Chemical wastewater treatment process abnormal condition biochemical potential evaluation method

Abstract

The application discloses a kind of chemical wastewater treatment process abnormal condition biochemical potential evaluation method.The application is based on the water quality parameter of chemical wastewater treatment plant to construct sample dataset, based on multiple-input multiple-output fuzzy neural network to establish wastewater treatment process ATP and BOD prediction model, using the water quality parameter sample of sample dataset to train and test wastewater treatment process ATP and BOD prediction model after, using the wastewater treatment process ATP and BOD prediction model after testing to predict ATP concentration value and BOD concentration value in real time, establish based on ATP and BOD real-time prediction abnormal condition discrimination matrix, accurately identify abnormal condition type, and combine ATP concentration value and BOD concentration value to evaluate the sludge condition and biochemical disposal potential of chemical wastewater treatment plant, adjust the load operation scheme of chemical wastewater treatment plant.The application realizes the accurate identification of sludge poisoning, carbon source deficiency and other abnormal conditions, is beneficial to timely grasp the abnormal condition of sludge activity in chemical wastewater treatment process, and improves the operation efficiency of wastewater treatment process.

Description

Technical Field

[0001] This invention relates to the field of wastewater treatment technology, specifically to a method for assessing the biochemical potential under abnormal operating conditions in chemical wastewater treatment processes. Background Technology

[0002] The pollutant composition of chemical wastewater is relatively stable, but it faces bottlenecks in treatment due to high organic matter concentration, poor biodegradability, and biological toxicity. Because the microorganisms in the biological treatment units are highly susceptible to impact, untimely control can easily lead to abnormal operating conditions at the wastewater treatment plant, and even wastewater exceeding discharge standards. Common abnormal operating conditions in chemical wastewater treatment processes include sludge bulking and sludge poisoning. Delayed detection and handling of abnormal conditions, as well as inadequate safety management during the wastewater treatment process, will directly affect the effluent quality of chemical wastewater treatment, jeopardize the operation of the entire wastewater treatment system, and in severe cases, even cause the entire chemical wastewater treatment process to collapse.

[0003] Most chemical wastewater treatment processes involve complex influent flow rates, influent composition, pollutant types, and organic matter concentrations, which can easily lead to large fluctuations in the microbial content of the sludge-water mixture in the biochemical reaction zone. At the same time, there are many types and numbers of activated sludge microorganisms, and the microorganisms used under different process operating conditions vary. The operating environment has a significant and uncertain impact on microorganisms.

[0004] To promptly identify abnormal operating conditions in chemical wastewater treatment processes, those skilled in the art have conducted relevant research. Chinese invention patent CN 112363391 B discloses a sludge bulking suppression method based on adaptive segmented sliding mode control. This invention employs a fuzzy neural network to obtain water quality parameters during wastewater treatment operation, predicts the sludge volume index in real time, and determines the operating condition of the process. This provides a reference signal for controller switching. Furthermore, a segmented sliding mode controller with an adaptive switching mechanism is designed to regulate dissolved oxygen and nitrate nitrogen concentrations, coordinate biochemical reaction processes, improve sludge settling characteristics, and ensure that the wastewater treatment process returns to normal operating conditions when sludge bulking occurs. Chinese invention patent CN 115578015 B discloses a method, system, and storage medium for monitoring the entire wastewater treatment process based on the Internet of Things (IoT). Specifically, it includes: acquiring current wastewater treatment process information and setting key water quality monitoring indicators for water quality monitoring points; obtaining water quality monitoring results at each point through the key water quality monitoring indicators, constructing a full-process wastewater treatment model, and using the full-process model for real-time monitoring and control based on the water quality monitoring results; generating fault warning information for the wastewater treatment process, adding a fault detection model through deep learning for real-time learning, tracing the source of faults in the wastewater treatment process based on water quality monitoring results and multivariate operating condition data, and visualizing the full-process model. This invention uses IoT technology to remotely monitor the entire wastewater treatment process, achieving dynamic monitoring of water quality changes and identification of abnormal operating conditions during wastewater treatment, significantly improving the efficiency and accuracy of water quality monitoring.

[0005] However, existing research on abnormal operating conditions in chemical wastewater treatment processes mainly focuses on deep learning and other aspects, and has not yet fully analyzed the key variables describing activated sludge in the biochemical unit during chemical wastewater treatment. As a result, it is impossible to effectively identify abnormal operating conditions in chemical wastewater treatment processes, accurately assess the treatment potential of biochemical units, and ensure that wastewater treatment facilities reach a stable operating state.

[0006] The assessment of the biochemical sludge biological activity and treatment capacity in the biochemical unit of chemical wastewater treatment processes mainly involves judging the degradation capacity of activated sludge for pollutants. Sludge activity and wastewater treatment effectiveness are two primary evaluation criteria. On one hand, adenosine triphosphate (ATP) is not only the most direct energy source in organisms but also an important indicator of microbial life activities, and can be used to monitor sludge activity. When influent water quality changes or the concentration of toxic substances changes drastically, the microorganisms in the sludge may enter a dormant state, and their biological activity will decrease. However, ATP, as a marker of microbial cell life and vitality, can indirectly reflect microbial activity and thus their metabolic status. Therefore, ATP is also considered an important parameter reflecting microbial activity in wastewater treatment processes. On the other hand, biochemical oxygen demand (BOD) is an important water quality indicator in wastewater treatment processes. The level of BOD reflects the degree to which organic matter in the water body is degraded by microbial biochemical processes. Real-time monitoring of BOD is an important means of wastewater treatment. Therefore, achieving real-time prediction of ATP and BOD is key to understanding the biochemical sludge treatment capacity of the biochemical unit.

[0007] Therefore, there is an urgent need to propose a biochemical potential assessment method for abnormal operating conditions in chemical wastewater treatment processes based on real-time prediction of ATP and BOD, so as to accurately identify abnormal operating conditions in chemical wastewater treatment processes and ensure that the effluent quality meets standards under abnormal operating conditions. Summary of the Invention

[0008] To address the current inability to directly and accurately determine the activity of sludge in biochemical units, this invention proposes a method for assessing the biochemical potential under abnormal operating conditions in chemical wastewater treatment processes. By combining chemical wastewater management data with artificial intelligence algorithms, an abnormal operating condition discrimination matrix is ​​established to assess abnormalities in petrochemical wastewater treatment processes. This enables accurate assessment of the biochemical treatment potential of biochemical units in petrochemical wastewater treatment processes, effectively ensuring the effluent quality of petrochemical wastewater treatment processes.

[0009] The present invention specifically adopts the following technical solution:

[0010] A method for assessing the biochemical potential of abnormal operating conditions in chemical wastewater treatment processes, which evaluates the biochemical treatment potential of abnormal operating conditions of chemical wastewater by real-time prediction of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD), specifically includes the following steps:

[0011] Step 1: Obtain water quality parameters from the chemical wastewater treatment plant. After preprocessing, normalizing, and principal component analysis of the water quality parameters, multiple chemical wastewater water quality parameter samples are obtained. Each chemical wastewater water quality parameter sample is randomly assigned to the training sample set and the test sample set. The training sample set and the test sample set together constitute the sample dataset used to train the ATP and BOD prediction models for the wastewater treatment process.

[0012] Step 2: Establish ATP and BOD prediction models for the wastewater treatment process based on a multi-input multi-output fuzzy neural network;

[0013] Step 3: Use the training set to train the ATP and BOD prediction models for the wastewater treatment process to obtain the trained ATP and BOD prediction models for the wastewater treatment process.

[0014] Step 4: Use the test set to verify the prediction effect of the ATP and BOD prediction model for the wastewater treatment process after training. If the prediction effect of the ATP and BOD prediction model for the wastewater treatment process after training has reached the best, proceed to step 5. Otherwise, return to step 3 and continue to train the ATP and BOD prediction model for the wastewater treatment process using the training set.

[0015] Step 5: Using the ATP and BOD prediction model of the wastewater treatment process after testing, predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) in real time based on the water quality parameters of the chemical wastewater, generate an abnormal operating condition discrimination matrix, use the abnormal operating condition discrimination matrix to identify the type of abnormal operating condition, and evaluate the sludge status and biochemical treatment potential of the chemical wastewater treatment plant based on the ATP and BOD concentration values, and adjust the load operation plan of the chemical wastewater treatment plant.

[0016] Preferably, step 1 specifically includes the following steps:

[0017] Step 1.1: Obtain the chemical wastewater from the chemical wastewater treatment plant and measure the influent flow rate, water quality parameters, adenosine triphosphate (ATP), and biochemical oxygen demand (BOD).

