Method for controlling feeding of supernatant of kitchen waste as carbon source for denitrification based on three-dimensional fluorescence spectrum

Abstract

This invention discloses a method for controlling the addition of supernatant from kitchen waste as a carbon source for denitrification based on three-dimensional fluorescence spectroscopy. The method includes steps such as sample collection and detection, three-dimensional fluorescence spectral processing, establishment of a multiple linear regression model, and control of the carbon source dosage. By acquiring real-time online data on the changes in three-dimensional fluorescence spectra during the denitrification process of nitrate-containing wastewater and combining this with a multiple linear regression model to predict carbon source consumption, this invention achieves precise control of the carbon source dosage, ensuring a sufficient supply of carbon source during the denitrification stage while avoiding waste caused by excessive carbon source addition. This significantly improves the denitrification rate, reduces energy consumption and operating costs, and ensures that the effluent quality meets standards.

Description

Technical Field

[0001] This invention relates to the field of food waste treatment and wastewater treatment, specifically to an intelligent method for controlling the addition of food waste supernatant as a denitrification carbon source based on three-dimensional fluorescence spectral changes. Background Technology

[0002] Insufficient carbon sources have severely hampered the efficient nitrogen removal in wastewater treatment plants, and finding economical and sustainable carbon sources is one of the important ways to solve this problem.

[0003] Food waste is a common type of organic waste, rich in organic matter, which can be used as a raw material for producing soluble carbon. Utilizing the supernatant from food waste as an external carbon source can effectively supplement the carbon source needed by denitrifying bacteria in wastewater treatment plants, reducing the cost of external carbon sources, improving denitrification efficiency, and simultaneously achieving the resource utilization of food waste, thus achieving the goal of treating waste with waste. Furthermore, rationally controlling the amount of carbon source added is also a crucial step in achieving efficient denitrification of wastewater. Insufficient carbon source prevents denitrifying bacteria from carrying out the denitrification reaction normally, leading to insufficient nitrate nitrogen (NO3) in the effluent. - Increased NO3- concentration negatively impacts effluent quality. Excessive dosage increases treatment costs and may cause secondary pollution. Therefore, regular monitoring of effluent quality during the denitrification stage is crucial for ensuring a suitable carbon source. Detecting NO3- in the effluent... - -N, nitrite nitrogen (NO2) - Indicators such as -N can be used to determine the growth and metabolism of denitrifying bacteria, thereby determining whether the amount of carbon source added needs to be adjusted.

[0004] However, traditional wastewater quality analysis requires manual sampling in a laboratory using standard chemical methods, which is time-consuming, labor-intensive, has a low monitoring frequency, cannot achieve precise control, leads to high energy consumption, unstable effluent quality, and cannot cope with sudden wastewater discharge accidents. Based on this, many instrument manufacturers have developed continuous online monitoring equipment to replace manual wastewater quality analysis. However, online monitoring equipment for indicators such as total organic carbon (TOC), total nitrogen (TN), and total phosphorus (TP) is generally bulky, complex in structure, expensive, and can only monitor one indicator, making it difficult to widely apply in most wastewater treatment plants. Against this backdrop, how to quickly detect effluent quality and then accurately control carbon source dosage to achieve efficient nitrogen removal in the denitrification process has become a critical issue.

[0005] In recent years, three-dimensional fluorescence spectroscopy (3D-EEM) has been widely used for the detection of dissolved organic matter (DOM) in water bodies due to its advantages such as high sensitivity, high selectivity, non-destructive testing, rapid analysis, and ease of operation. 3D-EEM consists of a matrix spectrum of fluorescence intensity, excitation wavelength (Ex), and emission wavelength (Em). This technique allows for the acquisition of fluorescence information of specific components in a substance, enabling qualitative or quantitative analysis. Currently, there are no publicly available methods for controlling carbon source addition based on changes in three-dimensional fluorescence spectra. Summary of the Invention

[0006] This invention provides a method for controlling the addition of kitchen waste supernatant as a carbon source for denitrification based on three-dimensional fluorescence spectroscopy. This method utilizes the correlation between fluorescent substances and water pollution to establish a multiple linear regression model based on stepwise regression analysis or principal component analysis. Then, by acquiring the changes in three-dimensional fluorescence spectra of nitrate nitrogen-containing wastewater during the denitrification process in real time online, and combining the established model to predict carbon source consumption, intelligent and precise control of carbon source addition is achieved.

