A smart dosing control method and system for sewage treatment plants based on carbon neutrality

Abstract

The present invention relates to the field of intelligent control technology, and specifically to an intelligent dosing control method and system for a sewage treatment plant based on carbon neutrality. The method comprises: collecting a pH value data sequence, taking the average of the noise levels of all extreme points between any two extreme points as the noise level between any two extreme points, setting the initial interpolation density of the extreme point sequence, adaptively adjusting the initial interpolation density using the noise level to obtain the interpolation density between any two extreme points, performing interpolation fitting between any two extreme points using the adjusted interpolation density to generate upper and lower envelopes, performing subsequent denoising based on the envelope to obtain a denoised data sequence, predicting future pH values ​​based on the denoised data sequence, and controlling the input amount of a pH regulator, thereby improving the denoising effect. Based on the high-quality denoised data, the system can provide more accurate pH value prediction and control of the input amount of a reagent.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control technology, and more particularly to an intelligent dosing control method and system for a sewage treatment plant based on carbon neutrality. Background Art

[0002] In the sewage treatment process, the precise addition of pH adjusters (such as alkaline sodium hydroxide or acidic sulfuric acid) is the key to ensuring treatment efficiency and effectiveness. The coagulation process is directly affected by the pH value, because different pH values ​​will change the state and activity of the coagulant hydrolysis products. Only within the appropriate pH range can these products most effectively promote the coagulation and sedimentation of suspended particles. Therefore, real-time monitoring and purification of pH data to ensure its stability in the optimal working range is crucial to improving treatment quality and reducing chemical consumption. It also helps to reduce carbon emissions in the chemical treatment process and promote the realization of carbon neutrality goals.

[0003] In the process of real-time monitoring of pH value, due to the continuous stirring caused by the stirring effect, stirring will increase the contact between the liquid surface and the air, which may cause gases to dissolve in the water, especially oxygen and carbon dioxide. The dissolution of carbon dioxide will affect the pH value of water because it can react with water to form carbonic acid, thereby lowering the pH value, or because of the turbulence effect caused by stirring, certain chemicals diffuse and react faster in a turbulent environment, affecting the uniformity and stability of the pH value. Stirring will redistribute the solid particles suspended in the water. These particles may contain acidic or alkaline substances. When they are redispersed during stirring, they may affect the local pH value. These situations will cause the collected pH data to be inaccurate and have random noise.

[0004] The empirical mode decomposition algorithm (EMD) is usually used for data denoising. For example, the authorized patent document with the existing announcement number CN117368141B discloses an intelligent detection method for perchlorate wastewater concentration based on artificial intelligence, including: collecting the spectral curve of perchlorate wastewater and decomposing it into several IMF (intrinsic mode component) component curves, obtaining the fluctuation coefficient of each component curve according to the data difference between the curves, dividing all the curves into several non-repetitive curve segments, and obtaining the total fluctuation ratio difference of each component curve according to the difference in absorbance between the curve segments, thereby obtaining the optimal wavelet threshold, using the wavelet threshold denoising algorithm to obtain the denoised curve, and then using the EMD reconstruction algorithm to obtain the filtered denoising curve of the spectral curve.

[0005] However, in the process of denoising the pH value, the traditional empirical mode decomposition algorithm needs to interpolate and fit the upper and lower envelopes through the maximum and minimum points respectively to obtain the IMF components, and reconstruct each intrinsic mode component (IMF) after denoising to achieve the effect of denoising the pH value data. When fitting the envelope, the traditional interpolation method usually needs to set a fixed number of interpolation points in the entire interpolation interval or interpolate at fixed intervals. This rigid interpolation method cannot adaptively adjust the interpolation density according to the characteristics of the local signal, which easily leads to unstable or inaccurate envelopes after interpolation, thereby failing to accurately extract intrinsic mode components, reducing the denoising accuracy of the empirical mode decomposition algorithm, and further affecting the accuracy of dosing control. Summary of the invention