[0018] Step 1.2: Preprocess the water quality parameters of the chemical wastewater, remove abnormal data from each water quality parameter, normalize the preprocessed water quality parameters, and analyze the obtained water quality parameters using principal component analysis. Based on the correlation coefficient and contribution rate of each water quality parameter to the prediction of ATP and BOD in the wastewater treatment process, screen the water quality parameters and influent flow rate as input variables, and determine the output variables by combining the measured ATP concentration and BOD concentration of the chemical wastewater. Use the input and output variables to form a sample of water quality parameters for the chemical wastewater.

[0019] Step 1.3: Randomly allocate the water quality parameter samples of each chemical wastewater to the training sample set and the test sample set to generate the training sample set and the test sample set, and construct the sample dataset for training the ATP and BOD prediction model of the wastewater treatment process.

[0020] Preferably, when the chemical wastewater is wastewater from a propylene oxide plant and ethylene alkali residue wastewater, the water quality parameters include: influent chemical oxygen demand (COD) concentration, effluent COD concentration, formaldehyde concentration, cumene concentration, methanol concentration, propylene oxide concentration, propylene concentration, formic acid concentration, influent flow rate, sludge retention time, sludge MLSS concentration in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sulfide concentration in the biological treatment tank, volatile phenol concentration in the biological treatment tank, sludge settling ratio, sludge volume index, effluent biochemical oxygen demand (BOD) concentration, adenosine triphosphate (ATP) concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP).

[0021] When the chemical wastewater is purified terephthalic acid wastewater and ethylene alkali residue wastewater, the water quality parameters include influent chemical oxygen demand (COD) concentration, effluent COD concentration, formaldehyde concentration, terephthalic acid concentration, isophthalic acid concentration, acetic acid concentration, influent flow rate, sludge retention time, sludge MLSS concentration in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sludge settling ratio, sludge volume index, effluent biochemical oxygen demand (BOD) concentration, adenosine triphosphate (ATP) concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP).

[0022] When the chemical wastewater is propylene oxide wastewater, purified terephthalic acid wastewater, and ethylene alkali residue wastewater, the water quality parameters include influent chemical oxygen demand (COD) concentration, effluent COD concentration, formaldehyde concentration, cumene concentration, methanol concentration, propylene oxide concentration, propylene concentration, formic acid concentration, terephthalic acid concentration, isophthalic acid concentration, acetic acid concentration, influent flow rate, sludge retention time, sludge MLSS concentration in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sludge settling ratio, sludge volume index, effluent biochemical oxygen demand (BOD) concentration, adenosine triphosphate (ATP) concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP).

[0023] When the chemical wastewater is propylene oxide wastewater, aromatic hydrocarbon wastewater, purified terephthalic acid wastewater, and ethylene alkali residue wastewater, the water quality parameters include influent chemical oxygen demand (COD) concentration, effluent COD concentration, cyanide concentration, polycyclic aromatic hydrocarbons (PAHs), volatile phenol concentration, sulfide concentration, formaldehyde concentration, cumene concentration, methanol concentration, propylene oxide concentration, formic acid concentration, terephthalic acid concentration, isophthalic acid concentration, acetic acid concentration, influent flow rate, sludge retention time, MLSS concentration of sludge in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sludge settling ratio, sludge volume index, effluent biochemical oxygen demand (BOD) concentration, adenosine triphosphate (ATP) concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP).

[0024] Preferably, when the chemical wastewater is propylene oxide plant wastewater and ethylene alkali residue wastewater, the input variables include influent chemical oxygen demand (COD) concentration, MLSS concentration of sludge in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, sludge volume index, temperature, influent ammonia nitrogen concentration, propylene oxide concentration, formaldehyde concentration, and influent flow rate; the output variables include adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration.

[0025] When the chemical wastewater is purified terephthalic acid wastewater and ethylene alkali residue wastewater, the input variables include influent chemical oxygen demand (COD) concentration, MLSS concentration of sludge in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, sludge volume index, temperature, influent ammonia nitrogen concentration, formaldehyde concentration, terephthalic acid concentration, influent flow rate, and isophthalic acid concentration. The output variables include adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration.

[0026] When the chemical wastewater is propylene oxide wastewater, purified terephthalic acid wastewater, and ethylene alkali residue wastewater, the input variables include the influent chemical oxygen demand (COD) concentration, the sludge MLSS concentration in the biological treatment tank, the dissolved oxygen (DO) concentration in the biological treatment tank, the sludge volume index, the temperature, the influent ammonia nitrogen concentration, the formaldehyde concentration, the cumene concentration, the methanol concentration, the propylene oxide concentration, the terephthalic acid concentration, and the influent flow rate; the output variables include the adenosine triphosphate (ATP) concentration and the biochemical oxygen demand (BOD) concentration.

[0027] When the chemical wastewater is propylene oxide wastewater, aromatic hydrocarbon wastewater, purified terephthalic acid wastewater, and ethylene alkali residue wastewater, the input variables include the influent chemical oxygen demand (COD) concentration, the MLSS concentration of the sludge in the biological treatment tank, the concentration of polycyclic aromatic hydrocarbons, volatile phenols, sulfides, the dissolved oxygen (DO) concentration in the biological treatment tank, the sludge volume index, temperature, the influent ammonia nitrogen concentration, formaldehyde concentration, cumene concentration, propylene oxide concentration, terephthalic acid concentration, and influent flow rate. The output variables include the adenosine triphosphate (ATP) concentration and the biochemical oxygen demand (BOD) concentration.

[0028] Preferably, in step 2, the multi-input multi-output fuzzy neural network includes an input layer, an RBF layer, a rule layer, and an output layer;

[0029] The input layer of the multi-input multi-output fuzzy neural network is configured as follows:

[0030] x i =u i (1)

[0031] x = [x1, x2, ..., x k (2)

[0032] In the formula, i is the index of the input variable, i = 1, 2, ..., k, and k is the total number of input variables; x i Let u be the value of the i-th input variable. i Let x be the input value of the i-th input neuron, and let x be the input vector.

[0033] The RBF layer contains multiple neurons, and the output function value of each neuron in the RBF layer is:

[0034]

[0035] In the formula, φ j c is the output value of the j-th neuron in the RBF layer, where j is the neuron number (j = 1, 2, ..., Q) and Q is the total number of neurons in the RBF layer; ij σ is the center of the j-th neuron. ij The width of the j-th neuron;

[0036] The rule layer contains multiple neurons, and the output function value of each neuron in the rule layer is:

[0037]

[0038] In the formula, l is the inner layer number of the rule layer, l = 1, 2, ..., Q; v l φ is the output value of the l-th layer in the rule layer. l This is the output value of the l-th layer in the RBF layer;

[0039] The output value of the output layer is:

[0040] W = [w 1 ,w 2 ,…,w M ] T (5)

[0041] y = Wv (6)

[0042] In the formula, W is the weight matrix, w MLet be the output value of the Mth neuron in the output layer, where M is the total number of neurons in the output layer, and T is the transpose matrix; y is the abnormal condition discrimination matrix, y = [y1, y2], where y1 is the predicted value of adenosine triphosphate (ATP) concentration, and y2 is the predicted value of biochemical oxygen demand (BOD) concentration; v is the output matrix of the regular layer, v = [v1, v2, ..., v Q ] T v Q This is the output value of the Q-th layer in the rule layer.

[0043] Preferably, the number of neurons in the rule layer is equal to that in the RBF layer.

[0044] Preferably, in step 2, a multi-input multi-output fuzzy neural network for predicting ATP and BOD in wastewater treatment processes is trained and constructed using the Levenberg-Marquardt algorithm based on adaptive learning rate.

[0045] The update rule for the Levenberg-Marquardt algorithm based on adaptive learning rate is set as follows:

[0046] Θ(t+1)=Θ(t)+(Ψ(t)+λ(t)×I) -1 Ω(t) (7)

[0047] In the formula, t is time, Θ(t+1) is the variable vector at time t+1, Θ(t) is the variable vector at time t, Ψ(t) is the quasi-Hessian matrix at time t, λ(t) is the adaptive learning rate at time t, 0<λ(t)<1, I is the identity matrix, and Ω(t) is the gradient vector at time t.

