Method, device, equipment and medium for automatically monitoring enterprise emission reduction based on flow rate

Abstract

The application provides a method, device, equipment and medium for automatically monitoring enterprise emission reduction based on flow rate, and relates to the technical field of environmental protection data monitoring. The method comprises: pre-processing minute-level flow rate data; dividing the data into two clusters by K-means clustering with K=2, extracting the minimum value and the maximum value of the flow rate of the coke pushing state cluster as a coke pushing wave peak identification interval; only in the interval, using the AMPD algorithm to calculate the local maximum value through a multi-scale sliding window and screen the real wave peak, counting the actual coke pushing times, and calculating the number of coke pushing times based on the parameters of the emission permit; and determining whether the enterprise implements the emergency emission reduction measures by comparing the two sizes. It is verified that the overall identification accuracy reaches 97%, the application eliminates low-frequency peak interference, significantly improves the coke pushing time identification accuracy, and can be widely applied to the non-site supervision of heavy pollution emergency emission reduction in the coking industry and the like.

Description

Technical Field

[0001] This application relates to the field of environmental data monitoring technology, and in particular to a method, device, equipment and medium for automatic monitoring of enterprise emission reduction based on flow velocity. Background Technology

[0002] Air pollution control is a crucial aspect of ecological and environmental protection, and emergency emission reduction during heavy pollution alerts is one of the effective means to alleviate severe air pollution in winter. The coking industry is a key sector for air pollution control; its unique processes prevent complete shutdowns. Common emission reduction measures within the industry include delaying coking time to reduce the number of coke pushes, thereby lowering pollutant emissions.

[0003] With the development of information technology in environmental protection supervision, the legal basis for the application of automatic monitoring data of pollution sources has been gradually improved. Electronic data has become one of the effective evidences for ecological and environmental monitoring, providing data support for off-site supervision of coking enterprises. When the dust removal fans in the processes of coking, coal charging, and coke quenching are working, the automatic monitoring data of flue gas velocity will show obvious peaks. These peaks can reflect changes in the coking conditions of the enterprise. Therefore, by identifying the number of velocity peaks and counting the number of coking operations, it can be determined whether the enterprise has implemented emission reduction measures such as delayed coke discharge.

[0004] When identifying flow velocity peaks, two types of peaks are identified: low-frequency peaks and outlier peaks. Low-frequency peaks refer to weak flue gas flow within the flue gas duct even when the dust collector fan in a coking plant is not in a coke-pushing state (such as standby, natural convection in the duct, or slight air leakage). Due to the fan's operating environment or sensor drift, the flow velocity data may exhibit some periodic small fluctuations or low-frequency sensor noise. These fluctuations are usually low in amplitude but have the form of local peaks. Outlier peaks are caused by sudden failures in monitoring equipment, backflushing maintenance, instantaneous packet loss in data transmission, or electromagnetic interference. They are characterized by extremely abnormal and abrupt values ​​(flow velocity < 0 m / s), which physically do not conform to the flue gas flow velocity patterns during normal coking plant production.

[0005] Currently, common methods for peak identification include AMPD peak identification, first-order difference method, and second-order difference method. Among them, AMPD peak identification method has strong adaptability and noise resistance, but when directly applied to the identification of flow velocity data in coking enterprises, it is easy to misidentify low-frequency peaks and outliers as coke pushing peaks, resulting in distorted identification results. The second-order difference method has even lower identification accuracy and cannot meet the precision requirements of supervision. K-means clustering algorithm has the advantages of high efficiency, simplicity, and ability to handle non-uniform data, and has been widely used in fields such as quantity estimation, pollutant identification, and rockburst prediction, but its application in the environmental protection field is still lacking.

[0006] Current peak identification methods suffer from low accuracy due to interference from low-frequency peaks and outliers in identifying the number of coke pushes in coking enterprises. There is an urgent need for a composite identification method that combines machine learning algorithms to improve the accuracy of coke push identification and thus effectively determine the implementation of emission reduction measures during periods of heavy pollution in coking enterprises. Summary of the Invention

[0007] This application provides a method, device, equipment and medium for automatic flow velocity monitoring of enterprise emission reduction. It combines the K-means clustering algorithm with the AMPD peak identification method. First, the exclusive identification interval of the push peak is defined by K-means clustering to eliminate the interference of low-frequency peaks and outlier peaks. Then, the AMPD peak identification method is used to accurately identify the peaks within the interval.

[0008] In a first aspect, this application provides a method for automatic monitoring of enterprise emissions reduction based on flow velocity, the method comprising:

[0009] Obtain automatic monitoring flow velocity data of enterprise emission outlets and pollution discharge permit information containing coke oven parameters, and preprocess the automatic monitoring flow velocity data to obtain an effective flow velocity data sequence;

[0010] Based on the effective flow velocity data sequence, a clustering algorithm is used to divide the flow velocity data into two clusters. The cluster with the higher average flow velocity is identified as the focus-pushing state cluster. The minimum and maximum values ​​of all flow velocity data in the focus-pushing state cluster are extracted as the focus-pushing peak identification interval.