[0006] In order to solve the problem that when the empirical mode decomposition algorithm processes the real-time collected pH value data, the rigid interpolation method easily leads to unstable or inaccurate envelope after interpolation, reduces the denoising accuracy of the empirical mode decomposition algorithm, and affects the accuracy of dosing control, the present invention proposes a smart dosing control method and system for sewage treatment plants based on carbon neutrality, which specifically adopts the following scheme:

[0007] On the one hand, the present invention provides a smart dosing control method for a sewage treatment plant based on carbon neutrality, comprising:

[0008] Collecting a pH value data sequence, performing data fitting on data points within a local range of each data point in the data sequence, and taking a mean value of fitting loss as a noise level of the data point;

[0009] Obtaining an extreme value point sequence in the data sequence, and taking the average of the noise levels of all extreme value points between any two extreme value points in the extreme value point sequence as the noise level between any two extreme value points;

[0010] The inverse of the average time interval between all extreme value points of the extreme value point sequence is used as the initial interpolation density of the extreme value point sequence, and the initial interpolation density is adaptively adjusted using the noise level to obtain the interpolation density between any two extreme value points, and the interpolation density between any two extreme value points is negatively correlated with the noise level;

[0011] The EMD algorithm is used to perform interpolation fitting between any two extreme points based on the adjusted interpolation density, and denoising is performed after the interpolation fitting to obtain a denoised data sequence. The future pH value is predicted based on the denoised data sequence, and the input amount of the pH regulator is controlled.

[0012] The above technical solution can quantify the degree of noise by collecting the pH value data sequence and performing data fitting on the local range of each data point, which helps to more accurately identify and process outliers or noise. The analysis of the extreme point sequence and the evaluation of the noise level allow the system to adaptively adjust the interpolation density, so that the system can better cope with various operating conditions, including water quality fluctuations, load changes and external interference, enhance the stability and robustness of the system, and can more finely capture the characteristics of the signal to avoid data distortion caused by fixed interpolation density. The adjusted interpolation density is negatively correlated with the noise level, which means that the interpolation points are sparser in areas with greater noise, reducing the error introduced by overfitting and improving the fitting accuracy of the envelope of the EMD algorithm, so that the intrinsic modal components can be accurately extracted, and the subsequent denoising and intrinsic modal component reconstruction process can effectively improve the denoising effect and ensure data quality, and then provide accurate guidance for the dosage of the agent based on the denoised data.

[0013] Further, the initial interpolation density is adaptively adjusted using the noise level to obtain the interpolation density between any two extreme value points, including:

[0014]

[0015] In the formula, For the The extreme point and The interpolation density between extreme points, is the initial interpolation density, For the The extreme point and The noise level between the extreme points, is a natural exponential function.

[0016] The above technical solution ensures that a lower interpolation density is used in data sections with higher noise, thereby effectively avoiding overfitting and improving the generalization ability of the model. The natural exponential function in the formula serves as a decay factor, which smoothly reduces the number of interpolation points in high-noise areas. This not only improves the efficiency of data processing, but also ensures the true reflection of signal characteristics and enhances the accuracy of prediction. The method of dynamically adjusting the interpolation density enhances the flexibility and adaptability of the algorithm, enabling it to better cope with data fluctuations under different conditions, which is of great significance for maintaining stable control of pH values ​​in sewage treatment processes.

[0017] Furthermore, predicting the future pH value according to the denoised data sequence and controlling the input amount of the pH regulator includes:

[0018] Using the denoised data sequence and the historical input amount of pH regulator as input of a generalized predictive control algorithm;

[0019] The generalized predictive control algorithm predicts the future pH value based on the input content;

[0020] The functional relationship between the historical input amount of the pH adjuster and the corresponding pH value is analyzed by a statistical method, and the input amount of the pH adjuster required for the future pH value is determined based on the functional relationship.