[0048] In the update rule of the Levenberg-Marquardt algorithm based on adaptive learning rate, the formula for calculating the adaptive learning rate λ(t) is as follows:

[0049] λ(t)=μ(t)λ(t-1) (8)

[0050]

[0051] In the formula, μ(t) is the adaptive factor at time t, and λ(t-1) is the adaptive learning rate at time t-1; τ min (t) is the smallest eigenvalue of the quasi-Hessian matrix at time t, τ max (t) is the largest eigenvalue of the quasi-Hessian matrix at time t, 0 < τ min (t)<τ max (t);

[0052] The variable vector Θ(t) includes the connection weight vector w(t), the center vector c(t), and the width vector σ(t). The formula for calculating the variable vector is as follows:

[0053] Θ(t)=[w1(t),...,w J (t),c1(t),...,c J (t),...,σ1(t),...,σ J (t)] (10)

[0054] In the formula, w J (t) represents the J-th connection weight vector value at time t, c J (t) represents the J-th center vector value at time t, σ J (t) represents the J-th width vector value at time t;

[0055] The formulas for calculating the quasi-Hessian matrix Ψ(t) and the gradient vector Ω(t) are as follows:

[0056] Ψ(t)=j T (t)j(t) (10)

[0057] Ω(t)=j T (t)e(t) (11)

[0058] e(t) = y d (t)-y(t) (12)

[0059] in,

[0060] The Jacobian vector j(t) is:

[0061]

[0062] In the formula, j(t) is the Jacobian vector at time t, e(t) is the error value of the multi-input multi-output fuzzy neural network at time t, and y d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

[0063] Preferably, step 3 specifically includes the following steps:

[0064] Step 3.1, set the training precision value;

[0065] Step 3.2: Input the water quality parameter samples of chemical wastewater in the training set into the ATP and BOD prediction model of the wastewater treatment process constructed in Step 2. Use the ATP and BOD prediction model of the wastewater treatment process to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration of the chemical wastewater.

[0066] Step 3.3: Compare the predicted values ​​of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration from the ATP and BOD prediction model for the wastewater treatment process with the ATP and BOD concentration values ​​from the water quality parameter samples of chemical wastewater. Calculate the accuracy value of the ATP and BOD prediction model for the wastewater treatment process. If the accuracy value of the ATP and BOD prediction model for the wastewater treatment process is less than the preset accuracy value, proceed to step 3.4; otherwise, proceed to step 3.5.

[0067] Step 3.4: Update the adaptive learning rate of the ATP and BOD prediction model for the wastewater treatment process based on the Levenberg-Marquardt algorithm to obtain the updated ATP and BOD prediction model. Then, randomly select water quality parameter samples of chemical wastewater from the training set and input them into the updated ATP and BOD prediction model. Use the updated ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) of the chemical wastewater samples. After obtaining the predicted values ​​of ATP and BOD concentrations of the chemical wastewater samples, return to step 3.3.

[0068] Step 3.5: Complete the training of the ATP and BOD prediction models for the wastewater treatment process, and obtain the trained ATP and BOD prediction models for the wastewater treatment process.

[0069] Preferably, step 4 specifically includes the following steps:

[0070] Step 4.1: Input the water quality parameter samples of chemical wastewater from the test set into the trained wastewater treatment process ATP and BOD prediction model. Use the wastewater treatment process ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration in the chemical wastewater.

[0071] Step 4.2: Compare the predicted values ​​of ATP and BOD concentrations from the wastewater treatment process prediction model with the ATP and BOD concentrations in the chemical wastewater quality parameter samples. Calculate the accuracy of the wastewater treatment process prediction model. If the accuracy is not less than the preset accuracy, proceed to step 5; otherwise, return to step 3 and continue training the wastewater treatment process prediction model using the training set.

[0072] Preferably, the error values ​​of the ATP and BOD prediction models for the wastewater treatment process are:

[0073]

[0074] The accuracy values ​​of the ATP and BOD prediction models for wastewater treatment processes are:

[0075]

[0076] in,

[0077] e(t) = y d (t)-y(t) (16)

[0078] In the formula, RMSE(t) represents the error value of the ATP and BOD prediction models for the wastewater treatment process, t represents time (t = 1, 2, ..., N), N represents the total number of times, e(t) represents the error value of the multi-input multi-output fuzzy neural network at time t, and y represents the error value of the model. d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

[0079] Preferably, in step 5, the abnormal operating condition discrimination matrix is:

[0080]

[0081] In the formula, ATP is the concentration of adenosine triphosphate (ATP) predicted by the ATP and BOD prediction model for the wastewater treatment process, and BOD is the concentration of biochemical oxygen demand (BOD) predicted by the ATP and BOD prediction model for the wastewater treatment process; C1 is the first type of abnormal operating condition, C2 is the second type of abnormal operating condition, C3 is the third type of abnormal operating condition, and C4 is the fourth type of abnormal operating condition.

[0082] Preferably, in step 5, when determining the type of abnormal operating condition based on the abnormal operating condition discrimination matrix and according to the adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration:

[0083] When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are high, the current operating condition of the chemical wastewater treatment plant is determined to be the first type of abnormal operating condition, C1. This means that the influent concentration of the chemical wastewater treatment plant exceeds its treatment capacity, the BOD concentration is abnormal, and the biochemical treatment capacity is high.

[0084] When adenosine triphosphate (ATP) is high and biochemical oxygen demand (BOD) is low, the current operating condition of the chemical wastewater treatment plant is determined to be the second type of abnormal operating condition, C2, indicating that the biochemical treatment capacity of the chemical wastewater treatment plant is excessive.

[0085] When adenosine triphosphate (ATP) is low and biochemical oxygen demand (BOD) is high, the current operating condition of the chemical wastewater treatment plant is determined to be the third type of abnormal operating condition, C3. It is determined that the sludge in the current chemical wastewater treatment plant is being impacted by water flow. It is necessary to measure the adenosine triphosphate (ATP) in the sludge and analyze the abnormal operating condition in combination with the concentration value of adenosine triphosphate (ATP) in the sludge.

[0086] When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are low, the current operating condition of the chemical wastewater treatment plant is determined to be the fourth abnormal operating condition type C4, indicating that the biochemical treatment capacity of the chemical wastewater treatment plant is low.

[0087] Preferably, when the operating condition of the chemical wastewater treatment plant is determined to be the fourth abnormal operating condition type C4, the sludge disposal capacity of the chemical wastewater treatment plant can be improved by reducing the wastewater treatment volume, increasing the aeration volume, or supplementing the carbon source.

[0088] The beneficial effects of this invention are as follows:

[0089] This patent provides a method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process. By predicting the concentrations of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) in real time during the chemical wastewater treatment process, an abnormal operating condition discrimination matrix is ​​established to assess abnormalities in the petrochemical wastewater treatment process, and to determine in real time whether abnormal operating conditions occur in the activated sludge of the biochemical unit during the chemical wastewater treatment process.

[0090] This invention solves the problem that existing technologies cannot directly and accurately determine the activity of sludge in the biochemical unit during wastewater treatment. By applying a multi-input multi-output fuzzy neural network to the prediction of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration, an abnormal operating condition discrimination matrix is ​​established using the real-time predicted values ​​of ATP and BOD concentrations. This enables accurate identification of abnormal operating conditions such as sludge poisoning and carbon source depletion. By automatically pushing process adjustment plans and emergency remedial measures, the invention effectively ensures the stable compliance of effluent quality in chemical wastewater treatment under abnormal operating conditions.

[0091] This invention helps to promptly detect abnormal conditions of sludge activity during chemical wastewater treatment, effectively preventing sludge poisoning, sludge bulking, and other abnormal conditions, thereby improving the operational efficiency of the wastewater treatment process. Attached Figure Description

[0092] Figure 1 This is a schematic diagram of the structure of the multi-input multi-output fuzzy neural network of the present invention.

[0093] Figure 2 This is a flowchart illustrating the process of identifying abnormal operating conditions using an abnormal operating condition discrimination matrix, as described in this invention. Detailed Implementation

[0094] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0095] Example 1

[0096] This embodiment employs a biochemical potential assessment method for abnormal operating conditions in chemical wastewater treatment processes proposed in this invention. It assesses the biochemical treatment potential of abnormal chemical wastewater operating conditions by real-time prediction of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration. The method specifically includes the following steps:

[0097] Step 1: Obtain water quality parameters from the chemical wastewater treatment plant. In this embodiment, the chemical wastewater mainly consists of propylene oxide unit wastewater and ethylene alkali residue wastewater. After pretreatment, normalization, and principal component analysis, multiple chemical wastewater water quality parameter samples are obtained. These samples are then randomly assigned to training and testing sample sets. The training and testing sample sets together constitute the sample dataset used to train the ATP and BOD prediction models for the wastewater treatment process. The specific steps include:

[0098] Step 1.1: Obtain the chemical wastewater from the chemical wastewater treatment plant. Determine the influent flow rate, water quality parameters, adenosine triphosphate (ATP), and biochemical oxygen demand (BOD) of the chemical wastewater using online instruments or laboratory testing and measurement. The water quality parameters include influent COD concentration, effluent COD concentration, formaldehyde concentration, cumene concentration, methanol concentration, propylene oxide concentration, propylene concentration, formic acid concentration, influent flow rate, sludge retention time, sludge MLSS concentration in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sulfide concentration in the biological treatment tank, volatile phenol concentration in the biological treatment tank, sludge settling ratio, sludge volume index, effluent BOD concentration, ATP concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP).