[0007] To achieve its objectives, the present invention employs the following technical solution:

[0008] A method for controlling the addition of kitchen waste supernatant as a denitrification carbon source based on three-dimensional fluorescence spectroscopy includes the following steps:

[0009] Step 1: Sample collection and testing

[0010] During the denitrification process of nitrate-containing wastewater using kitchen waste supernatant as a carbon source, samples are taken at intervals to ensure a minimum of 15 samples; the main water quality parameters (preferably NO3) of each sample are tested. - -N concentration) and its three-dimensional fluorescence spectrum;

[0011] Step 2: Three-dimensional fluorescence spectroscopy processing

[0012] The three-dimensional fluorescence spectra of each sample were quantitatively analyzed using the fluorescence region integration (FRI) method. The three-dimensional fluorescence spectrum of each sample could be divided into five fluorescence regions representing tyrosine, tryptophan, fulvic acid, soluble microbial metabolites, and humic acid organic compounds, respectively. Quantitative analysis using FRI yielded the fluorescence intensity of each sample in each region, reflecting the concentration of the corresponding organic compound.

[0013] Furthermore, before performing quantitative analysis using the fluorescence region integration method, Raman correction and Rayleigh scattering preprocessing are required for the three-dimensional fluorescence spectrum.

[0014] Step 3: Establishing a Multiple Linear Regression Model

[0015] This invention uses the fluorescence intensity of five fluorescence regions obtained from analysis as independent variables and major water quality parameters as dependent variables to establish a multiple linear regression model. The model's parameters are estimated and optimized using training data. The specific steps are as follows:

[0016] (3-1) Correlation analysis was performed on the fluorescence intensity and the concentration of major water quality parameters in each region:

[0017] Fluorescence changes in the five regions showed varying degrees of correlation with the main water quality parameters. Using fluorescence intensity in each region as the independent variable and the main water quality parameters measured in step 1 as the dependent variable, linear equations were established between the main water quality parameters and fluorescence intensity in each region. The correlation coefficients for each linear equation were determined, and regions with insignificantly correlated coefficients were removed. Then, a multiple linear regression model was established using the fluorescence intensity of the remaining regions as the independent variable and the main water quality parameters as the dependent variable.

[0018] Specifically, SPSS was used to calculate the correlation coefficient matrix between each dependent and independent variable. A correlation coefficient ≥ 0.6 was defined as a significant correlation between variables, and regions with correlation coefficients < 0.6 were removed.

[0019] Nitrate nitrogen NO3 - Taking -N concentration as the main water quality parameter, the established multiple linear regression model can be expressed as follows:

[0020]

[0021] In equation (1), C1, C2, C3, C4, and C5 represent the relative fluorescence intensities of tyrosine, tryptophan, fulvic acid, soluble microbial metabolites, and humic acid, respectively. The coefficients of the substances corresponding to the regions removed from the model are 0.

[0022] (3-2) Solve the multicollinearity problem that may exist in the model.

[0023] The fluorescence intensities of various fluorescent regions (i.e., C1, C2, C3, C4, and C5 in equation (1)) often exhibit a high correlation, indicating a potential serious multicollinearity problem. This correlation typically implies that these regions share a large amount of information, causing the model to be influenced by other regions when attempting to capture changes in a specific region. This not only leads to inaccurate model predictions but also significantly reduces the model's explanatory power. The present invention addresses the multicollinearity problem in two ways:

[0024] When the number of independent variables m ≤ 3 in the established multiple linear regression model, stepwise regression analysis is used to address potential multicollinearity issues: Independent variables are introduced into the model one by one. After each introduction, the selected variables are tested individually. If an previously introduced variable becomes insignificant due to a newly introduced independent variable, it is removed. This process is repeated until no insignificant independent variables are selected into the model, and no significant variables are removed.