[0021] The above technical solution integrates the denoised pH value sequence and historical dosing data. The generalized predictive control algorithm can more accurately predict future pH value changes, thereby adjusting the dosing strategy in advance and enhancing the system's ability to respond quickly to water quality changes. Statistical methods reveal the quantitative relationship between the dosage and the pH value, enabling the system to accurately calculate the required dosage based on the predicted pH value changes, avoid overdosing or underdosing, and ensure that the pH value is stable within the optimal range.

[0022] Furthermore, the noise level of the data point can also be obtained according to the following method:

[0023] Replace the value of each data point in the data sequence with the average value of the adjacent data points on both sides of the data point, and replace the first data point and the last data point of the data sequence with the values ​​of the adjacent data points;

[0024] After the replacement, re-fitting the data points within the local range of the data points, and taking the mean value of the fitting loss as the noise level of the data points after the value replacement;

[0025] The noise level of the data point , For the The noise level of each data point is For the The noise level of the data point before the value replacement, For the The noise level of a data point after value replacement.

[0026] The above technical solution uses the difference in noise levels before and after the pH value data point replacement as the calculation factor for the final noise level. This method not only takes into account the noise performance of the original data, but also incorporates the evaluation after data smoothing, thereby providing a more comprehensive and detailed noise level measurement indicator.

[0027] Furthermore, the data fitting adopts the least square method.

[0028] Furthermore, the fitting loss is the difference between the fitting value and the actual value of the data point.

[0029] Furthermore, the extreme value point sequence in the data sequence includes a maximum value point sequence and a minimum value point sequence, and the maximum value point sequence and the minimum value point sequence are processed according to any of the above methods.

[0030] Furthermore, the method for obtaining the extreme point sequence is:

[0031] The data sequence is processed by the difference method to obtain the maximum value points and the minimum value points in the data sequence. According to the chronological order of acquisition time, the maximum value points constitute a maximum value point sequence, and the minimum value points constitute a minimum value point sequence.

[0032] The above technical solution processes the data sequence through the differential method, and can quickly identify the maximum and minimum points in the data sequence. This method simplifies the data processing process and improves the efficiency of data processing. The maximum and minimum points usually mark abnormal changes in the data sequence. By constructing the maximum point sequence and the minimum point sequence, the abnormal points in the data sequence can be more accurately identified and the ability of anomaly detection can be improved. The maximum point and the minimum point sequence can reflect the main trend changes of the data sequence, which is conducive to understanding the overall trend and change law of the data.

[0033] On the other hand, the present invention also provides a smart dosing control system for a sewage treatment plant based on carbon neutrality, the dosing control system comprising a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement any step of the dosing control method.

[0034] The present invention has the following effects:

[0035] When processing the pH value, the present invention optimizes the fitting process of the envelope by adaptively adjusting the interpolation density of the empirical mode decomposition algorithm, obtains a stable and accurate envelope, provides high-quality data for subsequent denoising and reconstruction, improves the denoising effect, and improves the stability and efficiency of the processing flow. Based on the high-quality data after denoising, the system can provide more accurate pH value prediction and drug input guidance. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0037] Figure 1 It is a schematic flow chart of the method of the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0039] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0040] Reference Figure 1 , a smart dosing control method for a sewage treatment plant based on carbon neutrality, comprising steps S1-S5, as follows:

[0041] S1: Obtain pH value data sequence during sewage treatment.

[0042] The first step to implementing accurate control and efficient management in modern wastewater treatment facilities is to obtain pH data, which is achieved through the following process:

[0043] Real-time monitoring and data collection: The pH dynamics of the treated water are continuously monitored by pH sensors. The sampling frequency of the sensor is set to 0.1 Hz, which means that 10 pH readings are collected per second to ensure the timeliness and continuity of the data.

[0044] Data transmission and integration: The collected pH data is transmitted to the central data processing platform in real time via the Internet of Things technology. The application of Internet of Things technology not only accelerates the data transmission efficiency, but also ensures the security and integrity of the data.