[0011] Within the coke pushing peak identification interval, peak identification is performed on the effective flow rate data sequence to statistically determine the actual number of coke pushing operations; based on the coke oven parameters in the wastewater discharge permit information and the preset delayed coke discharge time, the required number of coke pushing operations is calculated.

[0012] Compare the actual number of coke pushes with the number of coke pushes that should be pushed. If the actual number of coke pushes is greater than the number of coke pushes that should be pushed, it is determined that the enterprise has not implemented the emergency emission reduction measures for heavy pollution. If the actual number of coke pushes is less than or equal to the number of coke pushes that should be pushed, it is determined that the enterprise has implemented the emergency emission reduction measures for heavy pollution.

[0013] In one possible design, the preprocessing of the automatically monitored flow velocity data includes: replacing negative flow velocity values ​​with 0; and replacing abnormally large values ​​of flow velocity greater than or equal to 50 m / s with the average flow velocity during normal production.

[0014] In one possible design, a clustering algorithm is used to divide the flow velocity data into two clusters, specifically including:

[0015] Two flow velocity data points are randomly selected as initial cluster centers. The distance between each flow velocity data point and each initial cluster center is calculated. The data are classified according to the principle of minimizing distance. The cluster centers are iteratively updated until the sum of squared errors is locally minimized, and the clustering results are obtained.

[0016] In one possible design, peak identification is performed on the effective flow rate data sequence within the focus-pushing peak identification interval to statistically determine the actual number of focus pushes, specifically including:

[0017] Perform the following operations on the effective flow velocity data sequence only within the specified focus peak identification interval:

[0018] Construct a multi-scale sliding window sequence, where each window corresponds to a scale parameter;

[0019] For each scale parameter of the sliding window, the local maxima of the window center point relative to other data points in the window are calculated in turn to generate a binary matrix of local maxima at each scale.

[0020] The binary matrices of local maxima at all scales are summed to form a multi-scale feature matrix, where the value at each position represents the number of data points that are identified as local maxima.

[0021] The data points with the largest corresponding values ​​in the multi-scale feature matrix are selected as the true focus push peaks, and each true focus push peak represents one focus push operation.

[0022] The total number of all actual focus push peaks is counted as the actual number of focus pushes.

[0023] In one possible design, the formula for calculating the number of times coke should be pushed is: Number of times coke should be pushed by the enterprise = (Number of coke oven holes / Turnover time) × Delayed coke discharge time.

[0024] In one possible design, the method further includes: outputting key data and judgment conclusions for each enterprise identification period, wherein the key data includes the lower and upper limits of the defined focus push peak identification interval, the actual number of focus pushes, and the number of focus pushes that should be pushed.

[0025] In one possible design, after outputting the key data and judgment conclusions for each enterprise identification period, the method further includes: comparing the judgment conclusions with the on-site verification records to verify the identification accuracy.

[0026] Secondly, this application provides a device for automatically monitoring enterprise emission reduction based on flow velocity, the device comprising:

[0027] The data preprocessing module is configured to acquire automatic monitoring flow velocity data of the enterprise's discharge outlet and discharge permit information containing coke oven parameters, and preprocess the automatic monitoring flow velocity data to obtain an effective flow velocity data sequence.

[0028] The interval identification module is configured to divide the flow velocity data into two clusters based on the effective flow velocity data sequence using a clustering algorithm, identify the cluster with the higher average flow velocity as the focus-pushing state cluster, and extract the minimum and maximum values ​​of all flow velocity data in the focus-pushing state cluster as the focus-pushing peak identification interval.

[0029] The coke pushing count calculation module is configured to perform peak identification on the effective flow rate data sequence within the coke pushing peak identification interval to obtain the actual number of coke pushing counts; and to calculate the number of coke pushing counts based on the coke oven parameters in the wastewater discharge permit information and the preset delayed coke discharge time.

[0030] The judgment module is configured to compare the actual number of times the coke is pushed with the number of times the coke should be pushed. If the actual number of times the coke is pushed is greater than the number of times the coke should be pushed, it is determined that the enterprise has not implemented the emergency emission reduction measures for heavy pollution. If the actual number of times the coke is pushed is less than or equal to the number of times the coke should be pushed, it is determined that the enterprise has implemented the emergency emission reduction measures for heavy pollution.

[0031] Thirdly, embodiments of this application provide an electronic device, including: at least one processor and a memory; the memory stores computer execution instructions; the at least one processor executes the computer execution instructions stored in the memory, causing the at least one processor to perform the method for enterprise emission reduction based on automatic flow rate monitoring as described in the first aspect and various possible designs of the first aspect.

