted from the V channel of the corresponding HSV image.

[0074] (1) Average brightness

[0075] Extract from HSV images V For each channel component, calculate its average brightness using the following formula:

[0076] (1);

[0077] in, Average brightness The first flame region in the flame image The brightness of each pixel This represents the number of pixels in the flame region of the flame image.

[0078] (2) Standard deviation of luminance

[0079] The standard deviation of luminance can measure the dispersion degree of luminance, and to some extent reflect the non-uniformity of flame temperature. The calculation formula of the standard deviation of luminance is as follows:

[0080] (2);

[0081] Wherein, is the standard deviation of luminance.

[0082] (3) Average pulsation frequency

[0083] By extracting all the images within 5 seconds before the current time, the average luminance sequence is calculated. The average luminance sequence is processed by discrete Fourier transform, and the direct current component is removed. The power spectrum average pulsation frequency is analyzed and processed, and the calculation formula is as follows:

[0084] (3);

[0085] In the formula, is the average pulsation frequency; N represents the total number of frequency components in the discrete Fourier transform; is the first frequency component in the discrete Fourier transform, with the unit of Hz; is the square of the amplitude component of the frequency domain curve.

[0086] (4) Pulsation area rate

[0087] The pulsation area rate is calculated based on the three-frame difference method to reduce noise interference. Specifically, the V channel data of the adjacent 3 images (i.e. the current flame image and the previous and next two flame images) are extracted, the first two frames and the last two frames are respectively differentiated, the two times of difference results are logically AND operated, and finally the V value pulsation data is obtained, which reflects the change of pixel brightness. The pixels with pulsation size exceeding 3 are defined as significant pulsation pixels, and the ratio of the number of significant pulsation pixels to the number of pixels in the flame area is the pulsation area rate. The calculation formula is as follows:

[0088] (4);

[0089] (5);

[0090] In the formula, is the V channel data of the first frame, is the V channel data of the first -2 frame, is the V channel data of the first -1 frame, Vp is the pulsation data, H is the step function, Vth is the pulsation threshold, N is the number of pixels in the flame region, A is the pulsation area ratio, Nrow represents the number of pixels in the horizontal direction of the image, Ncol represents the number of pixels in the vertical direction of the image, (x, y) represents the pixel coordinates, Vp is the calculated , ∩ represents the intersection operation, and denotes the logical AND.

[0091] (5) Average velocity

[0092] The average velocity is calculated based on the Lucas-Kanada optical flow method. The gradient equation can be obtained from the L-K optical flow method:

[0093] (6);

[0094] In the formula, and represent the velocities in the x and y directions, and directions, , and represent the partial derivatives of the luminance in the x, y, and z directions, respectively. , and

[0095] Taking a 10x10 window, each point in the window is considered to have the same moving direction. Applying all the points to the gradient equation gives:

[0096] (7).

[0097] Based on the least squares method, formula (7) can be solved to obtain , and the velocity calculation formula is:

[0098] (8).

[0099] Finally, the average velocity of all pixel points in the flame region is calculated by taking the average value, which is the average velocity of the flame image at the current time. Similarly, the average direction is calculated based on the equidistant sampling method.

[0100] (6) Average direction

[0101] The velocity direction calculation formula of a single pixel point is as follows:

[0102] (9).

[0103] ​The average of the velocity directions of all the pixel points in the flame region is calculated to obtain the average direction at the current moment.

[0104] S103, the flame static characteristics and the flame dynamic characteristics, and the state data and the flue gas emission monitoring data are all taken as features, and for any kind of feature, the Z-score, the linear instability index and the threshold instability index of the corresponding feature at each moment are calculated.

[0105] Instability index extraction: the feature data set includes the aforementioned flame static characteristics and flame dynamic characteristics, and the state data and the flue gas emission monitoring data, and the stability is defined as the change of the parameter relative to the historical data. The parameter stability means that the parameter fluctuates in a small range over time. The instability includes: 1, the parameter increases or decreases sharply, which deviates from the historical data range obviously; 2, the parameter continuously increases or decreases, which has a clear linear trend; 3, the parameter exceeds the specified threshold, which exceeds the upper threshold or lower threshold based on experience. Three instability indexes are defined for the three cases: Z-score, linear instability index and threshold instability index.