[0045] Sequence construction and preliminary screening: On the terminal device, the received pH data is arranged in order to form a continuous time series. During the sequence construction process, preliminary data screening and formatting are carried out simultaneously, laying a solid foundation for subsequent data analysis and processing.

[0046] S2: Evaluates the noise level between extreme points in the extreme point sequence.

[0047] In order to reduce the influence of noise on the interpolation fitting of the upper and lower envelopes in the EMD decomposition process through the interpolation density between adaptive extreme points, improve the quality of IMF decomposition and thus improve the final denoising effect, this step evaluates the noise level between the extreme points in the extreme point sequence, and then uses the noise level to adaptively interpolate the density to achieve adaptive interpolation density and improve the denoising effect.

[0048] S21: Evaluate the noise level of each data point.

[0049] The noise level of each data point in the collected original pH value data sequence is evaluated. In the field of signal processing and data analysis, noise usually manifests itself as abnormal mutations in the data stream or deviations from normal trends. In order to accurately quantify the noise level of the data point, the mutation of the data point within the local range of each data point can be used to reflect the noise level of the data point.

[0050] For the local range, in order to accurately capture the local features of the data points, an appropriate window size should be selected to define the local range of each data point. If the window is too small, important trend information may be ignored, while if the window is too large, the real signal change may be misjudged as noise. Specifically, this step sets each data point and w data points on its front and back sides as its local range. In this embodiment, w is 5. For example, the zth data point in the data sequence has a local range of the zth data point and 5 data points on its front and back sides, a total of 11 data points. If one side is less than 5, the other side is aligned with it. For example, if there are only 3 data points on the front side and 6 data points on the back side, 3 data points are also taken on the back side. At this time, the value of w becomes 3.

[0051] The mutation of data points can be reflected by the curve fitting loss in the local range. In statistics and signal processing, fitting loss refers to the difference between the data point and the fitting curve, that is, the absolute value of the difference between the actual value of the data point and the fitted value. If the fitting loss of a data point is significantly higher than that of other data points in its local range, it indicates that the data point exhibits higher mutation, which in turn reflects that the data point has a higher degree of noise.

[0052] Here are two ways to assess the noise level of each data point:

[0053] The first method is:

[0054] The least squares method is used to fit all data points in the local range of each data point to obtain a fitting curve. The horizontal axis of the fitting curve is the serial number of the data point in the local range, and the vertical axis is the fitting value of each data point. The fitting loss of each data point in the local range can be calculated, and then the mean of the fitting loss of all data points in the local range is used to quantify the noise level of the data point, which is regarded as the degree of deviation of the data point from the surrounding trend, reflecting the potential noise impact. Indicates the first The noise level of each data point.

[0055] The second method is an optimization of the first method. Taking into account that there may be multiple mutation points at the same time in the local range of the data sequence, the evaluation of the noise level should comprehensively consider the fitting loss of all mutation points rather than focusing on a single extreme value. Therefore, the value of each data point in the data sequence is optimized. Specifically, the value of each data point is replaced with the average value of the adjacent data points on both sides, and the first and last points are replaced with the values ​​of their adjacent points respectively. Based on the data sequence after numerical replacement, the fitting process in the first method is repeated, and the mean of the fitting loss is calculated as the noise level of the data point after numerical replacement.

[0056] By comparing the changes in the noise level of the data points before and after the numerical replacement (mean fitting loss), the difference in the noise level before and after can be regarded as the credibility factor of the noise level calculated by the first method. A larger difference indicates that the data point has significant mutation, which is more likely to be attributed to noise rather than the signal itself. The noise level of the data point depends not only on the mean fitting loss before and after the numerical replacement, but also on its credibility factor. If the change in the noise level before and after the numerical replacement is large, the larger the value of the credibility factor, the more likely the mutation of the point is caused by noise, and the noise level assessment should be given more attention.