[0032] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions. When a processor executes the computer-executable instructions, it implements the method for enterprise emission reduction based on automatic flow rate monitoring as described in the first aspect and various possible designs of the first aspect.

[0033] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the method for enterprise emission reduction based on automatic flow rate monitoring as described in the first aspect and various possible designs of the first aspect.

[0034] The method, apparatus, equipment, and medium for enterprise emission reduction based on automatic flow velocity monitoring provided in this application have at least the following beneficial effects:

[0035] (1) This application automatically divides the preprocessed effective flow velocity data sequence into a coking state cluster and a non-coking state cluster using K-means clustering (K=2). The minimum and maximum values ​​of the flow velocity in the coking state cluster are extracted as personalized coking peak identification intervals, effectively eliminating the interference of low-frequency peaks and outlier peaks. The AMPD algorithm is used only within this interval to identify the real coking peaks and count the actual number of coking operations, solving the problem that traditional AMPD or second-order difference methods are prone to misjudging low-frequency disturbances as coking peaks, leading to identification distortion. Verified by multiple coking enterprises, the overall identification accuracy rate reached 97%, with the identification accuracy rate of enterprises that have implemented emission reduction measures reaching 98% and that of enterprises that have not implemented them reaching 96%.

[0036] (2) Based on the distribution characteristics of flow rate data under the actual production conditions of different coking enterprises, this application can adaptively calculate the upper and lower limits of the unique coking peak identification range of each enterprise through K-means clustering, thereby adapting to the coking process, monitoring data characteristics and emission outlet differences of different enterprises. There is no need to manually set a fixed threshold, which is suitable for large-scale off-site monitoring in concentrated areas of coking enterprises.

[0037] (3) This application achieves automatic judgment of the implementation of emission reduction measures based on automatic monitoring of flow velocity data, without the need for real-time manual monitoring. The average processing time for a single enterprise from data input to output judgment conclusion is about 30 seconds. Although it is slightly slower than the traditional AMPD or second-order difference method, the pre-clustering step greatly removes interference peaks, which improves the judgment accuracy from less than 40% of the traditional method to 97%. Overall, it significantly improves the accuracy and efficiency of law enforcement supervision and can be widely applied to the field of off-site monitoring of emergency emission reduction in the coking industry. Attached Figure Description

[0038] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0039] Figure 1 A flowchart illustrating a method for enterprise emission reduction based on automatic flow velocity monitoring, provided in this application embodiment;

[0040] Figure 2 The image provided here is a result of identifying Enterprise I using the method described in this application, as shown in the embodiments of this application.

[0041] Figure 3 The image provided in this application shows the result of identifying enterprise I using the existing AMPD algorithm model;

[0042] Figure 4 The image provided here is a result of identifying Enterprise II using the method described in this application, as shown in the embodiments of this application.

[0043] Figure 5 The image provided in this application shows the result of identifying Enterprise II using the existing AMPD algorithm model;

[0044] Figure 6 The image provided here is a result of identifying Enterprise III using the method described in this application, as shown in the embodiments of this application.

[0045] Figure 7 The image provided in this application shows the result of identifying Enterprise III using the existing AMPD algorithm model;

[0046] Figure 8 This is a structural diagram of an enterprise emission reduction device based on automatic flow velocity monitoring, provided in an embodiment of this application.

[0047] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concepts of this application to those skilled in the art through reference to specific embodiments. Detailed Implementation

[0048] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0049] The collection, storage, use, processing, transmission, provision, and disclosure of relevant data and information in the technical solution of this application all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0050] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0051] The technical solution of this application and how it solves the above-mentioned technical problems will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will be described below with reference to the accompanying drawings.

[0052] To address the limitations of existing peak identification methods, such as weak anti-interference capabilities and low accuracy, this application provides a method for enterprise emission reduction based on automatic flow velocity monitoring, such as... Figure 1 As shown, the method for enterprise emission reduction based on automatic flow velocity monitoring includes the following steps S10~S40.

[0053] S10: Obtain automatic monitoring flow velocity data from the enterprise's emission outlet and the discharge permit information containing coke oven parameters, and preprocess the automatic monitoring flow velocity data to obtain an effective flow velocity data sequence.

[0054] Step S10 obtains minute-level flow velocity data from the coking plant's emission outlets, as well as coke oven parameters from the discharge permit, including but not limited to the number of coke ovens, the number of holes per oven, and turnaround time. The raw flow velocity data often contains outliers, negative values, and other contaminants, which can interfere with the accuracy of subsequent clustering and peak identification. An exemplary preprocessing method is to replace negative flow velocity values ​​with 0 to eliminate invalid information caused by negative values; and to replace outliers with flow velocities greater than or equal to 50 m / s with the average flow velocity during the plant's normal production period, thereby avoiding the skewed effect of extreme outliers on the calculation of cluster centers. After the above correction process, a clean, continuous, and effective flow velocity data sequence suitable for subsequent algorithm analysis is obtained. Step S10 provides reliable input data for subsequent K-means clustering and is fundamental to ensuring the accuracy of the entire method.