[0106] Z-score: measures the distance of a data point from the average value of the data set, which is in standard deviation units, and the calculation formula is:

[0107] (10);

[0108] Among them, according to the sliding window method, the feature set of the current moment feature is obtained, for example, the sliding window is 1 hour 60 data points of features, and the feature set includes the features of the previous 1 hour from the current moment; at this time, Z-score of the current moment feature, the value of the current moment feature, the average value of the feature set, the standard deviation of the feature set.

[0109] In one embodiment, the linear instability index of the corresponding feature at each moment is calculated, including: for any kind of feature, according to the sliding window method, the feature sequence of the current moment is obtained; the linear fitting of the feature sequence is determined by linear regression, and the slope and goodness of fit of the linear fitting are determined; the absolute value of the product of the slope and the goodness of fit of the linear fitting is determined as the linear instability index of the feature.

[0110] The way of obtaining the feature sequence of the current moment in this embodiment is the same as the way of obtaining the feature set in the above embodiment, which will not be described here.

[0111] Linear instability index The calculation formula is:

[0112] (11);

[0113] wherein, is the slope of the linear fit, is the goodness of fit of the linear fit.

[0114] In one embodiment, the threshold instability index of each feature at each time point is calculated, including: for any feature, if the value corresponding to the feature is greater than or equal to the lower threshold of the feature and less than or equal to the upper threshold of the feature, determining the threshold instability index of the feature as 0; if the value corresponding to the feature is greater than the upper threshold of the feature or less than the lower threshold of the feature, determining the threshold instability index of the feature according to the value of the feature, the upper threshold of the feature and the lower threshold of the feature. Optionally, in the case where the value corresponding to the feature is greater than the upper threshold of the feature, the calculation formula of the threshold instability index of the feature is:

[0115] (12);

[0116] wherein, denotes the threshold instability index, denotes the value of the feature, denotes the upper threshold of the feature, denotes the lower threshold of the feature.

[0117] In the case where the value corresponding to the feature is less than the lower threshold of the feature, the calculation formula of the threshold instability index of the feature is:

[0118] (13).

[0119] S104, determining the comprehensive instability index of each feature at each time point according to the Z-score, the linear instability index and the threshold instability index of each feature at each time point.

[0120] Optionally, the comprehensive instability index of each feature at each time point is determined according to the Z-score, the linear instability index and the threshold instability index of each feature at each time point, including: for any feature at each time point, adding the Z-score, the linear instability index and the threshold instability index of the feature to determine the comprehensive instability index of the feature.

[0121] The comprehensive instability index of the feature measures the instability degree of the current data relative to the historical data.

[0122] S105, clustering the data of all time points according to the comprehensive instability index of each feature at each time point to obtain a stability clustering center.

[0123] Optionally, based on the comprehensive instability index of each feature at each time point, the data at all times are clustered to obtain stable cluster centers. This includes: clustering the data at all times using multiple different clustering algorithms based on the comprehensive instability index of each feature at each time point, obtaining the clustering results of each clustering algorithm; the clustering results include stable data clusters and unstable data clusters in a two-dimensional plane; determining the optimal clustering algorithm based on the clustering results of each clustering algorithm; and determining the cluster centers of the stable data clusters of the optimal clustering algorithm as stable cluster centers.

[0124] Specifically, four clustering algorithms can be used: K-means clustering, hierarchical clustering, Gaussian mixture model, and density-based spatial clustering of applications with noise (DBSCAN) model. These algorithms are used to perform cluster analysis to obtain stable and unstable data clusters. Then, principal component analysis (PCA) is used to reduce the overall instability index of multiple features at each time step to two dimensions. For example, if there are N features at a time step, the overall instability index is reduced from N dimensions to two dimensions. The clustering results of each clustering algorithm are displayed on a two-dimensional plane. The clustering effects are compared based on the distribution of stable and unstable data clusters to determine the optimal clustering algorithm.