[0025] When the number of independent variables (m) in the established multiple linear regression model is greater than 3, principal component analysis (PCA) is used to address potential multicollinearity issues: multiple variables are reduced to a smaller number of important composite variables through linear transformation. In PCA, a cumulative contribution rate of principal components exceeding 95% is considered a good result.

[0026] (3-3) The correlation coefficient of the regression equation is continuously optimized using a multiple linear regression algorithm until the goodness of fit no longer increases significantly. The final multiple linear regression equation obtained is then used as the final solution of the model. The specific steps are as follows:

[0027] The first linear regression was performed on the model using SPSS software to determine the coefficients of the linear regression equation. (Using NO3...) - Taking -N concentration as an example, the first linear regression equation is:

[0028]

[0029] The correlation coefficient, Pearson coefficient, and residual obtained from equation (2) are denoted as R1. 2 P1 and S1. The closer the correlation coefficient is to 1, the better the fit. When the fit is not ideal, outliers are removed using the residuals, and the fit is repeated. This process is repeated until the goodness of fit does not improve significantly in the Nth iteration, and the final multiple linear regression equation is taken as the final solution. (Final NO3) - The multivariate linear relationship between -N concentration and fluorescence intensity in the five regions is as follows:

[0030]

[0031] Step 4: Control of carbon source dosage

[0032] During the denitrification process of nitrate-containing wastewater, the three-dimensional fluorescence spectrum of the effluent is sampled and monitored at regular intervals. Based on the model established in step 3, the main water quality parameters of the effluent are calculated and obtained. The values ​​of the main water quality parameters of the effluent when carbon source needs to be added are set and recorded as set values.

[0033] Next batch of influent and carbon source addition: When the main water quality parameters calculated by the model at three consecutive time points are less than or equal to the set value, it indicates that this batch of denitrification has ended, the denitrification tank begins to discharge water, and the next batch of influent and carbon source addition is carried out at the same time.

[0034] Secondary carbon source addition: When the main water quality parameters calculated by the model at three consecutive time points are greater than the set value, and the difference between the three values ​​is within 10%, it indicates that the main water quality parameters of the influent are higher than those of the normal batch. The model will issue an early warning and, through parameter adjustment, output a secondary carbon source addition signal to perform secondary carbon source addition.

[0035] Furthermore, the sampling time interval between steps 1 and 4 is 10 to 20 minutes.

[0036] Furthermore, in step 4, to ensure the quality of the effluent, the amount of kitchen waste supernatant added as a carbon source for denitrification is guaranteed to have the optimal carbon-to-nitrogen ratio in the system.

[0037] Compared with the prior art, the beneficial effects of the present invention are reflected in:

[0038] 1. This invention utilizes the supernatant of kitchen waste as a carbon source for denitrification, which can reduce dependence on commercial carbon sources, thereby reducing operating costs, while realizing the resource utilization of kitchen waste and achieving the goal of "treating waste with waste and treating both wastes simultaneously".

[0039] 2. This invention achieves precise control of carbon source dosage by acquiring the three-dimensional fluorescence spectrum changes of nitrate nitrogen-containing wastewater during the denitrification process in real time online and combining it with a multiple linear regression model to predict carbon source consumption. This ensures a sufficient supply of carbon source during the denitrification stage, while avoiding waste caused by excessive carbon source addition. It greatly improves the denitrification rate, reduces energy consumption and operating costs, and ensures that the effluent water quality meets the standards. Attached Figure Description

[0040] Figure 1 This is a flowchart of the method for adding kitchen waste supernatant as a denitrification carbon source based on three-dimensional fluorescence spectroscopy according to the present invention.

[0041] Figure 2 This is a distribution map of the fluorescence regions of five representative organic compounds obtained by FRI technology in an embodiment of the present invention.