[0099] Step 1.2: Preprocess the water quality parameters of the chemical wastewater, remove abnormal data, and then normalize the preprocessed water quality parameters to eliminate the influence of dimensional and order-of-magnitude differences on the water quality parameters. Principal component analysis is then used to analyze the obtained water quality parameters. Based on the correlation coefficients and contribution rates of each water quality parameter to the prediction of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration in the wastewater treatment process, the parameters are screened. The screened water quality parameters and influent flow rate are used as input variables. The output variables are determined by combining the measured ATP and BOD concentration values ​​of the chemical wastewater. A sample of chemical wastewater water quality parameters is then formed using the input and output variables.

[0100] In this embodiment, the input variables include the influent chemical oxygen demand (COD) concentration, the sludge MLSS concentration in the biological treatment tank, the dissolved oxygen (DO) concentration in the biological treatment tank, the sludge volume index, the temperature, the influent ammonia nitrogen concentration, the propylene oxide concentration, the formaldehyde concentration, and the influent flow rate. The output variables include the adenosine triphosphate (ATP) concentration and the biochemical oxygen demand (BOD) concentration.

[0101] Step 1.3: Randomly allocate the water quality parameter samples of each chemical wastewater to the training sample set and the test sample set to generate the training sample set and the test sample set, and construct the sample dataset for training the ATP and BOD prediction model of the wastewater treatment process.

[0102] In this embodiment, a total of 350 sets of chemical wastewater quality parameter samples were obtained, of which the training sample set included 200 sets of chemical wastewater quality parameter samples and the test sample set included 150 sets of test samples.

[0103] Step 2: Establish ATP and BOD prediction models for the wastewater treatment process based on a multi-input multi-output fuzzy neural network.

[0104] In this embodiment, the multi-input multi-output fuzzy neural network includes an input layer, an RBF layer, a rule layer, and an output layer, such as... Figure 1 As shown.

[0105] The input layer of the multi-input multi-output fuzzy neural network is configured as follows:

[0106] x i =u i (1)

[0107] x = [x1, x2, ..., x k (2)

[0108] In the formula, i is the index of the input variable, i = 1, 2, ..., k, and k is the total number of input variables. In this embodiment, k = 9; x i Let u be the value of the i-th input variable. iLet x be the input value of the i-th input neuron, and let x be the input vector.

[0109] The RBF layer contains multiple neurons, and the output function value of each neuron in the RBF layer is:

[0110]

[0111] In the formula, φ j c is the output value of the j-th neuron in the RBF layer, where j is the neuron number (j = 1, 2, ..., Q) and Q is the total number of neurons in the RBF layer (Q = 6 in this embodiment). ij σ is the center of the j-th neuron. ij The width of the j-th neuron;

[0112] The number of neurons in the rule layer and the RBF layer are equal, and the output function value of the neurons in the rule layer is:

[0113]

[0114] In the formula, l is the inner layer number of the rule layer, l = 1, 2, ..., Q; v l φ is the output value of the l-th layer in the rule layer. l This is the output value of the l-th layer in the RBF layer;

[0115] The output value of the output layer is:

[0116] W = [w 1 ,w 2 ,…,w M ] T (5)

[0117] y = Wv (6)

[0118] In the formula, W is the weight matrix, w M Let be the output value of the Mth neuron in the output layer, where M is the total number of neurons in the output layer, and T is the transpose matrix; y is the abnormal condition discrimination matrix, y = [y1, y2], where y1 is the predicted value of adenosine triphosphate (ATP) concentration, and y2 is the predicted value of biochemical oxygen demand (BOD) concentration; v is the output matrix of the regular layer, v = [v1, v2, ..., v Q ] T v Q This is the output value of the Q-th layer in the rule layer.

[0119] In this embodiment, a multi-input multi-output fuzzy neural network for predicting ATP and BOD in wastewater treatment processes is trained using the Levenberg-Marquardt algorithm based on adaptive learning rate. By combining the second-order algorithm with the adaptive learning rate, the convergence speed of the Levenberg-Marquardt algorithm is effectively improved by utilizing the adaptive learning rate.

[0120] The update rule for the Levenberg-Marquardt algorithm based on adaptive learning rate is set as follows:

[0121] Θ(t+1)=Θ(t)+(Ψ(t)+λ(t)×I) -1 Ω(t) (7)

[0122] In the formula, t is time, Θ(t+1) is the variable vector at time t+1, Θ(t) is the variable vector at time t, Ψ(t) is the quasi-Hessian matrix at time t, λ(t) is the adaptive learning rate at time t, 0<λ(t)<1, I is the identity matrix, and Ω(t) is the gradient vector at time t.

[0123] In the update rule of the Levenberg-Marquardt algorithm based on adaptive learning rate, the formula for calculating the adaptive learning rate λ(t) is as follows:

[0124] λ(t)=μ(t)λ(t-1) (8)

[0125]

[0126] In the formula, μ(t) is the adaptive factor at time t, and λ(t-1) is the adaptive learning rate at time t-1; τ min (t) is the smallest eigenvalue of the quasi-Hessian matrix at time t, τ max (t) is the largest eigenvalue of the quasi-Hessian matrix at time t, 0 < τ min (t)<τ max (t).

[0127] The variable vector Θ(t) includes the connection weight vector w(t), the center vector c(t), and the width vector σ(t). The formula for calculating the variable vector is as follows:

[0128] Θ(t)=[w1(t),...,w J (t),c1(t),...,c J (t),...,σ1(t),...,σ J (t)] (10)

[0129] In the formula, w J (t) represents the J-th connection weight vector value at time t, cJ (t) represents the J-th center vector value at time t, σ J (t) represents the J-th width vector value at time t.

[0130] The formulas for calculating the quasi-Hessian matrix Ψ(t) and the gradient vector Ω(t) are as follows:

[0131] Ψ(t)=j T (t)j(t) (10)

[0132] Ω(t)=j T (t)e(t) (11)

[0133] e(t) = y d (t)-y(t) (12)

[0134] in,

[0135] The Jacobian vector j(t) is:

[0136]

[0137] In the formula, j(t) is the Jacobian vector at time t, e(t) is the error value of the multi-input multi-output fuzzy neural network at time t, and y d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

[0138] Step 3: Train the ATP and BOD prediction models for the wastewater treatment process using the training set to obtain the trained ATP and BOD prediction models for the wastewater treatment process. This specifically includes the following steps:

[0139] Step 3.1: Set the training precision value.

[0140] Step 3.2: Input the water quality parameter samples of chemical wastewater from the training set into the ATP and BOD prediction model of the wastewater treatment process constructed in Step 2. Use the ATP and BOD prediction model of the wastewater treatment process to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration of the chemical wastewater.

[0141] Step 3.3: Compare the predicted values ​​of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration from the ATP and BOD prediction model for the wastewater treatment process with the ATP and BOD concentration values ​​from the water quality parameter samples of chemical wastewater. Calculate the accuracy value of the ATP and BOD prediction model for the wastewater treatment process. If the accuracy value of the ATP and BOD prediction model for the wastewater treatment process is less than the preset accuracy value, proceed to step 3.4; otherwise, proceed to step 3.5.

[0142] Step 3.4: Update the adaptive learning rate of the ATP and BOD prediction model for the wastewater treatment process based on the Levenberg-Marquardt algorithm to obtain the updated ATP and BOD prediction model. Then, randomly select water quality parameter samples of chemical wastewater from the training set and input them into the updated ATP and BOD prediction model. Use the updated ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) of the chemical wastewater samples. After obtaining the predicted values ​​of ATP and BOD concentrations of the chemical wastewater samples, return to step 3.3.

[0143] Step 3.5: Complete the training of the ATP and BOD prediction models for the wastewater treatment process, and obtain the trained ATP and BOD prediction models for the wastewater treatment process.

[0144] Step 4: Validate the prediction performance of the trained ATP and BOD prediction models for wastewater treatment processes using the test set. If the prediction performance of the trained ATP and BOD prediction models for wastewater treatment processes has reached its optimal level, proceed to Step 5. Otherwise, return to Step 3 and continue training the ATP and BOD prediction models for wastewater treatment processes using the training set. This specifically includes the following steps:

[0145] Step 4.1: Input the water quality parameter samples of chemical wastewater from the test set into the trained wastewater treatment process ATP and BOD prediction model. Use the wastewater treatment process ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration in the chemical wastewater.

[0146] Step 4.2: Compare the predicted values ​​of ATP and BOD concentrations from the wastewater treatment process prediction model with the ATP and BOD concentrations in the chemical wastewater quality parameter samples. Calculate the accuracy of the wastewater treatment process prediction model. If the accuracy is not less than the preset accuracy, proceed to step 5; otherwise, return to step 3 and continue training the wastewater treatment process prediction model using the training set.