[0055] In one exemplary embodiment, taking multiple coking enterprises as an example, automatic monitoring data of flue gas velocity during heavy pollution warnings is collected from the enterprise's industrial control computer or data acquisition instrument at a sampling frequency of 1 time / min. Simultaneously, coke oven parameters are read from the enterprise's discharge permit; for example, a certain enterprise has P coke ovens, each with Q carbonization chambers, and a turnover time of T hours (h). The collected velocity data is iterated and checked. If the velocity value of a data point is negative, it is reset to 0; if the velocity value of a data point is greater than or equal to 50 m / s, the average value of all velocity data for the enterprise during non-warning periods (normal production) is calculated, and this average value replaces the abnormal maximum value. After all replacements are completed, a valid velocity data sequence suitable for subsequent processing is obtained.

[0056] S20: Based on the effective flow velocity data sequence, a clustering algorithm is used to divide the flow velocity data into two clusters. The cluster with the higher average flow velocity is identified as the focus-pushing state cluster. The minimum and maximum values ​​of all flow velocity data in the focus-pushing state cluster are extracted as the focus-pushing peak identification interval.

[0057] Step S20 executes the K-means clustering algorithm on the preprocessed effective flow velocity data sequence, setting the number of clusters K to be equal to 2. Specifically, two flow velocity data are randomly selected as initial cluster centers. The Euclidean distance between each flow velocity data and the two centers is calculated, and the data is assigned to the nearest class. Then, the mean of each class is recalculated as the new cluster center. This process is iterated until the sum of squared errors reaches a local minimum, ultimately outputting two clusters. The cluster with a higher mean flow velocity and a periodic fluctuation corresponds to the flue gas velocity change during normal coke pushing operations and is called the coke pushing state cluster. The other cluster with a lower mean flow velocity corresponds to non-coke pushing or interference states and is called the non-coke pushing state cluster. The minimum and maximum values ​​of all flow velocity data in the coke pushing state cluster are extracted to obtain the coke pushing peak identification interval corresponding to the enterprise. This interval eliminates interference from low-frequency noise and outlier peaks, defining a dedicated search range for subsequent AMPD peak identification, allowing the algorithm to find the true coke pushing peak only within the effective interval, significantly reducing the false identification rate.

[0058] In one exemplary implementation, for a coking plant, its effective flow velocity data sequence is input into a K-means clustering algorithm, with K equal to 2. After initializing two random cluster centers, the algorithm converges to two clusters after multiple rounds of distance calculation and center updates. The mean flow velocity of each cluster is calculated; for example, if the mean of cluster A is V1 m / s and the mean of cluster B is V2 m / s, and V1 > V2, then cluster A is identified as the coking state cluster. The specific values ​​of all flow velocity data in cluster A are extracted, and the minimum value V is found. min and maximum value V max Then the identification range of the coking peak of the enterprise is (V min V max All flow rate peaks generated by the focus pushing operation fall within this range, while peak values ​​generated by periodic low-frequency fluctuations or sensor noise are typically below V. min It was excluded from the recognition range.

[0059] S30: Within the coke pushing peak identification interval, peak identification is performed on the effective flow rate data sequence to obtain the actual number of coke pushing times; based on the coke oven parameters in the pollutant discharge permit information and the preset delayed coke discharge time, the number of coke pushing times to be calculated is obtained.

[0060] Step S30 specifically includes two parallel subtasks.

[0061] First, within the peak identification interval defined in step S20, the Automatic Multi-Scale Peak Identification (AMPD) algorithm is used to identify peaks in the effective flow velocity data sequence. The specific process is as follows: a multi-scale sliding window sequence is constructed, with each window corresponding to different scale parameters; for each scale, the sliding window sequentially calculates the local maxima of the window's center point relative to the remaining data points within the window, generating a binary matrix at that scale; the binary matrices of all scales are summed to form a multi-scale feature matrix, where the value at each position indicates how many scales the data point is considered a local maxima; the data point with the largest corresponding value in the multi-scale feature matrix is ​​selected as the true peak, with each true peak representing one peak; the total number of all true peaks is counted, thus obtaining the actual number of peaks pushed by the enterprise during the warning period.

[0062] Second, based on the coke oven parameters in the pollution discharge permit and the delayed coking time required by the heavy pollution warning, the number of times the enterprise should push coke is calculated according to the formula "Number of times the enterprise should push coke = (Number of coke oven holes / Turnover time) × Delayed coking time".