[0125] Specifically, the clustering results are compared based on the distribution of stable and unstable data clusters to determine the best clustering algorithm. This includes: calculating the silhouette coefficient of the entire dataset corresponding to each clustering algorithm based on the stable and unstable data clusters, and determining the clustering algorithm with the largest silhouette coefficient as the best clustering algorithm.

[0126] like Figure 2 As shown, Figure 2 Figure 1 shows the clustering results of four clustering algorithms in a two-dimensional plane. (a) Figure 2 shows the clustering results of KMeans clustering in a two-dimensional plane, (b) Figure 3 shows the clustering results of hierarchical clustering in a two-dimensional plane, (c) Figure 4 shows the clustering results of Gaussian mixture model in a two-dimensional plane, and (d) Figure 5 shows the clustering results of DBSCAN model in a two-dimensional plane.

[0127] S106. Based on the distance between the data at each time point and the stability cluster center, determine the stability assessment result of the combustion process of the incinerator at each time point.

[0128] Optionally, the stability evaluation result of the combustion process of the incinerator at each time point is determined according to the distance between the data at each time point and the stability clustering center, including: determining the incineration system stability index at each time point according to the distance between the data at each time point and the stability clustering center and a preset distance factor; for any time point, if the incineration system stability index is within a first preset range, it is determined that the stability evaluation result is that the combustion process of the incinerator is stable; if the incineration system stability index is within a second preset range, it is determined that the stability evaluation result is that the combustion process of the incinerator is relatively stable; if the incineration system stability index is within a third preset range, it is determined that the stability evaluation result is that the combustion process of the incinerator is unstable; all values within the first preset range are greater than all values within the second preset range, and all values within the second preset range are greater than all values within the third preset range.

[0129] Optionally, the incineration system stability index is defined based on a Sigmoid function, and a calculation formula of the incineration system stability index is:

[0130] (14);

[0131] wherein, represents the incineration system stability index, represents the distance between the data point and the stability clustering center, represents the distance factor. For any time point, the data point here is a two-dimensional coordinate point after the comprehensive instability index of each feature at the corresponding time point is reduced in dimension after clustering, and the stability clustering center is also a two-dimensional coordinate point at this time.

[0132] The first preset range can be , the second preset range can be , and the third preset range can be .

[0133] The incineration system stability index is between , the combustion is stable, and the parameters normally fluctuate; the incineration system stability index is between , the combustion is relatively stable, and the parameters fluctuate in a small range; and the incineration system stability index is between , the combustion is unstable, and some parameters may sharply fluctuate or exceed the normal working condition range.

[0134] The stability index of the incineration system constructed by the present application can evaluate the overall stability of the incineration process in real time, and provide guidance for combustion optimization and pollutant control. When the index abnormally decreases, it can be used as a warning signal to indicate that the incineration system may have problems such as abnormal pollutant emission or deviation of the temperature in the furnace. At the same time, the present application further provides the instability index of each feature parameter, which helps to identify the source of the abnormality, and can be combined with process requirements and historical data to optimize and adjust these parameters. For example, when the stability index of the incineration system is abnormal, if the instability index of the concentration of NOx or CO is large, the combustion temperature or denitration strategy can be optimized.

[0135] When the stability evaluation method of the combustion process provided by the present application is applied, the order of execution of each step shown in the above formula (1) can not be executed according to the order of execution of each step shown in the above formula (1), and the order of execution of each step can be determined according to the needs, and the present application does not limit this. Figure 1

[0136] The above is the stability evaluation method of the combustion process provided by one or more embodiments of the present application. Based on the same idea, the present application also provides a corresponding stability evaluation device of the combustion process, which comprises:

[0137] The acquisition module is configured to acquire data collected at a plurality of continuous time points, and each time point data comprises a flame image of the incinerator in the combustion process, and state data and flue gas emission monitoring data.

[0138] The extraction module is configured to extract the flame static feature and the flame dynamic feature of each time point from the flame image collected at each time point, respectively.

[0139] The calculation module is configured to take the flame static feature and the flame dynamic feature, and the state data and the flue gas emission monitoring data as features, and calculate the Z-score, the linear instability index and the threshold instability index of the corresponding feature at each time point for any kind of feature.

