A sewage treatment method, device and system in a deep processing process of coal tar

Abstract

The present application relates to the technical field of sewage treatment, in particular to a sewage treatment method, device and system in the deep processing of coal tar. The present application firstly obtains historical data and real-time data of various indexes; further predicts the pollution degree coefficient of the adjusting pool at the corresponding next moment according to the change of the feedforward index; further screens out the abnormal fluctuation section of the pollution degree coefficient and the data of each feedback index; further obtains the response fluctuation section corresponding to the abnormal fluctuation section of the pollution degree coefficient; and finally adjusts the inflow rate of high-concentration wastewater. The present application predicts the wastewater concentration in the adjusting pool through the feedforward index, and combines the response characteristics of the feedback index to predict the influence degree of the wastewater concentration on the sewage treatment, so as to adjust the inflow rate based on the current wastewater concentration change, dynamically predict the system load, accurately control the inflow amount, and effectively improve the sewage treatment efficiency and system stability.

Description

Technical Field

[0001] This invention relates to the field of wastewater treatment technology, specifically to a wastewater treatment method, apparatus, and system for the deep processing of coal tar. Background Technology

[0002] Wastewater generated during the deep processing of coal tar has a complex composition and usually contains highly toxic and recalcitrant pollutants (such as phenols, ammonia nitrogen, sulfides, cyanides, etc.). It requires multi-stage treatment to achieve compliant discharge or reuse.

[0003] Because the water quality of coal tar wastewater fluctuates greatly, in order to avoid the instantaneous impact of highly toxic or high-load wastewater on microbial activity, which could lead to sludge poisoning or mass mortality, water quality regulation (i.e., mixing and regulating high-concentration wastewater with low-concentration wastewater) is used to stabilize the water quality zone, thereby avoiding load shocks, reducing the difficulty of physicochemical treatment, and improving treatment efficiency.

[0004] Existing methods typically use real-time water quality data to adaptively adjust the flow rate of the influent pump for high-concentration wastewater, ensuring uniform mixing of wastewater with varying concentrations. However, in actual operation, the concentration of wastewater discharged from different coal tar processing steps varies greatly, and the switching is frequent. Existing methods struggle to adjust the influent pump flow rate in a timely manner based on these frequent changes, potentially leading to adjustment anomalies and affecting the treatment effectiveness of subsequent biochemical and physicochemical steps. Summary of the Invention

[0005] To address the technical problem that existing water quality regulation methods are unable to cope with the frequent changes in coal tar wastewater concentration, easily leading to regulation anomalies and affecting wastewater treatment efficiency, the present invention aims to provide a wastewater treatment method, apparatus, and system for the deep processing of coal tar. The specific technical solution adopted is as follows:

[0006] A wastewater treatment method in the deep processing of coal tar, the method comprising:

[0007] Historical and real-time data of preset feedforward and feedback indicators in the wastewater treatment process are acquired and divided into preset periods.

[0008] Obtain the growth segments of the feedforward indicators; based on the growth trend of each growth segment and the data value at each time point in the segment, obtain the growth coefficient for each time point; based on the growth coefficients of all feedforward indicators at each time point, and combined with the correlation characteristics between the various feedforward indicators, obtain the pollution level coefficient of the regulating pool at the next time point.

[0009] Within each historical preset period, the data of pollution degree coefficient and various feedforward indicators are divided into time series based on adjacent extreme points, and fluctuation segments are obtained for each. Based on the fluctuation amplitude and salience within each fluctuation segment, anomaly indices are obtained and abnormal fluctuation segments are screened out. Based on the time interval between the abnormal fluctuation segments of the pollution degree coefficient and the abnormal fluctuation segments of the subsequent feedforward indicators, combined with the similarity of the abnormal indices, the response fluctuation segments corresponding to the abnormal fluctuation segments of the pollution degree coefficient are obtained.

[0010] Based on the overall response characteristics of the pollution degree coefficient to the response fluctuations of various feedforward indicators in historical data, and combined with the changing trend of the pollution degree coefficient at the current moment, the influent disturbance index at the current moment is obtained; the influent flow rate of high-concentration wastewater is adjusted based on the influent disturbance index at the current moment and the pollution degree coefficient at the next moment.

[0011] Furthermore, the method for obtaining the growth coefficient includes:

[0012] Based on the number of first-order difference values ​​in the growth segment, the overall characteristics of the first-order difference values, and the data value at each time step in the segment, the growth coefficient corresponding to each time step is obtained.

[0013] Furthermore, the method for obtaining the pollution degree coefficient includes:

[0014] Select any time as the target time; based on all data from the target time to the start time of the preset period to which it belongs, obtain the Pearson correlation coefficient matrix between each feedforward index at the target time and obtain the feature vector matrix;

[0015] Based on the vector elements corresponding to the eigenvectors of each feedforward index at the maximum eigenvalue, the growth coefficients of each feedforward index at the target time are weighted and summed, and the weighted sum is used as the pollution degree coefficient at the next time step at the target time.

[0016] Furthermore, the method for obtaining the anomaly index includes:

[0017] Based on the overall fluctuation amplitude of the fluctuation segment and the overall difference between the data of the fluctuation segment and all historical data in the same dimension, an anomaly index corresponding to the fluctuation segment is obtained; the overall fluctuation amplitude is positively correlated with the anomaly index.

[0018] Furthermore, the method for obtaining the response fluctuation segment includes:

[0019] Select any of the aforementioned feedback indicators as the feedback target indicator, select any of the aforementioned pollution degree coefficients’ abnormal fluctuation segments as the target fluctuation segments, and within the preset period to which the target fluctuation segment belongs, and in the time domain after the start time of the target fluctuation segment, each abnormal fluctuation segment of the feedback target indicator is taken as the fluctuation segment to be analyzed.

[0020] Based on the time interval between the start time of each fluctuation segment to be analyzed and the target fluctuation segment, and combined with the similarity of the anomaly index, a matching coefficient is obtained for each fluctuation segment to be analyzed; the fluctuation segment to be analyzed with the largest matching coefficient is selected as the response fluctuation segment of the target fluctuation segment.

[0021] Furthermore, the method for obtaining the water inflow disturbance index includes:

[0022] Based on the increase or decrease of the abnormal fluctuation segment of the pollution degree coefficient, all the response fluctuation segments are divided into two categories;

[0023] For each type of response fluctuation segment, the predicted response duration is obtained by combining the time interval between the start time of the type of response fluctuation segment and the time interval between the end time of the type of response fluctuation segment and the start time of the abnormal fluctuation segment of the corresponding pollution degree coefficient.