[0057] Specifically, the noise level of each data point satisfies the following relationship:

[0058]

[0059] In the formula, For the The noise level of each data point is For the The noise level of the data point before the value replacement, For the The noise level of the data point after the value replacement, It reflects the change in the mean value of the fitting loss before and after data replacement, and takes it as The credibility factor, combined with , in order to make an accurate assessment of the noise level of each data point.

[0060] S22: Calculate the noise level between extreme points based on the noise level of each data point.

[0061] The difference method is used to detect extreme points (trend turning points) in the pH value data sequence. The difference method can effectively identify the turning points of the upward or downward trend in the sequence by comparing the values ​​of adjacent data points. Then, according to the maximum and minimum values ​​identified by the difference method, the maximum point sequence and the minimum point sequence are obtained as independent sequences. The EMD algorithm interpolates and fits the maximum point sequence to generate the upper envelope, and interpolates and fits the minimum point sequence to generate the lower envelope.

[0062] Taking into account that the EMD algorithm is not limited to adjacent extreme points when constructing the upper and lower envelopes, but needs to consider the global characteristics of the signal, for the maximum point sequence, the average of the noise levels of all maximum points between any two maximum points in the maximum point sequence (excluding the two maximum points themselves) is used as the noise level between the two maximum points; for the minimum point sequence, the average of the noise levels of all minimum points between any two minimum points in the minimum point sequence (excluding the two minimum points themselves) is used as the noise level between the two minimum points.

[0063] At this point, the noise level between any two maximum points in the maximum point sequence and the noise level between any two minimum points in the minimum point sequence are obtained. In this embodiment, the processing method for the maximum point sequence and the minimum point sequence is the same, so in the subsequent steps, only the maximum value sequence is used as an example for description.

[0064] S3: Set the initial interpolation density of the extreme point sequence, and use the noise level between the extreme points to adaptively adjust the initial interpolation density to obtain the interpolation density between the extreme points.

[0065] The construction of the upper and lower envelopes in the conventional EMD algorithm relies on interpolation between extreme points. However, traditional interpolation methods often use fixed density or equally spaced interpolation points, which may lead to overfitting or underfitting of the envelope in areas with complex signals or significant noise. Especially when the noise level is high, the fixed interpolation density may not accurately capture the true trend of the signal. To overcome the above limitations, this step introduces an adaptive interpolation density mechanism. The core is to dynamically adjust the number and distribution of interpolation points according to the local characteristics of the signal. In areas with large noise or complex signal changes, a lower interpolation density is used to avoid non-physical fluctuations of the envelope. In areas with stable signals or low noise, the interpolation density is appropriately increased to ensure that the envelope can accurately reflect the signal details.

[0066] By adaptively adjusting the interpolation density, the EMD algorithm can more accurately separate the IMF components, thereby effectively removing the non-essential changes introduced by the measurement noise while maintaining the essential characteristics of the signal. This step is crucial for the subsequent denoising operation, ensuring the reliability and authenticity of the denoised signal. Compared with the traditional EMD method, the EMD algorithm with adaptive interpolation density can improve the denoising effect, especially when processing noise-intensive pH data series, and can more finely restore the true dynamics of the signal.

[0067] Among them, the method of setting the initial interpolation density of the extreme point sequence is:

[0068] The inverse of the average time interval corresponding to all data points of the maximum point sequence is used as the initial interpolation density of the maximum point sequence. The formula is expressed as follows:

[0069]

[0070] In the formula, represents the initial interpolation density of the maximum point sequence, The number of data points representing the maximum point sequence, , Respectively represent the i-th and i+1-th maximum points in the maximum point sequence, , Respectively , The corresponding collection time is represents the time interval between these two points, Indicates this The number of time intervals corresponding to the maximum points, then the formula is to use the reciprocal of the average time interval of the maximum points in the maximum point sequence to represent the initial interpolation density of the maximum point sequence.