[0140] The comprehensive module is configured to determine the comprehensive instability index of each feature at each time point according to the Z-score, the linear instability index and the threshold instability index of each feature at each time point.

[0141] The clustering module is configured to cluster the data of all time points according to the comprehensive instability index of each feature at each time point, and obtain a stability clustering center.

[0142] The determination module is configured to determine the stability evaluation result of the combustion process of the incinerator at each time point according to the distance between the data at each time point and the stability clustering center.

[0143] ​The specific definition of the combustion process stability evaluation device can refer to the definition of the combustion process stability evaluation method in the above, which will not be repeated here. Each module in the above combustion process stability evaluation device can be realized by software, hardware and their combination in whole or in part. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so that the processor calls to execute the operation corresponding to each module.

[0144] The application further provides a computer readable storage medium, which stores a computer program, and the computer program can be used to execute the above Figure 1 The application provides a combustion process stability evaluation method.

[0145] The application further provides a computer readable storage medium, which stores a computer program, and the computer program can be used to execute the above Figure 3 The structure diagram of the computer device is shown in the figure, and the computer device includes a processor, an internal bus, a network interface, a memory and a nonvolatile memory. Figure 3 The computer device includes a processor, an internal bus, a network interface, a memory and a nonvolatile memory, and of course can further include other hardware required by the business. The processor reads the corresponding computer program from the nonvolatile memory into the memory and then runs to realize the above Figure 1 The application provides a combustion process stability evaluation method.

[0146] Those skilled in the art can understand that all or part of the above-mentioned embodiment methods can be completed by the computer program to instruct the related hardware, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, the computer program can include the flow of the above-mentioned embodiment method. In the embodiments of the present application, any reference to the memory, storage, database or other medium can include at least one of the non-volatile and volatile memory. The non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory. The volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, the RAM can be various forms such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0147] The technical features of the above embodiments can be combined in any manner. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described, however, as long as the combinations of the technical features do not contradict each other, they should be considered to be within the scope of the present application.

Claims

1. A method for stability assessment of a combustion process, characterized by, The method comprises the following steps: acquiring data collected at a plurality of continuous time points, the data at each time point comprising a flame image of the incinerator in a combustion process, and state data and flue gas emission monitoring data; extracting flame static features and flame dynamic features at each time point respectively according to the flame image collected at each time point; taking the flame static features, the flame dynamic features, the state data and the flue gas emission monitoring data as features, and calculating a Z-score, a linear instability index and a threshold instability index of the corresponding feature at each time point for any kind of feature; adding the Z-score, the linear instability index and the threshold instability index of the feature at each time point for any kind of feature to determine a comprehensive instability index of the feature; performing clustering on the data at all time points according to the comprehensive instability index of each feature at each time point to obtain a stability clustering center; determining a stability evaluation result of the combustion process of the incinerator at each time point according to the distance between the data at each time point and the stability clustering center; calculating the linear instability index of the corresponding feature at each time point, comprising: for any kind of feature, acquiring a feature sequence at the current time point according to a sliding window; performing linear fitting on the feature sequence by linear regression to determine a slope of linear fitting and a goodness of fit; and determining the absolute value of the product of the slope of linear fitting and the goodness of fit as the linear instability index of the feature; calculating the threshold instability index of the corresponding feature at each time point, comprising: for any kind of feature, determining the threshold instability index of the feature as 0 in the case that the value of the feature is greater than or equal to a lower threshold and less than or equal to an upper threshold; and determining the threshold instability index of the feature according to the value of the feature, the upper threshold and the lower threshold in the case that the value of the feature is greater than the upper threshold or less than the lower threshold.