[0042] Figure 3 This is a correlation diagram showing the changes in fluorescence intensity and nitrate nitrogen concentration in five major regions in an embodiment of the present invention.

[0043] Figure 4 This diagram illustrates the effect of using the supernatant from kitchen waste as a denitrification carbon source to treat nitrate-containing wastewater in an embodiment of the present invention. Detailed Implementation

[0044] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention. These all fall within the scope of protection of the present invention.

[0045] Example 1

[0046] This embodiment predicts the carbon source consumption during the denitrification process of nitrate-containing wastewater based on three-dimensional fluorescence spectral changes, and intelligently controls the addition of kitchen waste supernatant as a carbon source for denitrification. The specific process is as follows:

[0047] (1) During the denitrification process of nitrate-containing wastewater using kitchen waste supernatant as a carbon source, samples were taken at intervals to ensure that the number of samples was not less than 15; the main water quality parameters (NO3) of each sample were tested. - -N concentration) and its three-dimensional fluorescence spectrum.

[0048] Specifically, in this embodiment, NO3 is configured. - For nitrate-containing wastewater with a nitrogen concentration of 30 mg / L, to ensure effluent quality, kitchen waste supernatant was added as a carbon source for denitrification at the optimal carbon-to-nitrogen ratio (C / N = 5). Samples were taken and filtered every 20 minutes, and NO3 was measured according to standard methods. - -N concentration changes were studied, and the three-dimensional fluorescence spectra of the samples were measured using a fluorescence spectrometer at room temperature (25°C).

[0049] (2) The obtained three-dimensional fluorescence spectra are first preprocessed with Raman correction and Rayleigh scattering removal, and then quantitative analysis is performed using FRI. For example... Figure 2 As shown ( Figure 2 (a) and (b) in the image represent two samples. Through FRI quantitative analysis, the three-dimensional fluorescence spectrum of each sample can be resolved into five fluorescence regions (C1 to C5) representing tyrosine, tryptophan, fulvic acid, soluble microbial metabolites, and humic acid organic compounds, respectively. FRI quantitative analysis yields the fluorescence intensity of each sample in each region, reflecting the concentration of the corresponding organic compound.

[0050] (3) Figure 3 As shown, the fluorescence changes in five regions are related to NO3. - The changes in NO3- concentration showed varying degrees of correlation. Using the fluorescence intensity of each region as the independent variable and the NO3- concentration measured in step 1 as the NO3- concentration, the correlation was analyzed. - With NO3- concentration as the dependent variable, establish NO3- concentration separately. -The correlation equations between NO3- concentration and fluorescence intensity in various regions were established, and the correlation coefficients of each equation were determined. Regions with insignificant correlation coefficients were then removed (a correlation coefficient ≥ 0.6 was considered statistically significant, and regions with correlation coefficients < 0.6 were removed). The fluorescence intensity of the remaining regions was then used as the independent variable and NO3- as the dependent variable. - A multiple linear regression model was established based on the -N concentration.

[0051] Specifically, in this embodiment, NO3 is obtained through correlation coefficient analysis. - The correlation matrix between NO3- concentration and C1-C5, where NO3- - The correlation coefficients between -N concentration and C1, C4, and C5 were 0.64, 0.84, and 0.66, respectively, while the correlation coefficients with C2 and C3 were less than 0.6; therefore, C2 and C3 were removed. The established multiple linear regression model is (R0...). 2 =0.8790):

[0052]

[0053] Next, the multicollinearity problem in the equation was solved using stepwise regression analysis. The first fitted multicollinear regression equation is as follows:

[0054]

[0055] Among them, the correlation coefficient R 2 It is 0.8837.