[0063] Step S30 limits the AMPD algorithm to operate within the personalized recognition range, which not only leverages the advantages of AMPD's multi-scale noise resistance but also avoids its inherent defect of misjudging low-frequency disturbances as focus peaks. At the same time, it calculates the theoretically allowed number of focus pushes by combining license parameters, providing accurate data support for subsequent judgments.

[0064] In one exemplary implementation, for the same enterprise, only values ​​falling within (V) of the time series of flow rate data per minute are selected. min V max The data points within the range are subjected to AMPD analysis. Specifically, the scale parameter of the multi-scale sliding window is set from 2 to N, where N is usually 1 / 10 of the length of the velocity data sequence. For a window with a certain scale parameter k, the sliding window calculates whether each position is a local maximum in its neighborhood, marking it as 1 or 0, forming a binary matrix at that scale. This operation is repeated until all scales are completed, and all binary matrices are summed to obtain the multi-scale feature matrix. The velocity data point corresponding to the position with the largest accumulated value in the matrix is ​​found and marked as the true coking peak. Assuming that a total of M true peaks are found during the warning period, the actual number of coking operations is M. At the same time, according to the pollution discharge permit information: the enterprise has P coke ovens, each with Q holes, and a turnaround time of T hours. The delayed coking time required by the heavy pollution warning is D hours (h). Among them, the number of coke oven holes is P × Q holes, divided by the turnaround time T, and then multiplied by the delayed coking time D to obtain the number of coking operations required. The calculation result can be rounded down as needed, usually by omitting the decimal part of the calculated number of times and keeping only the integer part as the number of times to be pushed.

[0065] S40: Compare the actual number of coke pushes with the number of coke pushes that should be pushed. If the actual number of coke pushes is greater than the number of coke pushes that should be pushed, it is determined that the enterprise has not implemented the emergency emission reduction measures for heavy pollution. If the actual number of coke pushes is less than or equal to the number of coke pushes that should be pushed, it is determined that the enterprise has implemented the emergency emission reduction measures for heavy pollution.

[0066] Step S40 is implemented based on the actual number of coke pushes and the required number of coke pushes obtained in step S30. The actual number of coke pushes is identified and statistically analyzed by the AMPD algorithm within the coke push peak identification range, while the required number of coke pushes is calculated based on the discharge permit parameters and the delayed coke discharge time. The determination of the implementation of emission reduction measures is completed by comparing the actual number of coke pushes and the required number of coke pushes. During periods of heavy pollution warnings, coking enterprises are required to reduce the number of coke pushes by delaying coke discharge; therefore, if the actual number of coke pushes exceeds the allowed required number, it means that the enterprise has not complied with emission reduction regulations. The determination rule is logically clear and quantifiable, avoiding the bias of subjective human judgment. This determination result can be directly used as the basis for off-site supervision. Verification showed that in a practical application involving 100 coking enterprises, the overall identification accuracy rate reached 97%, with 98% accuracy for enterprises that had implemented the measures and 96% accuracy for those that had not.

[0067] In one exemplary embodiment, taking a certain enterprise as an example, let M be the actual number of coke pushing operations calculated in step S30, and E be the required number of coke pushing operations. If M > E, it is determined that the enterprise has not implemented emergency emission reduction measures for heavy pollution and should be given attention or subject to on-site verification. If M ≤ E, it is determined that the enterprise has implemented emergency emission reduction measures for heavy pollution.

[0068] In one exemplary embodiment, after step S40, the step further includes step S50: outputting and verifying the results.

[0069] In this embodiment, after the determination is completed in step S40, two levels of final results are output. The first level of final results consists of key data for each company within the identified time period. For example, the key data can be presented in the form of a result table, which includes the company identifier and the lower limit V of the coke push flow rate range. min With upper limit V max The actual number of focus pushes (M), the number of focus pushes that should have been made (E), and the conclusion regarding whether or not the work was carried out.

[0070] The second level's final result is the verification statistics, including the overall identification accuracy, the accuracy rates of enterprises that have implemented the measures, and the average processing time per enterprise. The verification method uses on-site inspection records as a benchmark, comparing the judgment conclusion output in step S40 with the number of coke pushes and the emission reduction implementation status in the on-site inspection records for the corresponding time period, and calculating the accuracy accordingly. This verification directly uses the actual number of coke pushes obtained in step S30 to check the overlap with the on-site records, forming a closed-loop feedback with the coke push peak identification range output in step S20 and the peak count output in step S30.

[0071] In summary, compared with single AMPD or second-order difference method, the method of this application effectively eliminates low-frequency interference by defining personalized focus push flow rate range by front-end K-means clustering, thereby increasing the overall recognition accuracy from 20% of second-order difference method and 40% of AMPD method to 97%.