2. The method of claim 1, wherein, The flame static features comprise average brightness and brightness standard deviation; and the flame dynamic features comprise average pulsation frequency, pulsation area rate, average speed and average direction. extracting the flame static features and the flame dynamic features at each time point according to the flame image collected at each time point, comprising: calculating the average brightness and the brightness standard deviation of the flame image for any flame image collected at any time point; calculating the average pulsation frequency of the flame image according to the flame image and a preset frame of flame image collected before the flame image; calculating the pulsation area rate of the flame image according to the current flame image and the flame images of the two adjacent frames before and after the current flame image; calculating the speed of each pixel point in the flame image by the optical flow method, and taking the average value of the speeds of all pixel points in the flame image as the average speed of the flame image; calculating the speed direction of each pixel point in the flame image, and taking the average value of the speed directions of all pixel points in the flame image as the average direction of the flame image.

3. The method of claim 1, wherein, In the case that the value of the feature is greater than the upper threshold of the feature, the calculation formula of the threshold instability index of the feature is: ； wherein, represents a threshold instability index, represents a value of a feature, represents an upper threshold value of a feature, represents a lower threshold value of a feature; In the case that the value of the feature is less than the lower threshold of the feature, the calculation formula of the threshold instability index of the feature is: 。 4. The method of claim 1, wherein, According to the comprehensive instability index of each feature at each time, the data of all times is clustered to obtain a stability clustering center, including: According to the comprehensive instability index of each feature at each time, the data of all times is clustered by a plurality of different clustering algorithms to obtain a clustering result of each clustering algorithm; the clustering result includes stable data clusters and unstable data clusters in a two-dimensional plane; According to the clustering result of each clustering algorithm, a best clustering algorithm is determined; The clustering center of the stable data cluster of the best clustering algorithm is determined as the stability clustering center.

5. The method of claim 4, wherein, According to the distance between the data of each time and the stability clustering center, a stability evaluation result of the combustion process of the incinerator at each time is determined, including: According to the distance between the data of each time and the stability clustering center, and a preset distance factor, an incineration system stability index of each time is determined; For any time, if the incineration system stability index is within a first preset range, it is determined that the stability evaluation result is that the combustion process of the incinerator is stable; If the incineration system stability index is within a second preset range, it is determined that the stability evaluation result is that the combustion process of the incinerator is relatively stable; If the incineration system stability index is within a third preset range, it is determined that the stability evaluation result is that the combustion process of the incinerator is unstable; all values within the first preset range are greater than all values within the second preset range, and all values within the second preset range are greater than all values within the third preset range.

6. The method of claim 5, wherein, The calculation formula of the incineration system stability index is: ； wherein, represents the incineration system stability index, represents the distance between a data point and a stability cluster center, represents the distance factor.

7. A stability assessment device for a combustion process, characterized in that, including: An acquisition module is configured to acquire data collected at a plurality of continuous times, the data of each time including a flame image in a combustion process of the incinerator, and state data and flue gas emission monitoring data; An extraction module is configured to extract flame static features and flame dynamic features of each time from the flame image collected at each time, respectively; A calculation module is configured to take the flame static features, the flame dynamic features, the state data and the flue gas emission monitoring data as features, and calculate, for any feature, a Z-score, a linear instability index and a threshold instability index of the corresponding feature at each time; the calculation of the linear instability index of the corresponding feature at each time includes: for any feature, a feature sequence of a current time is acquired in a sliding window manner; a linear fitting slope and a fitting goodness of the feature sequence are determined by linear regression; an absolute value of a product of the linear fitting slope and the fitting goodness is determined as the linear instability index of the feature; the calculation of the threshold instability index of the corresponding feature at each time includes: for any feature, in a case where a value corresponding to the feature is greater than or equal to a lower threshold value and less than or equal to an upper threshold value, the threshold instability index of the feature is determined as 0; in a case where the value corresponding to the feature is greater than the upper threshold value or less than the lower threshold value, the threshold instability index of the feature is determined according to the value, the upper threshold value and the lower threshold value of the feature; The comprehensive module is configured to determine, for each feature at each time, a comprehensive instability index of the feature as a sum of the Z-score, the linear instability index and the threshold instability index of the feature; The clustering module is configured to cluster data of all times according to the comprehensive instability index of each feature at each time, to obtain a stability clustering center; The determining module is configured to determine, according to a distance between data at each time and the stability clustering center, a stability evaluation result of the combustion process of the incinerator at each time.

Images