[0024] Based on the distribution characteristics of the fluctuation amplitude of a certain type of response fluctuation segment, and combined with the predicted response duration, the inflow disturbance index of a certain type of response fluctuation segment is obtained.

[0025] Based on the current trend of the pollution level coefficient, the corresponding influent disturbance index is selected.

[0026] Furthermore, the method for obtaining the influent flow rate includes:

[0027] If the pollution level coefficient at the current moment shows an increasing trend, an adjustment factor is obtained by combining the influent disturbance index at the current moment and the pollution level coefficient at the next moment; both the influent disturbance index and the pollution level coefficient at the next moment are negatively correlated with the adjustment factor.

[0028] If the pollution level coefficient at the current moment shows a decreasing trend, an adjustment factor is obtained by combining the influent disturbance index at the current moment and the pollution level coefficient at the next moment; the influent disturbance index is positively correlated with the adjustment factor; the pollution level coefficient at the next moment is negatively correlated with the adjustment factor.

[0029] The influent flow rate at the current moment is adjusted based on the adjustment factor.

[0030] Furthermore, the time domain length of the growth segment is at least 3.

[0031] This invention also proposes a wastewater treatment system for the deep processing of coal tar, the system comprising:

[0032] Data acquisition module: acquires historical and real-time data of preset feedforward and feedback indicators during the wastewater treatment process, and divides them into preset periods;

[0033] Data prediction module: Obtain the growth segment of the feedforward indicator; based on the growth trend of each growth segment and the data value at each time point in the segment, obtain the growth coefficient corresponding to each time point; based on the growth coefficient of all feedforward indicators at each time point, and combined with the correlation characteristics between various feedforward indicators, obtain the pollution degree coefficient of the regulating pool at the next time point.

[0034] Response matching module: Within each historical preset period, the data of pollution degree coefficient and various feedforward indicators are divided into time series based on adjacent extreme points, and fluctuation segments are obtained for each. Based on the fluctuation amplitude and salience within each fluctuation segment, anomaly indices are obtained and abnormal fluctuation segments are screened out. Based on the time interval between the abnormal fluctuation segments of the pollution degree coefficient and the abnormal fluctuation segments of the subsequent feedforward indicators, combined with the similarity of the abnormal indices, the response fluctuation segments corresponding to the abnormal fluctuation segments of the pollution degree coefficient are obtained.

[0035] Influent regulation module: Based on the overall response characteristics of the pollution degree coefficient to the response fluctuations of various feedforward indicators in historical data, and combined with the current trend of the pollution degree coefficient, the influent disturbance index at the current moment is obtained; the influent flow rate of high-concentration wastewater is adjusted based on the influent disturbance index at the current moment and the pollution degree coefficient at the next moment.

[0036] The present invention also proposes a wastewater treatment device in the deep processing of coal tar, the device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the steps of the wastewater treatment method in the deep processing of coal tar.

[0037] The present invention has the following beneficial effects:

[0038] This invention first acquires historical and real-time data for various indicators to provide a data foundation. Then, based on the growth trend of each growth segment and the data value at each moment within that segment, it quantifies the growth degree at each growth moment, providing a basis for subsequent prediction of pollution levels by integrating multiple feedforward indicators. Next, it obtains the pollution level coefficient of the equalization tank at the next corresponding moment, predicting the system load in advance and facilitating subsequent adjustment of the influent flow rate. Furthermore, it filters out abnormal fluctuation segments of the pollution level coefficient and various feedforward indicators, matching them based on time intervals and the similarity of abnormal indices to obtain the response fluctuation segments corresponding to the abnormal fluctuation segments of the pollution level coefficient. This facilitates subsequent analysis of the response of feedforward indicators to abnormal fluctuations in the pollution level coefficient and predicts the impact of pollution levels on subsequent water purification steps. Finally, based on the overall response characteristics of the response fluctuation segments of various feedforward indicators to the pollution level coefficient in historical data, it analyzes the response patterns, and then, combined with the current trend of the pollution level coefficient and the pollution level coefficient at the next moment, predicts the interference of wastewater concentration on wastewater treatment and adjusts the influent flow rate of high-concentration wastewater. This invention predicts the wastewater concentration in the equalization tank using feedforward indicators and combines the response characteristics of feedback indicators to predict the impact of wastewater concentration on wastewater treatment. Based on the current changes in wastewater concentration, the influent flow rate is adjusted, the system load is dynamically predicted, and the influent volume is precisely controlled, effectively improving wastewater treatment efficiency and system stability. Attached Figure Description

[0039] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 A flowchart illustrating wastewater treatment for deep processing of coal tar, provided as an embodiment of the present invention;

[0041] Figure 2 A flowchart illustrating a wastewater treatment method in the deep processing of coal tar, as provided in one embodiment of the present invention;

[0042] Figure 3 This is a system block diagram of a wastewater treatment system in the deep processing of coal tar, provided as an embodiment of the present invention. Detailed Implementation

[0043] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation methods, structures, features, and effects of a wastewater treatment method, apparatus, and system in the deep processing of coal tar according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0045] The following description, in conjunction with the accompanying drawings, details a specific scheme for a wastewater treatment method, apparatus, and system in the deep processing of coal tar provided by the present invention.

[0046] Please see Figure 1 It illustrates a flowchart of wastewater treatment for deep processing of coal tar according to an embodiment of the present invention. Figure 1 The diagram shows that the wastewater undergoes three main treatment stages. The incoming water first enters the pretreatment stage, passing through the oil-removal tank, oil-water separator, phenol removal and ammonia stripping tower, and equalization tank. Then, it enters the biological treatment stage, passing through the hydrolysis acidification tank, UASB / IC reactor, A / O process, and MBR membrane tank. Finally, it enters the advanced treatment stage, passing through coagulation sedimentation, adsorption, ultrafiltration, and reverse osmosis. Depending on whether the treatment meets the standards, the wastewater is determined to be either effluent or reused.

[0047] It can be briefly described as: pretreatment (oil separator, equalization tank) → biochemical treatment (A / O, MBR) → advanced treatment (coagulation sedimentation, advanced oxidation) → resource recovery.

[0048] The pretreatment of high-concentration wastewater includes: ammonia stripping and phenol removal, oil removal, and water quality conditioning (balancing water quality and quantity through a conditioning tank to avoid system shock). The water quality conditioning process is as follows: high-concentration wastewater (usually wastewater after ammonia stripping and phenol removal and oil removal) and low-concentration wastewater (usually circulating cooling water, flushing water, etc.) are combined at the inlet of the conditioning tank using an online static mixer and thoroughly mixed by a stirring device.