[0072] To fully demonstrate the feasibility and advancement of the method in this application, this embodiment collects flow rate monitoring data and coke oven parameters from 100 coking enterprises and summarizes the results obtained after processing through steps S10-S50 above. Regarding the implementation details of each step, in step S10, the flow rate data is preprocessed by replacing negative flow rate values ​​with 0 and replacing abnormally large values ​​with flow rates ≥50m / s with the average flow rate during normal production, thus eliminating interference from dirty data for subsequent identification. In step S20, the K-means clustering algorithm is executed on the preprocessed flow rate data, setting the number of clusters K=2. The specific clustering process has been described in the above embodiment and will not be repeated here. In step S30, AMPD peak identification counts the actual number of coke pushes and calculates the number of coke pushes that should be performed during periods of heavy pollution. In step S40, the actual number of coke pushes for each enterprise is compared with the number of coke pushes that should be performed. If the actual number of coke pushes > the number of coke pushes that should be performed, it is determined that emission reduction measures have not been implemented; if the actual number of coke pushes ≤ the number of coke pushes that should be performed, it is determined that emission reduction measures have been implemented.

[0073] The method described in this application was used to identify 100 coking enterprises in 2024. The results are shown in Table 1. Among them, 50 enterprises were identified by the model, and 49 of them were accurately identified, with an accuracy rate of 98%. 50 enterprises were not identified, and 48 of them were accurately identified, with an accuracy rate of 96%.

[0074] Table 1 Model Recognition Accuracy

[0075]

[0076] Therefore, compared with single AMPD or second-order difference method, the method of this application effectively eliminates low-frequency interference by defining personalized focus push flow rate range by front-end K-means clustering, thereby increasing the overall recognition accuracy from 20% of second-order difference method and 40% of AMPD method to 97%.

[0077] Based on the verification results of 100 coking enterprises, the method in this application has achieved a significant improvement in overall recognition accuracy. To further illustrate the difference in recognition performance between the method in this application and the single AMPD algorithm on specific enterprise flow velocity data, the following combines... Figures 2-7 Please provide an explanation.

[0078] Figure 2 The image provided in this application shows the result of identifying Enterprise I using the method of this application. During periods of heavy pollution, Enterprise I should push coke 30 times a day in accordance with regulations. According to the K-means composite algorithm model, 39 times of coke pushing were identified within the specified time period, which is consistent with the actual number of times on site. The model identification result is accurate and reliable.

[0079] Figure 3 The image provided in this application shows the results of identifying company I using the existing AMPD algorithm model. Directly using the AMPD algorithm to identify the number of times the company pushed the focus during the research period yielded 47 times, which deviates significantly from the actual value. The reason for this result is that using the AMPD algorithm alone to identify peak values ​​will treat some low-frequency peak values ​​as push-the-focus peak values, resulting in an increased number of identified peak values ​​and distorted identification results.

[0080] Figure 4 The image provided in this application illustrates the identification results of Enterprise II using the method described in this application. Enterprise II's compliant coking limit during emission reduction periods is 45 times. According to the K-means composite algorithm model, the actual coking during the specified period was 74 times, exceeding the production limit. The identification data matches the on-site records, and the model's judgment is accurate.

[0081] Figure 5 The image provided in this application shows the result of identifying Enterprise II using the existing AMPD algorithm model. The AMPD algorithm was used to identify the number of times the enterprise pushed the focus during this period, and the result was 88 times. The reason for this is still due to the local low-frequency peaks, which caused the number of times the focus was identified to increase.

[0082] Figure 6 The image shows the results of identifying Enterprise III using the method described in this application. Enterprise III complied with the emission reduction requirements by pushing coke 40 times during the emission reduction period. The K-means composite algorithm model identified 101 actual coke pushes during the period, which far exceeded the limit. Therefore, it was determined that the enterprise did not comply with the emission reduction requirements. The identification results are consistent with the actual situation on site.

[0083] Figure 7 The image shown is a result of identifying Enterprise III using the existing AMPD algorithm model provided in this application embodiment. The AMPD algorithm was used to identify the number of times the enterprise pushed the focus during this period, and the result was 117 times. The reason for this is still due to the local low-frequency peaks, which caused the number of times the focus was identified to increase.

[0084] AMPD Algorithm Model Source Explanation: Scholkmann et al. proposed an automatic multi-scale peak detection (AMPD) algorithm for quasi-periodic signals with motion interference. The algorithm detects peak points by calculating the local maximum scale value of the signal. This method has achieved good results in ECG signal peak detection. However, the main peak of the pulse wave signal under complex morphology is not necessarily the local maximum value point. The algorithm may cause peak points to be missed. For details, please refer to "Scholkmann F, BossJ, Wolf M. An efficient algorithm for automatic peak detection in noisyperiodic and quasi-periodic signals[J]. algorithms, 2012, 5 (4):588-603."