[0147] In this embodiment, the error values ​​of the ATP and BOD prediction models for the wastewater treatment process are:

[0148]

[0149] The accuracy values ​​of the ATP and BOD prediction models for wastewater treatment processes are:

[0150]

[0151] in,

[0152] e(t) = y d (t)-y(t) (16)

[0153] In the formula, RMSE(t) represents the error value of the ATP and BOD prediction models for the wastewater treatment process, t represents time (t = 1, 2, ..., N), N represents the total number of times, e(t) represents the error value of the multi-input multi-output fuzzy neural network at time t, and y represents the error value of the model. d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

[0154] Step 5: Using the ATP and BOD prediction model of the wastewater treatment process after testing, predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) in real time based on the water quality parameters of the chemical wastewater, and generate an abnormal operating condition discrimination matrix, such as... Figure 2 As shown, the abnormal operating condition discrimination matrix is ​​used to identify the type of abnormal operating condition, and the sludge condition and biochemical treatment potential of the chemical wastewater treatment plant are evaluated based on the concentration values ​​of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD), so as to adjust the load operation plan of the chemical wastewater treatment plant.

[0155] The abnormal operating condition discrimination matrix is ​​as follows:

[0156]

[0157] In the formula, ATP is the concentration of adenosine triphosphate (ATP) predicted by the ATP and BOD prediction model for the wastewater treatment process, and BOD is the concentration of biochemical oxygen demand (BOD) predicted by the ATP and BOD prediction model for the wastewater treatment process; C1 is the first type of abnormal operating condition, C2 is the second type of abnormal operating condition, C3 is the third type of abnormal operating condition, and C4 is the fourth type of abnormal operating condition.

[0158] In this embodiment, when the ATP concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is greater than 1.2, it is characterized as high ATP; when the ATP concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is less than 1.2, it is characterized as low ATP. Similarly, when the BOD concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is greater than 1500, it is characterized as high BOD; when the BOD concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is less than 1500, it is characterized as low BOD.

[0159] When determining the type of abnormal operating condition based on the abnormal operating condition discrimination matrix, according to the concentration values ​​of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD):

[0160] When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are high, the current operating condition of the chemical wastewater treatment plant is determined to be the first type of abnormal operating condition, C1. The influent concentration of the chemical wastewater treatment plant exceeds its treatment capacity, the BOD concentration is abnormal, but the sludge activity is relatively high, and the biochemical treatment capacity is high.

[0161] When adenosine triphosphate (ATP) is high and biochemical oxygen demand (BOD) is low, the current operating condition of the chemical wastewater treatment plant is determined to be the second type of abnormal operating condition, C2. The operating condition is good, the biochemical treatment capacity of the chemical wastewater treatment plant is sufficient, and the aeration rate can be appropriately reduced or the treatment load increased.

[0162] When adenosine triphosphate (ATP) is low and biochemical oxygen demand (BOD) is high, the current operating condition of the chemical wastewater treatment plant is determined to be the third type of abnormal operating condition, C3. This indicates that the sludge in the chemical wastewater treatment plant is being impacted by water flow. It is necessary to measure the ATP concentration in the sludge and analyze the abnormal operating condition in conjunction with the ATP concentration value of the sludge to determine the possible abnormal state of the sludge.

[0163] When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are low, the current operating condition of the chemical wastewater treatment plant is determined to be the fourth abnormal operating condition type C4, indicating low biochemical treatment potential and low biochemical treatment capacity. The sludge treatment capacity of the chemical wastewater treatment plant can be improved by reducing the wastewater treatment volume, increasing the aeration volume, or supplementing carbon sources.

[0164] Therefore, the biochemical potential assessment method for abnormal operating conditions in the chemical wastewater treatment process in this embodiment establishes an abnormal operating condition discrimination matrix by using the real-time predicted values ​​of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration. This enables accurate identification of abnormal operating conditions such as sludge poisoning and carbon source depletion. By automatically pushing process adjustment plans and emergency remedial measures, it effectively ensures that the effluent quality of chemical wastewater treatment meets the standards under abnormal operating conditions.

[0165] Example 2

[0166] This embodiment employs a biochemical potential assessment method for abnormal operating conditions in chemical wastewater treatment processes proposed in this invention. It assesses the biochemical treatment potential of abnormal chemical wastewater operating conditions by real-time prediction of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration. The method specifically includes the following steps:

[0167] Step 1: Obtain water quality parameters from the chemical wastewater treatment plant. In this embodiment, the chemical wastewater mainly consists of purified terephthalic acid wastewater and ethylene alkali residue wastewater. After pretreatment, normalization, and principal component analysis, multiple chemical wastewater water quality parameter samples are obtained. These samples are then randomly assigned to training and testing sample sets. The training and testing sample sets together constitute the sample dataset used to train the ATP and BOD prediction models for the wastewater treatment process. The specific steps include:

[0168] Step 1.1: Obtain the chemical wastewater from the chemical wastewater treatment plant. Determine the influent flow rate, water quality parameters, adenosine triphosphate (ATP), and biochemical oxygen demand (BOD) of the chemical wastewater using online instruments or laboratory testing and measurement. The water quality parameters include influent COD concentration, effluent COD concentration, formaldehyde concentration, terephthalic acid concentration, isophthalic acid concentration, acetic acid concentration, influent flow rate, sludge retention time, sludge MLSS concentration in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sludge settling ratio, sludge volume index, effluent BOD concentration, ATP concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP).

[0169] Step 1.2: Preprocess the water quality parameters of the chemical wastewater, remove abnormal data, and then normalize the preprocessed water quality parameters to eliminate the influence of dimensional and order-of-magnitude differences on the water quality parameters. Principal component analysis is then used to analyze the obtained water quality parameters. Based on the correlation coefficients and contribution rates of each water quality parameter to the prediction of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration in the wastewater treatment process, the parameters are screened. The screened water quality parameters and influent flow rate are used as input variables. The output variables are determined by combining the measured ATP and BOD concentration values ​​of the chemical wastewater. A sample of chemical wastewater water quality parameters is then formed using the input and output variables.

[0170] In this embodiment, the input variables include the influent chemical oxygen demand (COD) concentration, the sludge MLSS concentration in the biological treatment tank, the dissolved oxygen (DO) concentration in the biological treatment tank, the sludge volume index, the temperature, the influent ammonia nitrogen concentration, the formaldehyde concentration, the terephthalic acid concentration, the influent flow rate, and the isophthalic acid concentration. The output variables include the adenosine triphosphate (ATP) concentration and the biochemical oxygen demand (BOD) concentration.

[0171] Step 1.3: Randomly allocate the water quality parameter samples of each chemical wastewater to the training sample set and the test sample set to generate the training sample set and the test sample set, and construct the sample dataset for training the ATP and BOD prediction model of the wastewater treatment process.

[0172] In this embodiment, a total of 300 sets of chemical wastewater quality parameter samples were obtained, of which the training sample set included 150 sets of chemical wastewater quality parameter samples and the test sample set included 150 sets of test samples.

[0173] Step 2: Establish ATP and BOD prediction models for the wastewater treatment process based on a multi-input multi-output fuzzy neural network.

[0174] In this embodiment, the multi-input multi-output fuzzy neural network includes an input layer, an RBF layer, a rule layer, and an output layer, such as... Figure 1 As shown.

[0175] The input layer of the multi-input multi-output fuzzy neural network is configured as follows:

[0176] x i =u i (1)

[0177] x = [x1, x2, ..., x k (2)

[0178] In the formula, i is the index of the input variable, i = 1, 2, ..., k, and k is the total number of input variables. In this embodiment, k = 10; x i Let u be the value of the i-th input variable. iLet x be the input value of the i-th input neuron, and let x be the input vector.

[0179] The RBF layer contains multiple neurons, and the output function value of each neuron in the RBF layer is:

[0180]

[0181] In the formula, φ j c is the output value of the j-th neuron in the RBF layer, where j is the neuron number (j = 1, 2, ..., Q) and Q is the total number of neurons in the RBF layer (Q = 7 in this embodiment). ij σ is the center of the j-th neuron. ij Let be the width of the j-th neuron.

[0182] The number of neurons in the rule layer and the RBF layer are equal, and the output function value of the neurons in the rule layer is:

[0183]

[0184] In the formula, l is the inner layer number of the rule layer, l = 1, 2, ..., Q; v l φ is the output value of the l-th layer in the rule layer. l This is the output value of the l-th layer in the RBF layer.

[0185] The output value of the output layer is:

[0186] W = [w 1 ,w 2 ,…,w M ] T (5)

[0187] y = Wv (6)

[0188] In the formula, W is the weight matrix, w M Let be the output value of the Mth neuron in the output layer, where M is the total number of neurons in the output layer, and T is the transpose matrix; y is the abnormal condition discrimination matrix, y = [y1, y2], where y1 is the predicted value of adenosine triphosphate (ATP) concentration, and y2 is the predicted value of biochemical oxygen demand (BOD) concentration; v is the output matrix of the regular layer, v = [v1, v2, ..., v Q ] T v Q This is the output value of the Q-th layer in the rule layer.