[0049] When the treated water meets the standards for industrial circulating cooling water, it is reused in production. When the water quality is good but does not meet the standards for direct reuse, it is transferred back to the equalization tank, thereby realizing water circulation.

[0050] Please see Figure 2 The diagram illustrates a flowchart of a wastewater treatment method in the deep processing of coal tar according to an embodiment of the present invention, specifically including:

[0051] Step S1: Obtain historical and real-time data of preset feedforward and feedback indicators during the wastewater treatment process, and divide them into preset periods.

[0052] In this embodiment of the invention, before wastewater conditioning, the changes in pollutant concentration in the conditioning tank are predicted by combining historical data and current actual monitoring indicators, and the degree of influence of wastewater concentration on wastewater treatment is analyzed.

[0053] In one embodiment of the present invention, the monitoring indicators and data collection methods include:

[0054] ① The sulfur content (%) of the raw materials was monitored using the ZDL-9 intelligent sulfur analyzer;

[0055] ② The temperature in the ammonia stripping and phenol removal unit is monitored (°C) using a temperature sensor;

[0056] ③Use an ultraviolet fluorescence sensor to monitor the oil content (mg / L) during the degreasing process;

[0057] ④ Use a TOC analyzer to monitor COD (chemical oxygen demand) (mg / L) in the equalization tank and biochemical unit;

[0058] ⑤ The sludge concentration (mg / L) of the biochemical unit was monitored using an online sludge concentration monitoring instrument based on infrared scattering.

[0059] ⑥ The dissolved oxygen concentration (mg / L) in the biochemical unit was monitored using a fluorescence DO sensor.

[0060] ⑦ The residual amount of oxidant in the deep treatment unit (mg / L) was monitored using an online chemiluminescence analyzer.

[0061] The system uses a 24-hour period as a preset cycle, with each day constituting one cycle, to avoid excessively long data sets that could obscure subtle data features. The data collection frequency is once every 5 seconds, acquiring the historical data for the most recent 30 days for the current day, and also acquiring various data for the current day in real time.

[0062] It should be noted that the monitoring data can be preprocessed by normalization during collection. Normalization is performed under the corresponding data dimension. The normalization methods used in the embodiments of this invention can all adopt this method. For example, linear normalization is a technical means well known to those skilled in the art, and will not be described in detail here.

[0063] Step S2: Obtain the growth segment of the feedforward index; based on the growth trend of each growth segment and the data value at each time point in the segment, obtain the growth coefficient corresponding to each time point; based on the growth coefficient of all feedforward indices at each time point, and combined with the correlation characteristics between various feedforward indices, obtain the pollution degree coefficient of the regulating pool at the next time point.

[0064] In the ammonia stripping and phenol removal process, increased temperature accelerates the cracking of polycyclic aromatic hydrocarbons and phenols in the coal tar, leading to higher COD and phenol concentrations in the wastewater. Simultaneously, increased sulfur content in the feedstock (coal tar) raises the concentration of thiols in the wastewater, potentially reducing the organic matter degradation rate and causing COD accumulation. This results in further increases in COD in the equalization tank. Furthermore, a significant increase in oil content in the equalization tank indicates the influx of high-concentration organic phases, further amplifying the pollutant load. Therefore, increases in these monitoring indicators lead to a faster increase in the severity of water pollution in the equalization tank, allowing for the prediction of the pollution level in the equalization tank based on these indicators.

[0065] In one embodiment of the present invention, the preset feedforward indicators include at least distillation temperature (ZT), raw material sulfur content (S), COD (C), and oil content (O).

[0066] Obtain the growth segment of the feedforward index to predict the pollution degree coefficient based on the growth of the monitored feedforward index.

[0067] In one embodiment of the present invention, the extreme points of the feedforward index within each preset period are obtained, and the interval between the minimum point and the next maximum point is taken as a growth segment. In order to eliminate the influence of short-term data fluctuations, the time domain length of the growth segment is limited to at least 3.

[0068] In other embodiments of the present invention, the implementer may also obtain the first-order difference of the data and take the data segment corresponding to two or more consecutive difference values ​​greater than 0 as a growth segment.

[0069] Using the same method, obtain all growth segments of all target data.

[0070] Considering that the growth situation varies in different growth segments and at different times within a growth segment, the growth coefficient corresponding to each time moment is obtained by combining the growth trend of each growth segment with the data value of each time moment in the segment. This quantifies the growth degree at each growth moment and provides a basis for subsequent prediction of pollution level by integrating multiple feedforward indicators.

[0071] Preferably, in one embodiment of the present invention, considering that the more first-order difference values ​​there are in the growth segment, the longer the duration of the growth segment; the larger the overall first-order difference values, the greater the growth rate; and the larger the data value at the selected time, the greater the impact of the continued growth of the data on the adjustment pool, the growth coefficient corresponding to each time moment is obtained based on the number of first-order difference values ​​in the growth segment, the overall characteristics of the first-order difference values, and the data value at each time moment in the segment.

[0072] As an example, considering the differences in data distribution patterns across different dimensions, the growth characteristics can be highlighted by the ratio of the data value to the mean of all historical data moments. Therefore, after normalizing the product of the ratio of the data value at each moment to the mean of all historical data moments in the same dimension, the number of first-order difference values ​​in the growth segment, and the mean of the first-order difference values, the normalized result is used as the growth coefficient for each corresponding moment.

[0073] By using the mean to represent the overall characteristics of the first-order difference, and by leveraging the number of first-order difference values ​​and the mean of the first-order difference values ​​in the growth segment, the growth trend of the growth segment can be shown from the perspectives of growth duration and growth rate.

[0074] It should be noted that the growth coefficient for non-growth periods is set to zero; same dimension refers to the same dimension, such as the temperature data dimension.

[0075] Considering the interrelationship among multiple feedforward indicators and the varying impacts of their growth on the equalization tank, a more comprehensive and mutually validated predictive model can be constructed by combining the growth coefficients of all feedforward indicators at each time point with the correlation characteristics among them. This allows for the fusion of the growth coefficients of multiple feedforward indicators to obtain the pollution level coefficient of the equalization tank at the next corresponding time point, enabling early prediction of system load. This helps identify potential overload risks in advance, facilitates subsequent regulation of influent flow, enables peak shaving measures to be taken before pollution increases, and improves treatment efficiency when pollution decreases, thereby enhancing the system's resilience and stability.