[0071] The method of adaptively adjusting the initial interpolation density by using the noise level to obtain the interpolation density between any two extreme value points includes:

[0072]

[0073] In the formula, For the The extreme point and The adjusted interpolation density between extreme points, is the initial interpolation density, For the The extreme point and The noise level between the extreme points, is a natural exponential function.

[0074] The formula relates the interpolation density to the noise level, which ensures that in areas with higher noise, When it is larger, the interpolation density will decrease accordingly to avoid overfitting. On the contrary, in areas with low noise levels, the interpolation density is increased. By increasing the density of interpolation points in low-noise areas and reducing the density of interpolation points in high-noise areas, the difficulty of signal analysis is simplified, making the intrinsic pattern of the signal easier to identify and analyze, and being able to capture signal details more finely, which is crucial for accurate signal analysis and subsequent modal decomposition. Adaptive adjustment of the interpolation density can reduce the interpolation points in the noisy area and avoid noise being mistaken for signal features and retained, thereby improving the effect of signal denoising, ensuring the accuracy of subsequent processing (such as prediction and control), and providing a clearer and more accurate data basis for subsequent signal processing and analysis.

[0075] S4: Using the EMD algorithm to perform interpolation fitting between any two of the extreme value points based on the adjusted interpolation density, and performing subsequent denoising after the interpolation fitting to obtain a denoised data sequence.

[0076] Adaptive interpolation density is used to perform interpolation fitting in the maximum point sequence and the minimum point sequence respectively to generate the upper envelope and the lower envelope, and then modal decomposition is performed to obtain a series of IMF components. Each IMF represents a different frequency component of the data. For each obtained IMF component, wavelet threshold denoising or median filtering denoising is used to eliminate or weaken the influence of random noise and retain the essential characteristics of the signal. After the denoising of each IMF component, these components are recombined according to the reconstruction principle of the EMD algorithm to restore a denoised pH value data sequence.

[0077] The optimized and denoised pH data sequence is used to guide dosage control in the sewage treatment process, which can significantly reduce the control error caused by signal fluctuations, ensure more accurate pH adjustment, and thus improve the efficiency and quality of the entire sewage treatment process.

[0078] In summary, this method improves the interpolation fitting process of the EMD algorithm through adaptive interpolation density, obtains accurate envelopes, and thus improves the accuracy of modal decomposition and effectively filters out noise interference, which not only improves the availability of data, but also provides a more accurate basis for subsequent signal analysis and processing.

[0079] S5: Predict the future pH value based on the denoised data sequence and control the input amount of the pH regulator.

[0080] In deepening the intelligent and automated transformation of sewage treatment processes, the generalized predictive control (GPC) algorithm plays a vital role, especially in the precise control of pH value and the intelligent input of pH regulators. The specific steps are as follows:

[0081] Data fusion and intelligent input: The denoised pH data sequence is combined with the past records of pH adjuster input as the input of the generalized predictive control algorithm. This step ensures that the algorithm can make predictions based on high-quality, noise-free data, while taking into account historical operating experience, enhancing the accuracy and practicality of the predictions.

[0082] Prediction of future pH value: Using the powerful prediction ability of GPC algorithm, based on the input historical data, the trend of future pH value change can be accurately predicted. This prediction function provides solid data support for subsequent intelligent regulation and control, ensuring the foresight and flexibility of the control strategy.

[0083] Intelligent calculation of input amount: Through in-depth statistical analysis methods, the functional relationship between historical pH adjuster input and corresponding pH value is explored. This analysis not only reveals the quantitative relationship between the two, but also provides a scientific basis for determining the input amount of adjuster required for future pH value targets. Based on the established functional relationship, the input amount of pH adjuster required to achieve the future predicted pH value target can be accurately calculated;

[0084] Intelligent control and feedback mechanism: Combining the predicted results with the calculated input, the intelligent control system can dynamically adjust the real-time input of pH regulator to ensure that the pH value is stable within the ideal range, while optimizing resource use, reducing unnecessary chemical consumption, and lowering carbon emissions during chemical treatment.