[0189] In this embodiment, a multi-input multi-output fuzzy neural network for predicting ATP and BOD in wastewater treatment processes is trained using the Levenberg-Marquardt algorithm based on adaptive learning rate. By combining the second-order algorithm with the adaptive learning rate, the convergence speed of the Levenberg-Marquardt algorithm is effectively improved by utilizing the adaptive learning rate.

[0190] The update rule for the Levenberg-Marquardt algorithm based on adaptive learning rate is set as follows:

[0191] Θ(t+1)=Θ(t)+(Ψ(t)+λ(t)×I) -1 Ω(t) (7)

[0192] In the formula, t is time, Θ(t+1) is the variable vector at time t+1, Θ(t) is the variable vector at time t, Ψ(t) is the quasi-Hessian matrix at time t, λ(t) is the adaptive learning rate at time t, 0<λ(t)<1, I is the identity matrix, and Ω(t) is the gradient vector at time t.

[0193] In the update rule of the Levenberg-Marquardt algorithm based on adaptive learning rate, the formula for calculating the adaptive learning rate λ(t) is as follows:

[0194] λ(t)=μ(t)λ(t-1) (8)

[0195]

[0196] In the formula, μ(t) is the adaptive factor at time t, and λ(t-1) is the adaptive learning rate at time t-1; τ min (t) is the smallest eigenvalue of the quasi-Hessian matrix at time t, τ max (t) is the largest eigenvalue of the quasi-Hessian matrix at time t, 0 < τ min (t)<τ max (t).

[0197] The variable vector Θ(t) includes the connection weight vector w(t), the center vector c(t), and the width vector σ(t). The formula for calculating the variable vector is as follows:

[0198] Θ(t)=[w1(t),...,w J (t),c1(t),...,c J (t),...,σ1(t),...,σ J (t)] (10)

[0199] In the formula, w J (t) represents the J-th connection weight vector value at time t, cJ (t) represents the J-th center vector value at time t, σ J (t) represents the J-th width vector value at time t.

[0200] The formulas for calculating the quasi-Hessian matrix Ψ(t) and the gradient vector Ω(t) are as follows:

[0201] Ψ(t)=j T (t)j(t) (10)

[0202] Ω(t)=j T (t)e(t) (11)

[0203] e(t) = y d (t)-y(t) (12)

[0204] in,

[0205] The Jacobian vector j(t) is:

[0206]

[0207] In the formula, j(t) is the Jacobian vector at time t, e(t) is the error value of the multi-input multi-output fuzzy neural network at time t, and y d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

[0208] Step 3: Train the ATP and BOD prediction models for the wastewater treatment process using the training set to obtain the trained ATP and BOD prediction models for the wastewater treatment process. This specifically includes the following steps:

[0209] Step 3.1: Set the training precision value.

[0210] Step 3.2: Input the water quality parameter samples of chemical wastewater from the training set into the ATP and BOD prediction model of the wastewater treatment process constructed in Step 2. Use the ATP and BOD prediction model of the wastewater treatment process to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration of the chemical wastewater.

[0211] Step 3.3: Compare the predicted values ​​of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration from the ATP and BOD prediction model for the wastewater treatment process with the ATP and BOD concentration values ​​from the water quality parameter samples of chemical wastewater. Calculate the accuracy value of the ATP and BOD prediction model for the wastewater treatment process. If the accuracy value of the ATP and BOD prediction model for the wastewater treatment process is less than the preset accuracy value, proceed to step 3.4; otherwise, proceed to step 3.5.

[0212] Step 3.4: Update the adaptive learning rate of the ATP and BOD prediction model for the wastewater treatment process based on the Levenberg-Marquardt algorithm to obtain the updated ATP and BOD prediction model. Then, randomly select water quality parameter samples of chemical wastewater from the training set and input them into the updated ATP and BOD prediction model. Use the updated ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) of the chemical wastewater samples. After obtaining the predicted values ​​of ATP and BOD concentrations of the chemical wastewater samples, return to step 3.3.

[0213] Step 3.5: Complete the training of the ATP and BOD prediction models for the wastewater treatment process, and obtain the trained ATP and BOD prediction models for the wastewater treatment process.

[0214] Step 4: Validate the prediction performance of the trained ATP and BOD prediction models for wastewater treatment processes using the test set. If the prediction performance of the trained ATP and BOD prediction models for wastewater treatment processes has reached its optimal level, proceed to Step 5. Otherwise, return to Step 3 and continue training the ATP and BOD prediction models for wastewater treatment processes using the training set. This specifically includes the following steps:

[0215] Step 4.1: Input the water quality parameter samples of chemical wastewater from the test set into the trained wastewater treatment process ATP and BOD prediction model. Use the wastewater treatment process ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration in the chemical wastewater.

[0216] Step 4.2: Compare the predicted values ​​of ATP and BOD concentrations from the wastewater treatment process prediction model with the ATP and BOD concentrations in the chemical wastewater quality parameter samples. Calculate the accuracy of the wastewater treatment process prediction model. If the accuracy is not less than the preset accuracy, proceed to step 5; otherwise, return to step 3 and continue training the wastewater treatment process prediction model using the training set.

[0217] In this embodiment, the error values ​​of the ATP and BOD prediction models for the wastewater treatment process are:

[0218]

[0219] The accuracy values ​​of the ATP and BOD prediction models for wastewater treatment processes are:

[0220]

[0221] in,

[0222] e(t) = y d (t)-y(t) (16)

[0223] In the formula, RMSE(t) represents the error value of the ATP and BOD prediction models for the wastewater treatment process, t represents time (t = 1, 2, ..., N), N represents the total number of times, e(t) represents the error value of the multi-input multi-output fuzzy neural network at time t, and y represents the error value of the model. d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

[0224] Step 5: Using the ATP and BOD prediction model of the wastewater treatment process after testing, predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) in real time based on the water quality parameters of the chemical wastewater, and generate an abnormal operating condition discrimination matrix, such as... Figure 2 As shown, the abnormal operating condition discrimination matrix is ​​used to identify the type of abnormal operating condition, and the sludge condition and biochemical treatment potential of the chemical wastewater treatment plant are evaluated based on the concentration values ​​of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD), so as to adjust the load operation plan of the chemical wastewater treatment plant.

[0225] The abnormal operating condition discrimination matrix is ​​as follows:

[0226]

[0227] In the formula, ATP is the concentration of adenosine triphosphate (ATP) predicted by the ATP and BOD prediction model for the wastewater treatment process, and BOD is the concentration of biochemical oxygen demand (BOD) predicted by the ATP and BOD prediction model for the wastewater treatment process; C1 is the first type of abnormal operating condition, C2 is the second type of abnormal operating condition, C3 is the third type of abnormal operating condition, and C4 is the fourth type of abnormal operating condition.

[0228] In this embodiment, when the ATP concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is greater than 1.3, it is characterized as high ATP; when the ATP concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is less than 1.3, it is characterized as low ATP. Similarly, when the BOD concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is greater than 1300, it is characterized as high BOD; when the BOD concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is less than 1300, it is characterized as low BOD.

[0229] When determining the type of abnormal operating condition based on the abnormal operating condition discrimination matrix, according to the concentration values ​​of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD):

[0230] When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are high, the current operating condition of the chemical wastewater treatment plant is determined to be the first type of abnormal operating condition, C1. The influent concentration of the chemical wastewater treatment plant exceeds its treatment capacity, the BOD concentration is abnormal, but the sludge activity is relatively high, and the biochemical treatment capacity is high.

[0231] When adenosine triphosphate (ATP) is high and biochemical oxygen demand (BOD) is low, the current operating condition of the chemical wastewater treatment plant is determined to be the second type of abnormal operating condition, C2. The operating condition is good, the biochemical treatment capacity of the chemical wastewater treatment plant is sufficient, and the aeration rate can be appropriately reduced or the treatment load increased.

[0232] When adenosine triphosphate (ATP) is low and biochemical oxygen demand (BOD) is high, the current operating condition of the chemical wastewater treatment plant is determined to be the third type of abnormal operating condition, C3. This indicates that the sludge in the chemical wastewater treatment plant is being impacted by water flow. It is necessary to measure the ATP concentration in the sludge and analyze the abnormal operating condition in conjunction with the ATP concentration value of the sludge to determine the possible abnormal state of the sludge.

[0233] When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are low, the current operating condition of the chemical wastewater treatment plant is determined to be the fourth abnormal operating condition type C4, indicating low biochemical treatment potential and low biochemical treatment capacity. The sludge treatment capacity of the chemical wastewater treatment plant can be improved by reducing the wastewater treatment volume, increasing the aeration volume, or supplementing carbon sources.