[0076] Preferably, in one embodiment of the present invention, any time point is selected as the target time point; considering that the Pearson correlation coefficient matrix characterizes the correlation between variables and reflects the "cooperative change" pattern between feedforward indicators, the correlation characteristics between various feedforward indicators are expressed by using the Pearson correlation coefficient matrix, the linear correlation structure between feedforward indicators is analyzed, and the characteristic patterns of the same or opposite changes between indicators are expressed; and the eigenvector corresponding to the largest eigenvalue in the eigenvector matrix (i.e., the principal eigenvector) can be regarded as representing the direction of the main change trend of the system, and each element therein represents the "weight" or "contribution" of the original variable (feedforward indicator) in the principal direction, which can be used as the weighted weight of the growth coefficient of each monitored feedforward indicator;

[0077] Specifically, the correlation characteristics among the various feedforward indicators refer to the linear correlation between the feedforward indicators and the weight of each indicator in the direction of the main trend of change.

[0078] Based on this, using all data from the target time to the start time of the preset period, the Pearson correlation coefficients of a single feedforward index and the other feedforward indices are calculated respectively, and the Pearson correlation coefficient matrix between each feedforward index at the target time and the feature vector matrix are obtained.

[0079] Based on the vector elements corresponding to the eigenvectors of each feedforward index in the eigenvectors corresponding to the largest eigenvalue, the growth coefficients of each feedforward index at the target time are weighted and summed, and the weighted sum is used as the pollution degree coefficient for the next time step at the target time.

[0080] As an example, the vector elements corresponding to the eigenvectors of the largest eigenvalues ​​of each feedforward index are mapped using the softmax function. The mapped values ​​are then used as the weights of each feedforward index. Based on these weights, the growth coefficients at the target time are summed in a weighted manner, and the result is used as the pollution level coefficient for the next time step. This adaptively integrates the growth trends of each pollution feedforward index to accurately predict the future pollution level.

[0081] It should be noted that the Pearson correlation coefficient, the construction of the Pearson correlation coefficient matrix, the acquisition of the eigenvector matrix, the acquisition of the maximum eigenvalue, and the softmax function are all existing technologies and will not be elaborated upon here.

[0082] Iterate through all historical and real-time data to obtain the pollution level coefficient for the next time step at each time step.

[0083] Step S3: Within each historical preset period, divide the data of pollution degree coefficient and various feedforward indicators into time series based on adjacent extreme points to obtain fluctuation segments for each; based on the fluctuation amplitude and salience within each fluctuation segment, obtain the abnormal index and screen out the abnormal fluctuation segments; based on the time interval between the abnormal fluctuation segment of the pollution degree coefficient and the abnormal fluctuation segment of its subsequent feedforward indicator, and combined with the similarity of the abnormal index, obtain the response fluctuation segment corresponding to the abnormal fluctuation segment of the pollution degree coefficient.

[0084] Step S2 was used to predict the degree of wastewater in the equalization tank, and a pollution degree coefficient was given for each time point. The subsequent operation steps of the equalization tank include biochemical reactions, which have a lag in response to the wastewater concentration in the equalization tank. That is, when the wastewater concentration in the equalization tank changes, the monitoring indicators in the subsequent operation steps will take a long time to produce corresponding fluctuations in the wastewater concentration.

[0085] In one embodiment of the present invention, the preset feedback indicators include at least the COD content, sludge concentration, dissolved oxygen concentration in the biochemical unit, and the residual amount of oxidant in the advanced treatment unit.

[0086] Considering that excessive increases or decreases in wastewater concentration within the equalization tank are detrimental to maintaining stable concentration, abnormal fluctuation data segments of wastewater pollution levels within a single cycle are screened. Furthermore, abnormal fluctuations in pollution levels will also cause abnormal fluctuations in the data of feedforward indicators, facilitating subsequent analysis of the response of feedforward indicators to abnormal fluctuations in pollution level coefficients. Abnormal fluctuation data segments are also divided and screened for each feedforward indicator.

[0087] Since historical data is already established, which facilitates the analysis of the response of feedforward indicators to anomalies in the pollution degree coefficient, the analysis is conducted within a preset historical period. Considering that the data fluctuation trends between adjacent extreme points are consistent, within each preset period, the data of the pollution degree coefficient and various feedforward indicators are divided into time series based on adjacent extreme points, and fluctuation segments are obtained for each.

[0088] It should be noted that the method of obtaining the extreme points of time series data sequences is a well-known technique in the art. Implementers can also limit the minimum lower limit of the fluctuation segment, such as limiting the length of the fluctuation segment to a minimum of 3, to reduce the impact of short-term fluctuations caused by noise. This will not be elaborated further here.

[0089] Considering that the fluctuation amplitude and salience characteristics of the fluctuation segment both reflect the instability of the fluctuation segment and show abnormal characteristics, the abnormal index is obtained and abnormal fluctuation segments are screened according to the fluctuation amplitude and salience within each fluctuation segment. The pollution degree coefficient segmentation of the abnormal fluctuation and the segmentation of various feedforward index data are determined to facilitate subsequent analysis of the response of various feedforward indexes to the abnormal pollution degree coefficient.

[0090] Preferably, in one embodiment of the present invention, considering that the larger the overall fluctuation amplitude of a fluctuation segment, the greater the increase or decrease, the more unstable it is, and the higher the anomaly index; at the same time, the greater the difference between the data within the fluctuation segment and the conventional pattern of historical data in the same dimension, the more obvious the abnormal characteristics are, and the higher the anomaly index is.

[0091] Based on this, the abnormal index of the corresponding fluctuation segment is obtained by combining the overall difference between the fluctuation segment data and all historical data in the same dimension, according to the overall fluctuation amplitude of the fluctuation segment; the overall fluctuation amplitude is positively correlated with the abnormal index.

[0092] As an example, since the fluctuation segment is divided by adjacent extreme points, the data within the fluctuation segment is monotonous. The absolute value of the difference between the data values ​​at the beginning and end of the fluctuation segment is taken as the overall fluctuation amplitude of the fluctuation segment, representing the fluctuation amplitude within the fluctuation segment. The absolute value of the difference between the mean of the data value of the fluctuation segment and the mean of all historical data in the same dimension is taken as the numerator, and the mean of all historical data in the same dimension is taken as the denominator. The ratio of the fractions is taken as the fluctuation salience coefficient. With the mean of all historical data in the same dimension as the benchmark, it represents the regular pattern of the data in the same dimension, thus reflecting the overall difference between the data of the fluctuation segment and all historical data in the same dimension, showing the fluctuation salience within the fluctuation segment.

[0093] The product of the entire fluctuation range and the fluctuation prominence coefficient is mapped using linear normalization, and the mapped value is used as the anomaly index of the corresponding fluctuation range.