[0056] To obtain the optimal multiple linear regression model, this embodiment continuously optimizes the correlation coefficient of the regression equation using a multiple linear regression algorithm, removing outliers. This process is repeated until the goodness of fit no longer increases significantly, ultimately yielding the correlation coefficient R0. 2 The final NO3 was 0.9724. - The multivariate linear relationship between -N concentration and fluorescence intensity in the five regions is as follows:

[0057]

[0058] (4) During the denitrification process of nitrate-containing wastewater, the three-dimensional fluorescence spectrum of the effluent is monitored by taking samples at regular intervals (20 min intervals), and the NO3 in the effluent is calculated based on the model established in step (3). - -N concentration; setting the NO3 concentration in the effluent when a carbon source needs to be added. - The concentration of -N is denoted as the set value. In this embodiment, the standard for the effluent from the denitrification tank is set at 5 mg / L, i.e., set value = 5 mg / L.

[0059] Next batch of influent and carbon source addition: When the model calculates NO3 at three consecutive time points- When the NO3- concentration is less than or equal to the set value, it indicates that this batch of denitrification has ended, the denitrification tank begins to discharge water, and the next batch of influent and carbon source addition begins simultaneously. In actual operation: Wastewater influent NO3 - -N concentration is a (unit: mg / L). Starting from t = t1 (unit: min), the model outputs the signal "NO3" at three consecutive time points. - When "-N=a1≤Set value", this signal is immediately transmitted to the intelligent control system, and the system then starts the carbon source dosing pump, with a carbon source dosing amount of 5a. At this time (t=t2), the model outputs the signal "NO3". - -N = amg / L, C / N = n”, denitrification continues.

[0060] Secondary carbon source addition: When the NO3 calculated by the model at three consecutive time points - When the concentration of NO3-N is greater than the set value, and the difference between the three values ​​is within 10%, it indicates that the NO3-N concentration in the influent is high. - When the NO3- concentration is higher than usual, the model issues a warning and, through parameter adjustments, outputs a secondary carbon source addition signal to initiate secondary carbon source addition. In actual operation: when the wastewater treatment plant's water quality fluctuates significantly (t=t3), the model monitors NO3- concentrations at three consecutive time points. - The NO3 concentration (a = a3, a4, a5) differs by less than 10% and a > set value, indicating that this batch of influent NO3 - When the NO3- concentration is higher than usual, the model activates its automatic warning function. The intelligent control system receives the warning signal and detects NO3- in the influent immediately. - The system calculates the actual carbon source concentration (-N) and feeds this value back to the model. After parameter correction, the model outputs a "secondary carbon source addition" signal to the intelligent control system, which then restarts the carbon source addition pump.

[0061] To verify the model established in step (3) of this embodiment, 30 mg / L of nitrate-containing wastewater was prepared again, and kitchen waste supernatant was added as a carbon source for denitrification at a C / N ratio of 5. Samples were taken at 20-minute intervals, and the three-dimensional fluorescence spectra of the samples were monitored. The three-dimensional fluorescence spectral data of the samples were sequentially input into the model, and the model output the NO3 during the denitrification process in real time through the pre-set algorithm. - Changes in -N concentration and C / N ratio. For example... Figure 4 As shown, at t = 200, 220, and 240 min, the model outputs the signal "NO3". - When -N < 5 mg / L, it indicates that this batch of denitrification has ended, and the next batch of influent and carbon source will begin.

[0062] At t = 500 min, the denitrification tank began its third batch of influent, and at t = 620 min, the model began to continuously output the NO3 signal at three time points.- "-N = 19.75 mg / L > 5 mg / L", and the concentration difference among the three is within 10%, indicating that the NO3 in the third batch of influent is relatively high. - When the NO3- concentration in the influent is higher than usual, the model automatically activates the early warning function. This signal is immediately transmitted to the intelligent control system, which then detects the actual NO3 concentration in the influent. - The NO3- concentration was 40 mg / L, and this value was promptly fed back to the model. After parameter adjustment and optimization, the model output a "secondary carbon source dosage" signal, which was then transmitted to the intelligent control system. The system activated the carbon source dosing pump to add the secondary carbon source, with a dosage of (19.75 + (40 - 30)) * 5 = 148.75 mg / L. When t = 860 min, the model continuously output three "NO3" signals. - "-N<5mg / L" indicates that the third batch of influent denitrification has ended.