[0234] Therefore, the biochemical potential assessment method for abnormal operating conditions in the chemical wastewater treatment process in this embodiment establishes an abnormal operating condition discrimination matrix by using the real-time predicted values ​​of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration. This enables accurate identification of abnormal operating conditions such as sludge poisoning and carbon source depletion. By automatically pushing process adjustment plans and emergency remedial measures, it effectively ensures that the effluent quality of chemical wastewater treatment meets the standards under abnormal operating conditions.

[0235] Example 3

[0236] This embodiment employs a biochemical potential assessment method for abnormal operating conditions in chemical wastewater treatment processes proposed in this invention. It assesses the biochemical treatment potential of abnormal chemical wastewater operating conditions by real-time prediction of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration. The method specifically includes the following steps:

[0237] Step 1: Obtain water quality parameters from the chemical wastewater treatment plant. In this embodiment, the chemical wastewater mainly consists of propylene oxide wastewater, purified terephthalic acid wastewater, and ethylene alkali residue wastewater. After pretreatment, normalization, and principal component analysis, multiple chemical wastewater water quality parameter samples are obtained. These samples are then randomly assigned to training and testing sample sets. The training and testing sample sets together constitute the sample dataset used to train the ATP and BOD prediction models for the wastewater treatment process. The specific steps include:

[0238] Step 1.1: Obtain the chemical wastewater from the chemical wastewater treatment plant. Determine the influent flow rate, water quality parameters, adenosine triphosphate (ATP), and biochemical oxygen demand (BOD) of the chemical wastewater using online instruments or laboratory testing. The water quality parameters include influent COD concentration, effluent COD concentration, formaldehyde concentration, cumene concentration, methanol concentration, propylene oxide concentration, propylene concentration, formic acid concentration, terephthalic acid concentration, isophthalic acid concentration, acetic acid concentration, influent flow rate, sludge retention time, sludge MLSS concentration in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sludge settling ratio, sludge volume index, effluent BOD concentration, ATP concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP).

[0239] Step 1.2: Preprocess the water quality parameters of the chemical wastewater, remove abnormal data, and then normalize the preprocessed water quality parameters to eliminate the influence of dimensional and order-of-magnitude differences on the water quality parameters. Principal component analysis is then used to analyze the obtained water quality parameters. Based on the correlation coefficients and contribution rates of each water quality parameter to the prediction of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration in the wastewater treatment process, the parameters are screened. The screened water quality parameters and influent flow rate are used as input variables. The output variables are determined by combining the measured ATP and BOD concentration values ​​of the chemical wastewater. A sample of chemical wastewater water quality parameters is then formed using the input and output variables.

[0240] In this embodiment, the input variables include the influent chemical oxygen demand (COD) concentration, the sludge MLSS concentration in the biological treatment tank, the dissolved oxygen (DO) concentration in the biological treatment tank, the sludge volume index, the temperature, the influent ammonia nitrogen concentration, the formaldehyde concentration, the cumene concentration, the methanol concentration, the propylene oxide concentration, the terephthalic acid concentration, and the influent flow rate. The output variables include the adenosine triphosphate (ATP) concentration and the biochemical oxygen demand (BOD) concentration.

[0241] Step 1.3: Randomly allocate the water quality parameter samples of each chemical wastewater to the training sample set and the test sample set to generate the training sample set and the test sample set, and construct the sample dataset for training the ATP and BOD prediction model of the wastewater treatment process.

[0242] In this embodiment, a total of 400 sets of chemical wastewater quality parameter samples were obtained, of which the training sample set included 200 sets of chemical wastewater quality parameter samples and the test sample set included 200 sets of test samples.

[0243] Step 2: Establish ATP and BOD prediction models for the wastewater treatment process based on a multi-input multi-output fuzzy neural network.

[0244] In this embodiment, the multi-input multi-output fuzzy neural network includes an input layer, an RBF layer, a rule layer, and an output layer, such as... Figure 1 As shown.

[0245] The input layer of the multi-input multi-output fuzzy neural network is configured as follows:

[0246] x i =u i (1)

[0247] x = [x1, x2, ..., x k (2)

[0248] In the formula, i is the index of the input variable, i = 1, 2, ..., k, and k is the total number of input variables. In this embodiment, k = 12; x i Let u be the value of the i-th input variable.i Let x be the input value of the i-th input neuron, and let x be the input vector.

[0249] The RBF layer contains multiple neurons, and the output function value of each neuron in the RBF layer is:

[0250]

[0251] In the formula, φ j c is the output value of the j-th neuron in the RBF layer, where j is the neuron number (j = 1, 2, ..., Q) and Q is the total number of neurons in the RBF layer (Q = 7 in this embodiment). ij σ is the center of the j-th neuron. ij Let be the width of the j-th neuron.

[0252] The number of neurons in the rule layer and the RBF layer are equal, and the output function value of the neurons in the rule layer is:

[0253]

[0254] In the formula, l is the inner layer number of the rule layer, l = 1, 2, ..., Q; v l φ is the output value of the l-th layer in the rule layer. l This is the output value of the l-th layer in the RBF layer.

[0255] The output value of the output layer is:

[0256] W = [w 1 ,w 2 ,…,w M ] T (5)

[0257] y = Wv (6)

[0258] In the formula, W is the weight matrix, w M Let be the output value of the Mth neuron in the output layer, where M is the total number of neurons in the output layer, and T is the transpose matrix; y is the abnormal condition discrimination matrix, y = [y1, y2], where y1 is the predicted value of adenosine triphosphate (ATP) concentration, and y2 is the predicted value of biochemical oxygen demand (BOD) concentration; v is the output matrix of the regular layer, v = [v1, v2, ..., v Q ] T v Q This is the output value of the Q-th layer in the rule layer.

[0259] In this embodiment, a multi-input multi-output fuzzy neural network for predicting ATP and BOD in wastewater treatment processes is trained using the Levenberg-Marquardt algorithm based on adaptive learning rate. By combining the second-order algorithm with the adaptive learning rate, the convergence speed of the Levenberg-Marquardt algorithm is effectively improved by utilizing the adaptive learning rate.

[0260] The update rule for the Levenberg-Marquardt algorithm based on adaptive learning rate is set as follows:

[0261] Θ(t+1)=Θ(t)+(Ψ(t)+λ(t)×I) -1 Ω(t) (7)

[0262] In the formula, t is time, Θ(t+1) is the variable vector at time t+1, Θ(t) is the variable vector at time t, Ψ(t) is the quasi-Hessian matrix at time t, λ(t) is the adaptive learning rate at time t, 0<λ(t)<1, I is the identity matrix, and Ω(t) is the gradient vector at time t.

[0263] In the update rule of the Levenberg-Marquardt algorithm based on adaptive learning rate, the formula for calculating the adaptive learning rate λ(t) is as follows:

[0264] λ(t)=μ(t)λ(t-1) (8)

[0265]

[0266] In the formula, μ(t) is the adaptive factor at time t, and λ(t-1) is the adaptive learning rate at time t-1; τ min (t) is the smallest eigenvalue of the quasi-Hessian matrix at time t, τ max (t) is the largest eigenvalue of the quasi-Hessian matrix at time t, 0 < τ min (t)<τ max (t).

[0267] The variable vector Θ(t) includes the connection weight vector w(t), the center vector c(t), and the width vector σ(t). The formula for calculating the variable vector is as follows:

[0268] Θ(t)=[w1(t),...,w J (t),c1(t),...,c J (t),...,σ1(t),...,σ J (t)] (10)

[0269] In the formula, w J (t) represents the J-th connection weight vector value at time t, cJ (t) represents the J-th center vector value at time t, σ J (t) represents the J-th width vector value at time t.

[0270] The formulas for calculating the quasi-Hessian matrix Ψ(t) and the gradient vector Ω(t) are as follows:

[0271] Ψ(t)=j T (t)j(t) (10)

[0272] Ω(t)=j T (t)e(t) (11)

[0273] e(t) = y d (t)-y(t) (12)

[0274] in,

[0275] The Jacobian vector j(t) is:

[0276]

[0277] In the formula, j(t) is the Jacobian vector at time t, e(t) is the error value of the multi-input multi-output fuzzy neural network at time t, and y d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

[0278] Step 3: Train the ATP and BOD prediction models for the wastewater treatment process using the training set to obtain the trained ATP and BOD prediction models for the wastewater treatment process. This specifically includes the following steps:

[0279] Step 3.1: Set the training precision value.

[0280] Step 3.2: Input the water quality parameter samples of chemical wastewater from the training set into the ATP and BOD prediction model of the wastewater treatment process constructed in Step 2. Use the ATP and BOD prediction model of the wastewater treatment process to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration of the chemical wastewater.