[0094] Considering that the larger the abnormality index, the more obvious the abnormal characteristics of the fluctuation segment, and the more likely it is to be an abnormal fluctuation segment, the abnormal fluctuation threshold is set to 0.5 (empirical value). Fluctuation segments with an abnormality index greater than the abnormal fluctuation threshold are marked as abnormal fluctuation segments, and abnormal fluctuation segments are screened out.

[0095] The fluctuation ranges of pollution degree coefficients and various feedback indicators within all preset periods are analyzed to identify abnormal fluctuation ranges in each data dimension.

[0096] It should be noted that the normalization of the anomaly index for each type of data is performed within its respective data dimension; in other embodiments of the present invention, the implementer may also use the sigmoid function for normalization and adjust the anomaly fluctuation threshold, such as 0.8.

[0097] Considering that abnormal fluctuations in the pollution degree coefficient can affect subsequent feedforward indices in the time domain, causing similar abnormal fluctuations in the feedforward indices, and that the time interval also reflects the possibility that the abnormal fluctuations in the feedforward indices are caused by abnormal fluctuations in the pollution degree coefficient, we obtain the response fluctuation segment corresponding to the abnormal fluctuation segment of the pollution degree coefficient based on the time interval between the abnormal fluctuation segment of the pollution degree coefficient and the abnormal fluctuation segment of the subsequent feedforward indices, combined with the similarity of the abnormal indices. Matching the abnormal fluctuation segment of the pollution degree coefficient with the abnormal fluctuation segment of the feedforward indices facilitates the analysis of the impact characteristics of pollution degree on the feedforward indices, and provides a basis for subsequent prediction and adjustment of the influent flow rate of high-concentration wastewater to maintain the wastewater treatment effect at each stage.

[0098] Preferably, in one embodiment of the present invention, any feedback indicator is selected as the feedback target indicator, and any abnormal fluctuation segment of the pollution degree coefficient is selected as the target fluctuation segment, so as to facilitate the comparison and matching of the pollution degree coefficient with each feedback indicator dimension one by one.

[0099] Within the preset period to which the target fluctuation segment belongs, and in the time domain after the start time of the target fluctuation segment, each abnormal fluctuation segment of the feedback target indicator is taken as the fluctuation segment to be analyzed. First, the abnormal fluctuation segments in the target feedback indicators that can be compared are determined.

[0100] Considering that the closer the start time of the fluctuation segment to be analyzed is to the target fluctuation segment, the stronger the causal relationship is to the target fluctuation segment, and the more similar the anomaly index is to the target fluctuation segment, the matching coefficient corresponding to each fluctuation segment to be analyzed is obtained based on the time interval between the start time of each fluctuation segment to be analyzed and the target fluctuation segment, combined with the similarity of the anomaly index; the fluctuation segment to be analyzed with the largest matching coefficient is selected as the response fluctuation segment of the target fluctuation segment.

[0101] As an example, the reciprocal of the time interval between the start time of the fluctuation segment to be analyzed and the target fluctuation segment is used as the first matching factor; the absolute value of the difference between the abnormal index of the fluctuation segment to be analyzed and the target fluctuation segment is used as the numerator, the abnormal index of the target fluctuation segment is used as the denominator, the fractional ratio is used as the abnormal difference factor, and the reciprocal of the sum of the abnormal difference factor and a preset positive parameter divided by zero, such as 0.01, is used as the second matching factor.

[0102] The product of the first matching factor and the second matching factor is used as the matching coefficient corresponding to the fluctuation segment to be analyzed; the fluctuation segment to be analyzed with the largest matching coefficient is selected as the response fluctuation segment of the target fluctuation segment.

[0103] Since the fluctuations in the pollution degree coefficient are transmitted to subsequent wastewater treatment, the fluctuations in the feedforward index require time to propagate. Therefore, the starting time of the fluctuation segment to be analyzed must be after the starting time of the target fluctuation segment, so the time interval must be greater than zero. Considering that the abnormal indices may be the same, to prevent the denominator from being zero, a division operation is performed before taking the reciprocal. Using the abnormal index of the target fluctuation segment as a benchmark, the quantitative difference characteristics of the abnormal difference factor are obtained, and then negative correlation mapping is performed by taking the reciprocal to show the similarity of the abnormal indices, and finally the response fluctuation segment is obtained.

[0104] By changing the target feedback index, the response fluctuation segments of each feedback index corresponding to the target fluctuation segment are obtained; by traversing all abnormal fluctuation segments of the pollution degree coefficient in all historical preset periods, all response fluctuation segments are obtained.

[0105] Step S4: Based on the overall response characteristics of the pollution degree coefficient to the response fluctuations of various feedforward indicators in historical data, and combined with the changing trend of the pollution degree coefficient at the current moment, obtain the influent disturbance index at the current moment; adjust the influent flow rate of high-concentration wastewater based on the influent disturbance index at the current moment and the pollution degree coefficient at the next moment.

[0106] Considering that the changing trend of the pollution degree coefficient often causes fluctuations in the response of feedforward indicators in the time domain, by analyzing the abnormal fluctuations of the pollution degree coefficient in historical data and the resulting fluctuations in the response of feedforward indicators after time shift, typical behavioral patterns of pollution sources in response to system feedback can be extracted. Based on this response pattern, combined with the changing trend of the pollution degree coefficient at the current moment, the degree of disturbance risk currently faced by the system can be predicted.

[0107] Therefore, based on the response fluctuations of various feedforward indicators in historical data to the overall response characteristics of the pollution degree coefficient, and combined with the changing trend of the pollution degree coefficient at the current moment, the influent disturbance index at the current moment can be obtained, thereby enhancing the ability to identify system disturbance risks in advance and facilitating timely adjustment of the influent flow rate of high-concentration wastewater.

[0108] Preferably, in one embodiment of the present invention, considering that the pollution degree coefficient is mainly divided into two cases of abnormal increase and abnormal decrease, and different changes may have different effects on the feedforward index, based on the increase or decrease of the abnormal fluctuation segment of the pollution degree coefficient, all response fluctuation segments are divided into two categories and analyzed accordingly; one category is the increase type response fluctuation segment, and the other category is the decrease type response fluctuation segment.

[0109] Considering the overall characteristics of the time interval between the start time of a type of response fluctuation segment and the corresponding abnormal fluctuation segment of the pollution degree coefficient, it reflects the response time lag of the pollution disturbance in the system, that is, the time required for the overall response of various feedforward indicators; and combining the overall characteristics of the time interval between the end time of a type of response fluctuation segment and the start time of the corresponding abnormal fluctuation segment of the pollution degree coefficient, it reflects the time required for the overall response of various feedforward indicators to end, which can effectively estimate the predicted response duration of this type of disturbance in the feedforward indicators.