[0063] The above results demonstrate that by acquiring three-dimensional fluorescence spectral changes in real time and combining them with a multiple linear regression model, this invention can accurately predict carbon source consumption, achieve precise control of carbon source dosage, and avoid waste and insufficient supply caused by excessive carbon source dosage.

Claims

1. A method for controlling the addition of kitchen waste supernatant as a denitrification carbon source based on three-dimensional fluorescence spectroscopy, characterized in that, Includes the following steps: Step 1: Sample collection and testing During the denitrification process of nitrate-containing wastewater using kitchen waste supernatant as a carbon source, samples were taken at intervals to ensure a minimum of 15 samples. The main water quality parameters and three-dimensional fluorescence spectra of each sample were measured. The main water quality parameter was NO3. - -N concentration; Step 2: Three-dimensional fluorescence spectroscopy processing The fluorescence region integration method was used to quantitatively analyze the three-dimensional fluorescence spectra of each sample. The three-dimensional fluorescence spectrum of each sample can be divided into five fluorescence regions representing tyrosine, tryptophan, fulvic acid, soluble microbial metabolites and humic acid organic matter, respectively. The fluorescence intensity of each sample in each region was obtained to reflect the concentration of the corresponding organic matter. Step 3: Establishing a Multiple Linear Regression Model Correlation analysis was conducted between fluorescence intensity and main water quality parameters in each region: using fluorescence intensity in each region as the independent variable and the main water quality parameters measured in step 1 as the dependent variable, linear equations were established between the main water quality parameters and fluorescence intensity in each region, and the correlation coefficients of each linear equation were determined. Then, regions corresponding to coefficients with no significant correlation were removed. Finally, a multiple linear regression model was established using fluorescence intensity in the remaining regions as the independent variable and the main water quality parameters as the dependent variable. Then, the multicollinearity problem in the model is solved by stepwise regression analysis or principal component analysis; then, the correlation coefficient of the regression equation is continuously optimized by the multivariate linear regression algorithm until the goodness of fit no longer increases significantly, and the final multivariate linear regression equation is taken as the final solution of the model. Step 4: Control of carbon source dosage During the denitrification process of nitrate-containing wastewater, the three-dimensional fluorescence spectrum of the effluent is sampled and monitored at regular intervals, and the main water quality parameters of the effluent are calculated based on the model established in step 3. Set the values ​​of the main water quality parameters of the effluent when carbon source needs to be added, and record them as set values; When the main water quality parameters calculated by the model at three consecutive time points are less than or equal to the set value, it indicates that this batch of denitrification has ended and the next batch of influent and carbon source addition can begin. When the main water quality parameters calculated by the model at three consecutive time points are greater than the set values, and the difference between the three values ​​is within 10%, it indicates that the main water quality parameters of the influent are higher than those of the normal batch. The model will issue an early warning and, through parameter adjustment, output a secondary carbon source addition signal to perform secondary carbon source addition.

2. The method according to claim 1, characterized in that: The sampling time interval between steps 1 and 4 is 10-20 minutes.

3. The method according to claim 1, characterized in that: In step 2, before performing quantitative analysis using the fluorescence region integration method, Raman correction and Rayleigh scattering preprocessing are required for the three-dimensional fluorescence spectrum.

4. The method according to claim 1, characterized in that: In step 3, during the correlation analysis, it is stipulated that variables are significantly correlated when the correlation coefficient is ≥0.6, and regions with correlation coefficients <0.6 are eliminated.

5. The method according to claim 1, characterized in that, In step 3: when the number of independent variables in the established multiple linear regression model is no more than 3, stepwise regression analysis is used to solve the multicollinearity problem; when the number of independent variables in the established multiple linear regression model is more than 3, principal component analysis is used to solve the multicollinearity problem.

6. The method according to claim 1, characterized in that: In step 4, to ensure the quality of the effluent, the amount of kitchen waste supernatant added as a carbon source for denitrification is guaranteed to have the optimal carbon-to-nitrogen ratio in the system.

Images