[0085] This application also provides a device for automatically monitoring enterprise emission reduction based on flow velocity, used to implement the methods described in any of the above embodiments, such as... Figure 8 As shown, the device includes:

[0086] The data preprocessing module 801 is configured to acquire automatic monitoring flow velocity data of the enterprise's discharge outlet and discharge permit information containing coke oven parameters, and preprocess the automatic monitoring flow velocity data to obtain an effective flow velocity data sequence.

[0087] The interval identification module 802 is configured to divide the flow velocity data into two clusters based on the effective flow velocity data sequence using a clustering algorithm, identify the cluster with the higher average flow velocity as the focus-pushing state cluster, and extract the minimum and maximum values ​​of all flow velocity data in the focus-pushing state cluster as the focus-pushing peak identification interval.

[0088] The coke pushing count calculation module 803 is configured to perform peak identification on the effective flow rate data sequence within the coke pushing peak identification interval to obtain the actual number of coke pushing counts; and to calculate the number of coke pushing counts based on the coke oven parameters in the wastewater discharge permit information and the preset delayed coke discharge time.

[0089] The judgment module 804 is configured to compare the actual number of coke pushes with the number of coke pushes that should be pushed. If the actual number of coke pushes is greater than the number of coke pushes that should be pushed, it is determined that the enterprise has not implemented the emergency emission reduction measures for heavy pollution. If the actual number of coke pushes is less than or equal to the number of coke pushes that should be pushed, it is determined that the enterprise has implemented the emergency emission reduction measures for heavy pollution.

[0090] This application provides an electronic device. The electronic device may include a processor and a memory, wherein the processor and the memory can communicate; exemplarily, the processor and the memory communicate via a communication bus.

[0091] The processor executes computer execution instructions stored in memory, causing the processor to perform the scheme in the above embodiments. The processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0092] The communication bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The system bus can be divided into address bus, data bus, control bus, etc. Transceivers are used to enable communication between database access devices and other computers (e.g., clients, read-write libraries, and read-only libraries). Memory may include random access memory (RAM) and may also include non-volatile memory.

[0093] The electronic device provided in this application embodiment can be the terminal device described in the above embodiments.

[0094] This application also provides a computer-readable storage medium storing computer instructions. When the computer instructions are executed on a computer, the computer performs the technical solution of the above-described method for enterprise emission reduction based on automatic flow rate monitoring.

[0095] This application also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. At least one processor can read the computer program from the computer-readable storage medium. When the at least one processor executes the computer program, it can implement the technical solution of the enterprise emission reduction method based on flow rate automatic monitoring in the above embodiments.

[0096] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0097] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to implement the solution of this embodiment according to actual needs.

[0098] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The unit composed of the above modules can be implemented in hardware or in the form of hardware plus software functional units.

[0099] The integrated modules described above, implemented as software functional modules, can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods of the various embodiments of this application.

[0100] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly manifested as execution by a hardware processor, or execution by a combination of hardware and software modules within the processor.

[0101] The memory may include high-speed RAM, and may also include non-volatile storage (NVM), such as at least one disk storage device, and may also be a USB flash drive, external hard drive, read-only memory, disk or optical disc, etc.

[0102] Buses can be Industry Standard Architecture (ISA) buses, Peripheral Component Interconnect (PCI) buses, or Extended Industry Standard Architecture (EISA) buses, etc. Buses can be categorized into address buses, data buses, control buses, etc.

[0103] The aforementioned storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0104] An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Alternatively, the storage medium can be an integral part of the processor. The processor and storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and storage medium can exist as discrete components in an electronic control unit or main control device.

[0105] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for enterprise emission reduction based on automatic flow velocity monitoring, characterized in that, The method includes: Obtain automatic monitoring flow velocity data of enterprise emission outlets and pollution discharge permit information containing coke oven parameters, and preprocess the automatic monitoring flow velocity data to obtain an effective flow velocity data sequence; Based on the effective flow velocity data sequence, a clustering algorithm is used to divide the flow velocity data into two clusters. The cluster with the higher average flow velocity is identified as the focus-pushing state cluster. The minimum and maximum values ​​of all flow velocity data in the focus-pushing state cluster are extracted as the focus-pushing peak identification interval. Within the coke pushing peak identification interval, peak identification is performed on the effective flow rate data sequence to statistically determine the actual number of coke pushing operations; based on the coke oven parameters in the wastewater discharge permit information and the preset delayed coke discharge time, the required number of coke pushing operations is calculated. Compare the actual number of times the coke is pushed with the number of times the coke should be pushed. If the actual number of times the coke is pushed is greater than the number of times the coke should be pushed, it is determined that the enterprise has not implemented the emergency emission reduction measures for heavy pollution. If the actual number of times the coke is pushed is less than or equal to the number of times the coke should be pushed, it is determined that the enterprise has implemented the emergency emission reduction measures for heavy pollution. Within the focus-pushing peak identification interval, peak identification is performed on the effective flow rate data sequence to statistically determine the actual number of focus pushes, specifically including: Perform the following operations on the effective flow velocity data sequence only within the specified focus peak identification interval: Construct a multi-scale sliding window sequence, where each window corresponds to a scale parameter; For each scale parameter of the sliding window, the local maxima of the window center point relative to other data points in the window are calculated in turn to generate a binary matrix of local maxima at each scale. The binary matrices of local maxima at all scales are summed to form a multi-scale feature matrix, where the value at each position represents the number of data points that are identified as local maxima. The data point with the largest corresponding value in the multi-scale feature matrix is ​​selected as the true focus push peak, and each true focus push peak represents one focus push operation. The total number of all actual focus push peaks is counted as the actual number of focus pushes.