[0110] Based on this, for each type of response fluctuation segment, the predicted response duration is obtained by combining the time interval between the start time of a type of response fluctuation segment and the start time of the corresponding abnormal fluctuation segment of the pollution degree coefficient, and the time interval between the end time of a type of response fluctuation segment and the start time of the corresponding abnormal fluctuation segment of the pollution degree coefficient.

[0111] As an example, the average time interval between the start time of a type of response fluctuation segment and the start time of the corresponding abnormal fluctuation segment of the pollution degree coefficient is used as the predicted response start time; the average time interval between the end time of a type of response fluctuation segment and the start time of the corresponding abnormal fluctuation segment of the pollution degree coefficient is used as the predicted response end time; and the difference between the predicted response end time and the predicted response start time is used as the predicted response duration of this type of response fluctuation segment.

[0112] Considering that the larger the fluctuation amplitude of the first-class response fluctuation segment, the more sensitive it is to changes in the pollution degree coefficient; at the same time, the more concentrated the distribution of the fluctuation amplitude of the first-class response fluctuation segment, the more stable the trend of change, and the greater the influence of the pollution degree coefficient on the feedforward index; and the longer the predicted response duration, the longer the influence of the pollution degree coefficient on the feedforward index lasts, and the stronger the interference, the influent interference index of the first-class response fluctuation segment is obtained based on the distribution characteristics of the fluctuation amplitude of the first-class response fluctuation segment and the predicted response duration.

[0113] As an example, the product of the mean amplitude of a type of response fluctuation segment and the predicted response duration is used as the numerator, the range of the amplitude of such response fluctuation segments is used as the denominator, and the ratio of the fractions is used as the inflow disturbance index for such response fluctuation segments.

[0114] The fluctuation amplitude of each fluctuation segment is the absolute value of the difference between the data values ​​at the beginning and end of the fluctuation segment. The mean of the fluctuation amplitude represents the sensitivity of all feedback indicators to changes in a certain level of pollution. The range represents the discrete characteristics of the fluctuation amplitude. The smaller the range, the more stable the changes in all fluctuation segments, the stronger the interference of the pollution degree coefficient, and the larger the influent interference index. Combined with the predicted response duration, which represents the expected response duration of the feedback indicators, they together show the overall response characteristics of the feedback indicator's response fluctuation segment to the pollution degree coefficient.

[0115] Finally, based on the current trend of the pollution level coefficient, the corresponding influent disturbance index is selected.

[0116] It should be noted that, in one embodiment of the present invention, the increasing or decreasing trend of the pollution degree coefficient is determined based on the slope of the pollution degree coefficient at the current moment: the time series curve of the pollution degree coefficient for the current day is obtained, and when the slope of the time series curve at the current moment is greater than zero, the pollution degree coefficient at the current moment shows an increasing trend; when the slope of the time series curve at the current moment is less than zero, the pollution degree coefficient at the current moment shows a decreasing trend.

[0117] If the slope of the current time on the time-series curve is zero, the pollution level is considered stable, the current state of the system is stable, and no adjustment of the influent flow rate is required.

[0118] Considering that the current influent disturbance index represents the interference of the current influent state on wastewater treatment, and the pollution degree coefficient at the next moment represents the pollution degree that the system will face, the influent flow rate of high-concentration wastewater can be adjusted based on the current influent disturbance index and the pollution degree coefficient at the next moment. This improves the system's forward-looking and proactive regulation, reduces the impact of regulation on feedback indicators, and ensures the wastewater treatment effect.

[0119] Preferably, in one embodiment of the present invention, when considering that the pollution degree coefficient at the current moment is increasing, it is necessary to reduce the influent flow rate of high-concentration wastewater to avoid excessive load on subsequent biochemical and physicochemical steps and affect the wastewater treatment effect; at the same time, the larger the pollution degree coefficient predicted for the next moment, the greater the water purification pressure faced by the system, the smaller the influent flow rate needs to be, and the greater the reduction in influent flow rate should be.

[0120] Considering that the pollution level coefficient is decreasing at the current moment, it is necessary to increase the influent flow rate of high-concentration wastewater to maintain the treatment efficiency of subsequent biological and physicochemical steps and improve the overall wastewater treatment efficiency. At the same time, the higher the pollution level coefficient is predicted to be at the next moment, the greater the water purification pressure faced by the system, the smaller the influent flow rate needs to be, and the smaller the increase in the influent flow rate should be.

[0121] Based on this, if the pollution level coefficient at the current moment shows an increasing trend, the adjustment factor is obtained by combining the influent disturbance index at the current moment and the pollution level coefficient at the next moment; both the influent disturbance index and the pollution level coefficient at the next moment are negatively correlated with the adjustment factor.

[0122] If the pollution level coefficient at the current moment shows a decreasing trend, the adjustment factor is obtained by combining the influent disturbance index at the current moment and the pollution level coefficient at the next moment; the influent disturbance index is positively correlated with the adjustment factor; the pollution level coefficient at the next moment is negatively correlated with the adjustment factor.

[0123] The influent flow rate at the current moment is adjusted based on the adjustment factor.

[0124] As an example, the formula for calculating the adjusted influent flow rate of high-concentration wastewater includes:

[0125] ;

[0126] in, This represents the influent flow rate of the high-concentration wastewater before adjustment at the current i-th time point; The influent flow rate of the high-concentration wastewater after adjustment at the current time i; This represents the water inflow disturbance index at the current i-th moment; This represents the pollution level coefficient for the next time step after the current time step i. The slope of the time series curve representing the pollution level coefficient at the current i-th moment; This represents the linear normalization function.

[0127] In the formula, At that time, the influent disturbance index at the current moment and the pollution degree coefficient at the next moment are combined by addition. This is done by first adding, then normalizing, and finally subtracting the normalized value from the constant 1. Perform negative correlation mapping and normalization to obtain the regulation factor. Multiply and combine with the influent flow rate before adjustment at the current moment to reduce the influent flow rate and meet the adjustment logic when the pollution degree coefficient shows an increasing trend;

[0128] exist At that time, the influent disturbance index at the current moment and the pollution degree coefficient at the next moment are combined by subtraction. The adjustment factor is obtained by first subtracting, then normalizing, and finally adding a constant of 1. The influent flow rate is multiplied and combined with the influent flow rate before adjustment at the current moment to increase the influent flow rate and meet the adjustment logic when the pollution degree coefficient shows a decreasing trend.