[0085] The present invention also provides a smart dosing control system for a sewage treatment plant based on carbon neutrality, the control system comprising a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement any step of the detection method.

[0086] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0087] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A smart dosing control method for a sewage treatment plant based on carbon neutrality, characterized in that: include: Collecting a pH value data sequence, performing data fitting on data points within a local range of each data point in the data sequence, and taking a mean value of fitting loss as a noise level of the data point; Obtaining an extreme value point sequence in the data sequence, and taking the average of the noise levels of all extreme value points between any two extreme value points in the extreme value point sequence as the noise level between any two extreme value points; The inverse of the average time interval between all extreme points of the extreme point sequence is used as the initial interpolation density of the extreme point sequence, and the initial interpolation density is adaptively adjusted using the noise level to obtain the interpolation density between any two extreme points, including: ; In the formula, For the The extreme point and The interpolation density between extreme points, is the initial interpolation density, For the The extreme point and The noise level between the extreme points, is a natural exponential function, and the interpolation density between any two extreme value points is negatively correlated with the noise level; The EMD algorithm is used to perform interpolation fitting between any two extreme points based on the adjusted interpolation density, and denoising is performed after the interpolation fitting to obtain a denoised data sequence. The future pH value is predicted based on the denoised data sequence, and the input amount of the pH regulator is controlled.

2. The intelligent dosing control method for a sewage treatment plant based on carbon neutrality according to claim 1 is characterized in that: Predicting the future pH value according to the denoised data sequence and controlling the input amount of the pH regulator includes: Using the denoised data sequence and the historical input amount of pH regulator as input of a generalized predictive control algorithm; The generalized predictive control algorithm predicts the future pH value based on the input content; The functional relationship between the historical input amount of the pH adjuster and the corresponding pH value is analyzed by a statistical method, and the input amount of the pH adjuster required for the future pH value is determined based on the functional relationship.

3. The intelligent dosing control method for a sewage treatment plant based on carbon neutrality according to claim 1 is characterized in that: The noise level of the data point can also be obtained as follows: Replace the value of each data point in the data sequence with the average value of the adjacent data points on both sides of the data point, and replace the first data point and the last data point of the data sequence with the values ​​of the adjacent data points; After the replacement, re-fitting the data points within the local range of the data points, and taking the mean value of the fitting loss as the noise level of the data points after the value replacement; The noise level of the data point , For the The noise level of each data point is For the The noise level of the data point before the value replacement, For the The noise level of a data point after value replacement.

4. The intelligent dosing control method for a sewage treatment plant based on carbon neutrality according to claim 3 is characterized in that: The data were fitted using the least squares method.

5. The intelligent dosing control method for a sewage treatment plant based on carbon neutrality according to claim 3 is characterized in that: The fitting loss is the difference between the fitted value and the actual value of the data point.

6. The intelligent dosing control method for a sewage treatment plant based on carbon neutrality according to claim 3 is characterized in that: The extreme value point sequence in the data sequence includes a maximum value point sequence and a minimum value point sequence, and the maximum value point sequence and the minimum value point sequence are processed according to the method described in any one of claims 1 to 5.

7. The intelligent dosing control method for a sewage treatment plant based on carbon neutrality according to claim 6 is characterized in that: The method for obtaining the extreme point sequence is: The data sequence is processed by the difference method to obtain the maximum value points and the minimum value points in the data sequence. According to the chronological order of acquisition time, the maximum value points constitute a maximum value point sequence, and the minimum value points constitute a minimum value point sequence.

8. A smart dosing control system for a sewage treatment plant based on carbon neutrality, characterized in that: The control system comprises a memory and a processor, wherein a computer program is stored in the memory, and the processor executes the computer program to implement the steps of the control system according to any one of claims 1 to 7.

Images