2. The method for enterprise emission reduction based on automatic flow velocity monitoring according to claim 1, characterized in that, The preprocessing of the automatically monitored flow velocity data includes: replacing negative flow velocity values ​​with 0; and replacing abnormally large values ​​of flow velocity greater than or equal to 50 m / s with the average flow velocity during normal production.

3. The method for enterprise emission reduction based on automatic flow velocity monitoring according to claim 1, characterized in that, Clustering algorithms were used to divide the flow velocity data into two clusters, specifically including: Two flow velocity data points are randomly selected as initial cluster centers. The distance between each flow velocity data point and each initial cluster center is calculated. The data are classified according to the principle of minimizing distance. The cluster centers are iteratively updated until the sum of squared errors is locally minimized, and the clustering results are obtained.

4. The method for enterprise emission reduction based on automatic flow velocity monitoring according to claim 1, characterized in that, The formula for calculating the number of times coke should be pushed is: Number of times coke should be pushed by the enterprise = (Number of coke oven holes / Turnover time) × Delayed coke discharge time.

5. The method for enterprise emission reduction based on automatic flow velocity monitoring according to claim 1, characterized in that, The method further includes: outputting key data and judgment conclusions for each enterprise identification period, wherein the key data includes the lower and upper limits of the defined focus push peak identification interval, the actual number of focus pushes, and the number of focus pushes that should be pushed.

6. The method for enterprise emission reduction based on automatic flow velocity monitoring according to claim 5, characterized in that, After outputting the key data and judgment conclusions for each enterprise identification period, the method further includes: comparing the judgment conclusions with the on-site verification records to verify the identification accuracy.

7. A device for automatic flow velocity monitoring in enterprise emission reduction, characterized in that, The device includes: The data preprocessing module is configured to acquire automatic monitoring flow velocity data of the enterprise's discharge outlet and discharge permit information containing coke oven parameters, and preprocess the automatic monitoring flow velocity data to obtain an effective flow velocity data sequence. The interval identification module is configured to divide the flow velocity data into two clusters based on the effective flow velocity data sequence using a clustering algorithm, identify the cluster with the higher average flow velocity as the focus-pushing state cluster, and extract the minimum and maximum values ​​of all flow velocity data in the focus-pushing state cluster as the focus-pushing peak identification interval. The module for calculating the number of times coke should be pushed is configured to identify peaks in the effective flow rate data sequence within the coke pushing peak identification interval to statistically determine the actual number of times coke should be pushed; and to calculate the number of times coke should be pushed based on the coke oven parameters in the discharge permit information and the preset delayed coke discharge time; wherein, identifying peaks in the effective flow rate data sequence within the coke pushing peak identification interval to statistically determine the actual number of times coke should be pushed specifically includes: Perform the following operations on the effective flow velocity data sequence only within the specified focus peak identification interval: Construct a multi-scale sliding window sequence, where each window corresponds to a scale parameter; For each scale parameter of the sliding window, the local maxima of the window center point relative to other data points in the window are calculated in turn to generate a binary matrix of local maxima at each scale. The binary matrices of local maxima at all scales are summed to form a multi-scale feature matrix, where the value at each position represents the number of data points that are identified as local maxima. The data point with the largest corresponding value in the multi-scale feature matrix is ​​selected as the true focus push peak, and each true focus push peak represents one focus push operation. The total number of all actual focus push peaks is counted as the actual number of focus pushes. The judgment module is configured to compare the actual number of times the coke is pushed with the number of times the coke should be pushed. If the actual number of times the coke is pushed is greater than the number of times the coke should be pushed, it is determined that the enterprise has not implemented the emergency emission reduction measures for heavy pollution. If the actual number of times the coke is pushed is less than or equal to the number of times the coke should be pushed, it is determined that the enterprise has implemented the emergency emission reduction measures for heavy pollution.

8. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method for enterprise emission reduction based on automatic flow rate monitoring as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method for enterprise emission reduction based on automatic flow rate monitoring as described in any one of claims 1-6.

Images