[0129] It should be noted that in the calculation of adjusting the influent flow rate, because the calculation methods for the adjustment factors for increasing and decreasing pollution degree coefficients differ, two data dimensions are used during normalization, namely... One data dimension One data dimension.

[0130] In other embodiments of the present invention, the implementer may modify... and Normalization was performed separately before merging, so that... and To ensure consistency in the impact of adjustment factors and avoid influences based on orders of magnitude and dimensions, implementers can also add adjustment range coefficients to control the adjustment range of a single influent flow rate, such as... , To adjust the range coefficient, The adjustment range is then controlled within 80% to 120%.

[0131] In one embodiment of the present invention, The predicted influent flow rate of high-concentration wastewater for the next moment is transmitted to the PLC control system. Based on the control signal output by the PLC control system to the electric regulating valve, the influent flow rate of high-concentration wastewater for the next moment is pre-regulated by the electric regulating valve. The influent flow rate is stopped when the remaining capacity of the regulating tank is less than 10%.

[0132] The system also outputs the predicted influent flow rate of high-concentration wastewater at the next moment, as well as the measured values ​​of each indicator under the current condition.

[0133] One embodiment of the present invention also provides a wastewater treatment system in the deep processing of coal tar; please refer to [link to relevant documentation]. Figure 3The diagram illustrates a system block diagram of a wastewater treatment system in the deep processing of coal tar provided by an embodiment of the present invention, which specifically includes: a data acquisition module 101, a data prediction module 102, a response matching module 103, and an influent regulation module 104.

[0134] Data acquisition module 101: acquires historical and real-time data of preset feedforward and feedback indicators during the wastewater treatment process, and divides them into preset periods;

[0135] Data prediction module 102: Obtain the growth segment of the feedforward indicator; based on the growth trend of each growth segment and the data value of each moment in the segment, obtain the growth coefficient corresponding to each moment; based on the growth coefficient of all feedforward indicators at each moment, and combined with the correlation characteristics between various feedforward indicators, obtain the pollution degree coefficient of the regulating pool at the next moment.

[0136] Response matching module 103: Within each historical preset period, the data of pollution degree coefficient and various feedforward indicators are divided into time series based on adjacent extreme points to obtain fluctuation segments for each; based on the fluctuation amplitude and salience within each fluctuation segment, anomaly indices are obtained and abnormal fluctuation segments are selected; based on the time interval between the abnormal fluctuation segments of the pollution degree coefficient and the abnormal fluctuation segments of the subsequent feedforward indicators, combined with the similarity of the abnormal indices, the response fluctuation segment corresponding to the abnormal fluctuation segment of the pollution degree coefficient is obtained.

[0137] Influent regulation module 104: Based on the overall response characteristics of the pollution degree coefficient to the response fluctuations of various feedforward indicators in historical data, and combined with the changing trend of the pollution degree coefficient at the current moment, obtain the influent disturbance index at the current moment; and adjust the influent flow rate of high-concentration wastewater based on the influent disturbance index at the current moment and the pollution degree coefficient at the next moment.

[0138] The implementation of all modules of a wastewater treatment system in a coal tar deep processing process has been described in the wastewater treatment method in a coal tar deep processing process described in steps S1-S4, and will not be repeated here.

[0139] An embodiment of the present invention also provides a wastewater treatment device in the deep processing of coal tar. The device includes a memory, a processor, and a computer program. The memory is used to store the corresponding computer program, and the processor is used to run the corresponding computer program. When the computer program runs in the processor, it can implement the wastewater treatment method in the deep processing of coal tar described in steps S1-S4.

[0140] In summary, addressing the technical problem that existing water quality regulation methods struggle to cope with frequent changes in coal tar wastewater concentration, leading to regulation anomalies and impacting wastewater treatment efficiency, this invention proposes a wastewater treatment method, apparatus, and system for the deep processing of coal tar. This invention first obtains historical and real-time data for various indicators; then, based on changes in feedforward indicators, it predicts the pollution degree coefficient of the regulating tank at the next corresponding moment; further, it identifies abnormal fluctuation segments in the pollution degree coefficient and various feedforward indicators; further, it obtains the response fluctuation segments corresponding to the abnormal fluctuation segments of the pollution degree coefficient; finally, based on the overall response characteristics of the response fluctuation segments of various feedforward indicators in historical data to the pollution degree coefficient, combined with the current trend of the pollution degree coefficient and the pollution degree coefficient at the next moment, it adjusts the influent flow rate of high-concentration wastewater. This invention predicts the wastewater concentration in the regulating tank using feedforward indicators and, combined with the response characteristics of feedforward indicators, predicts the impact of wastewater concentration on wastewater treatment. Therefore, it adjusts the influent flow rate based on current wastewater concentration changes, dynamically predicts system load, precisely controls the influent flow rate, and effectively improves wastewater treatment efficiency and system stability.

[0141] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0142] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A wastewater treatment method in the deep processing of coal tar, characterized in that, The methods include: Historical and real-time data of preset feedforward and feedback indicators in the wastewater treatment process are acquired and divided into preset periods. The feedforward indicators include distillation temperature, raw material sulfur content, COD, and oil content. The feedback indicators include COD content, sludge concentration, dissolved oxygen concentration in the biochemical unit, and residual oxidant in the advanced treatment unit. Traverse all historical and real-time data at each moment: obtain the growth segment of the feedforward indicator; based on the growth trend of each growth segment and the data value of each moment in the segment, obtain the growth coefficient corresponding to each moment; based on the growth coefficient of all feedforward indicators at each moment, and combined with the correlation characteristics between various feedforward indicators, obtain the pollution degree coefficient of the regulating pool at the next moment. Within each historical preset period, the data of pollution degree coefficient and various feedforward indicators are divided into time series based on adjacent extreme points, and fluctuation segments are obtained for each. Based on the fluctuation amplitude and salience within each fluctuation segment, anomaly indices are obtained and abnormal fluctuation segments are screened out. Based on the time interval between the abnormal fluctuation segments of the pollution degree coefficient and the abnormal fluctuation segments of the subsequent feedforward indicators, combined with the similarity of the abnormal indices, the response fluctuation segments corresponding to the abnormal fluctuation segments of the pollution degree coefficient are obtained. Based on the response fluctuations of various feedforward indicators in historical data and their overall response characteristics to the pollution degree coefficient, combined with the current trend of the pollution degree coefficient, the influent disturbance index at the current moment is obtained. Adjust the influent flow rate of high-concentration wastewater based on the influent disturbance index at the current moment and the pollution degree coefficient at the next moment; The method for obtaining the response fluctuation range includes: Select any of the aforementioned feedback indicators as the feedback target indicator, select any of the aforementioned pollution degree coefficients' abnormal fluctuation segments as the target fluctuation segments, and within the preset period to which the target fluctuation segment belongs, and in the time domain after the start time of the target fluctuation segment, select each abnormal fluctuation segment of the feedback target indicator as a fluctuation segment to be analyzed; based on the time interval between the start time of each fluctuation segment to be analyzed and the target fluctuation segment, and combined with the similarity of the abnormal index, obtain the matching coefficient corresponding to each fluctuation segment to be analyzed; select the fluctuation segment to be analyzed with the largest matching coefficient as the response fluctuation segment of the target fluctuation segment.

2. The wastewater treatment method in the deep processing of coal tar according to claim 1, characterized in that, The method for obtaining the growth coefficient includes: Based on the number of first-order difference values ​​in the growth segment, the overall characteristics of the first-order difference values, and the data value at each time step in the segment, the growth coefficient corresponding to each time step is obtained.

3. The wastewater treatment method in the deep processing of coal tar according to claim 1, characterized in that, The method for obtaining the pollution degree coefficient includes: Select any time as the target time; based on all data from the target time to the start time of the preset period to which it belongs, obtain the Pearson correlation coefficient matrix between each feedforward index at the target time and obtain the feature vector matrix; Based on the vector elements corresponding to the eigenvectors of each feedforward index at the maximum eigenvalue, the growth coefficients of each feedforward index at the target time are weighted and summed, and the weighted sum is used as the pollution degree coefficient at the next time step at the target time.

4. The wastewater treatment method in the deep processing of coal tar according to claim 1, characterized in that, The method for obtaining the anomaly index includes: Based on the overall fluctuation amplitude of the fluctuation segment and the overall difference between the data of the fluctuation segment and all historical data in the same dimension, an anomaly index corresponding to the fluctuation segment is obtained; the overall fluctuation amplitude is positively correlated with the anomaly index.

5. A wastewater treatment method in the deep processing of coal tar according to claim 1, characterized in that, The method for obtaining the water inflow disturbance index includes: Based on the increase or decrease of the abnormal fluctuation segment of the pollution degree coefficient, all the response fluctuation segments are divided into two categories: increase and decrease. For each type of response fluctuation segment, the predicted response duration is obtained based on the time interval between the start time of each type of response fluctuation segment and the abnormal fluctuation segment of the corresponding pollution degree coefficient, combined with the time interval between the end time of each type of response fluctuation segment and the start time of the abnormal fluctuation segment of the corresponding pollution degree coefficient. Based on the distribution characteristics of the fluctuation amplitude of each type of response fluctuation segment, and combined with the predicted response duration, the inflow disturbance index of each type of response fluctuation segment is obtained. Based on the current trend of the pollution level coefficient, the corresponding influent disturbance index is selected.

6. A wastewater treatment method in the deep processing of coal tar according to claim 1, characterized in that, The method for obtaining the influent flow rate includes: If the pollution level coefficient at the current moment shows an increasing trend, an adjustment factor is obtained by combining the influent disturbance index at the current moment and the pollution level coefficient at the next moment; both the influent disturbance index and the pollution level coefficient at the next moment are negatively correlated with the adjustment factor. If the pollution level coefficient at the current moment shows a decreasing trend, an adjustment factor is obtained by combining the influent disturbance index at the current moment and the pollution level coefficient at the next moment; the influent disturbance index is positively correlated with the adjustment factor; the pollution level coefficient at the next moment is negatively correlated with the adjustment factor. The influent flow rate at the current moment is adjusted based on the adjustment factor.

7. A wastewater treatment method in the deep processing of coal tar according to claim 1, characterized in that, The time domain length of the growth segment is at least 3.

8. A wastewater treatment system for the deep processing of coal tar, characterized in that, The system includes: Data acquisition module: acquires historical and real-time data of preset feedforward and feedback indicators during the wastewater treatment process, and divides them according to preset periods; the feedforward indicators include distillation temperature, raw material sulfur content, COD and oil content; the feedback indicators include COD content, sludge concentration, dissolved oxygen concentration in the biochemical unit and residual oxidant in the advanced treatment unit. Data prediction module: Iterates through all historical and real-time data at each moment: obtains the growth segment of the feedforward indicator; based on the growth trend of each growth segment and the data value of each moment in the segment, obtains the growth coefficient corresponding to each moment; based on the growth coefficient of all feedforward indicators at each moment, combined with the correlation characteristics between various feedforward indicators, obtains the pollution degree coefficient of the regulating pool at the next moment. Response matching module: Within each historical preset period, the data of pollution degree coefficient and various feedforward indicators are divided into time series based on adjacent extreme points, and fluctuation segments are obtained for each. Based on the fluctuation amplitude and salience within each fluctuation segment, anomaly indices are obtained and abnormal fluctuation segments are screened out. Based on the time interval between the abnormal fluctuation segments of the pollution degree coefficient and the abnormal fluctuation segments of the subsequent feedforward indicators, combined with the similarity of the abnormal indices, the response fluctuation segments corresponding to the abnormal fluctuation segments of the pollution degree coefficient are obtained. Influent regulation module: Based on the overall response characteristics of the pollution degree coefficient to the response fluctuation range of various feedforward indicators in historical data, and combined with the current trend of the pollution degree coefficient, the influent disturbance index at the current moment is obtained; the influent flow rate of high-concentration wastewater is adjusted based on the influent disturbance index at the current moment and the pollution degree coefficient at the next moment. The method for obtaining the response fluctuation range includes: Select any of the aforementioned feedback indicators as the feedback target indicator, select any of the aforementioned pollution degree coefficients' abnormal fluctuation segments as the target fluctuation segments, and within the preset period to which the target fluctuation segment belongs, and in the time domain after the start time of the target fluctuation segment, select each abnormal fluctuation segment of the feedback target indicator as a fluctuation segment to be analyzed; based on the time interval between the start time of each fluctuation segment to be analyzed and the target fluctuation segment, and combined with the similarity of the abnormal index, obtain the matching coefficient corresponding to each fluctuation segment to be analyzed; select the fluctuation segment to be analyzed with the largest matching coefficient as the response fluctuation segment of the target fluctuation segment.

9. A wastewater treatment device for deep processing of coal tar, the device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the wastewater treatment method in the deep processing of coal tar as described in any one of claims 1 to 7.

Images