[0281] Step 3.3: Compare the predicted values ​​of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration from the ATP and BOD prediction model for the wastewater treatment process with the ATP and BOD concentration values ​​from the water quality parameter samples of chemical wastewater. Calculate the accuracy value of the ATP and BOD prediction model for the wastewater treatment process. If the accuracy value of the ATP and BOD prediction model for the wastewater treatment process is less than the preset accuracy value, proceed to step 3.4; otherwise, proceed to step 3.5.

[0282] Step 3.4: Update the adaptive learning rate of the ATP and BOD prediction model for the wastewater treatment process based on the Levenberg-Marquardt algorithm to obtain the updated ATP and BOD prediction model. Then, randomly select water quality parameter samples of chemical wastewater from the training set and input them into the updated ATP and BOD prediction model. Use the updated ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) of the chemical wastewater samples. After obtaining the predicted values ​​of ATP and BOD concentrations of the chemical wastewater samples, return to step 3.3.

[0283] Step 3.5: Complete the training of the ATP and BOD prediction models for the wastewater treatment process, and obtain the trained ATP and BOD prediction models for the wastewater treatment process.

[0284] Step 4: Validate the prediction performance of the trained ATP and BOD prediction models for wastewater treatment processes using the test set. If the prediction performance of the trained ATP and BOD prediction models for wastewater treatment processes has reached its optimal level, proceed to Step 5. Otherwise, return to Step 3 and continue training the ATP and BOD prediction models for wastewater treatment processes using the training set. This specifically includes the following steps:

[0285] Step 4.1: Input the water quality parameter samples of chemical wastewater from the test set into the trained wastewater treatment process ATP and BOD prediction model. Use the wastewater treatment process ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration in the chemical wastewater.

[0286] Step 4.2: Compare the predicted values ​​of ATP and BOD concentrations from the wastewater treatment process prediction model with the ATP and BOD concentrations in the chemical wastewater quality parameter samples. Calculate the accuracy of the wastewater treatment process prediction model. If the accuracy is not less than the preset accuracy, proceed to step 5; otherwise, return to step 3 and continue training the wastewater treatment process prediction model using the training set.

[0287] In this embodiment, the error values ​​of the ATP and BOD prediction models for the wastewater treatment process are:

[0288]

[0289] The accuracy values ​​of the ATP and BOD prediction models for wastewater treatment processes are:

[0290]

[0291] in,

[0292] e(t) = y d (t)-y(t) (16)

[0293] In the formula, RMSE(t) represents the error value of the ATP and BOD prediction models for the wastewater treatment process, t represents time (t = 1, 2, ..., N), N represents the total number of times, e(t) represents the error value of the multi-input multi-output fuzzy neural network at time t, and y represents the error value of the model. d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

[0294] Step 5: Using the ATP and BOD prediction model of the wastewater treatment process after testing, predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) in real time based on the water quality parameters of the chemical wastewater, and generate an abnormal operating condition discrimination matrix, such as... Figure 2 As shown, the abnormal operating condition discrimination matrix is ​​used to identify the type of abnormal operating condition, and the sludge condition and biochemical treatment potential of the chemical wastewater treatment plant are evaluated based on the concentration values ​​of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD), so as to adjust the load operation plan of the chemical wastewater treatment plant.

[0295] The abnormal operating condition discrimination matrix is ​​as follows:

[0296]

[0297] In the formula, ATP is the concentration of adenosine triphosphate (ATP) predicted by the ATP and BOD prediction model for the wastewater treatment process, and BOD is the concentration of biochemical oxygen demand (BOD) predicted by the ATP and BOD prediction model for the wastewater treatment process; C1 is the first type of abnormal operating condition, C2 is the second type of abnormal operating condition, C3 is the third type of abnormal operating condition, and C4 is the fourth type of abnormal operating condition.

[0298] In this embodiment, when the ATP concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is greater than 1.1, it is characterized as high ATP; when the ATP concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is less than 1.1, it is characterized as low ATP. Similarly, when the BOD concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is greater than 1200, it is characterized as high BOD; when the BOD concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is less than 1200, it is characterized as low BOD.

[0299] When determining the type of abnormal operating condition based on the abnormal operating condition discrimination matrix, according to the concentration values ​​of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD):

[0300] When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are high, the current operating condition of the chemical wastewater treatment plant is determined to be the first type of abnormal operating condition, C1. The influent concentration of the chemical wastewater treatment plant exceeds its treatment capacity, the BOD concentration is abnormal, but the sludge activity is relatively high, and the biochemical treatment capacity is high.

[0301] When adenosine triphosphate (ATP) is high and biochemical oxygen demand (BOD) is low, the current operating condition of the chemical wastewater treatment plant is determined to be the second type of abnormal operating condition, C2. The operating condition is good, the biochemical treatment capacity of the chemical wastewater treatment plant is sufficient, and the aeration rate can be appropriately reduced or the treatment load increased.

[0302] When adenosine triphosphate (ATP) is low and biochemical oxygen demand (BOD) is high, the current operating condition of the chemical wastewater treatment plant is determined to be the third type of abnormal operating condition, C3. This indicates that the sludge in the chemical wastewater treatment plant is being impacted by water flow. It is necessary to measure the ATP concentration in the sludge and analyze the abnormal operating condition in conjunction with the ATP concentration value of the sludge to determine the possible abnormal state of the sludge.

[0303] When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are low, the current operating condition of the chemical wastewater treatment plant is determined to be the fourth abnormal operating condition type C4, indicating low biochemical treatment potential and low biochemical treatment capacity. The sludge treatment capacity of the chemical wastewater treatment plant can be improved by reducing the wastewater treatment volume, increasing the aeration volume, or supplementing carbon sources.

[0304] Therefore, the biochemical potential assessment method for abnormal operating conditions in the chemical wastewater treatment process in this embodiment establishes an abnormal operating condition discrimination matrix by using the real-time predicted values ​​of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration. This enables accurate identification of abnormal operating conditions such as sludge poisoning and carbon source depletion. By automatically pushing process adjustment plans and emergency remedial measures, it effectively ensures that the effluent quality of chemical wastewater treatment meets the standards under abnormal operating conditions.

[0305] Example 4

[0306] This embodiment employs a biochemical potential assessment method for abnormal operating conditions in chemical wastewater treatment processes proposed in this invention. It assesses the biochemical treatment potential of abnormal chemical wastewater operating conditions by real-time prediction of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration. The method specifically includes the following steps:

[0307] Step 1: Obtain water quality parameters from the chemical wastewater treatment plant. In this embodiment, the chemical wastewater mainly consists of propylene oxide wastewater, aromatic hydrocarbon wastewater, purified terephthalic acid wastewater, and ethylene alkali residue wastewater. After pretreatment, normalization, and principal component analysis, multiple chemical wastewater water quality parameter samples are obtained. These samples are then randomly assigned to training and testing sample sets. Together, these sets constitute the sample dataset used to train the ATP and BOD prediction models for the wastewater treatment process. The specific steps include:

[0308] Step 1.1: Obtain the chemical wastewater from the chemical wastewater treatment plant. Determine the influent flow rate, water quality parameters, adenosine triphosphate (ATP), and biochemical oxygen demand (BOD) of the chemical wastewater using online instruments or laboratory testing and measurement. The water quality parameters include influent COD concentration, effluent COD concentration, cyanide concentration, polycyclic aromatic hydrocarbons (PAHs), volatile phenol concentration, sulfide concentration, formaldehyde concentration, cumene concentration, methanol concentration, propylene oxide concentration, formic acid concentration, terephthalic acid concentration, isophthalic acid concentration, acetic acid concentration, influent flow rate, sludge retention time, sludge MLSS concentration in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sludge settling ratio, sludge volume index, effluent BOD concentration, ATP concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP).

[0309] Step 1.2: Preprocess the water quality parameters of the chemical wastewater, remove abnormal data, and then normalize the preprocessed water quality parameters to eliminate the influence of dimensional and order-of-magnitude differences on the water quality parameters. Principal component analysis is then used to analyze the obtained water quality parameters. Based on the correlation coefficients and contribution rates of each water quality parameter to the prediction of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration in the wastewater treatment process, the parameters are screened. The screened water quality parameters and influent flow rate are used as input variables. The output variables are determined by combining the measured ATP and BOD concentration values ​​of the chemical wastewater. A sample of chemical wastewater water quality parameters is then formed using the input and output variables.

[0310] In this embodiment, the input variables include the influent chemical oxygen demand (COD) concentration, the MLSS concentration of the biological treatment tank sludge, polycyclic aromatic hydrocarbons (PAHs), volatile phenols, sulfides, dissolved oxygen (DO) concentration in the biological treatment tank, sludge volume index, temperature, influent ammonia nitrogen concentration, formaldehyde concentration, cumene concentration, propylene oxide concentration, terephthalic acid concentration, and influent flow rate. The output variables include the adenosine triphosphate (ATP) concentration and the biochemical oxygen demand (BOD) concentration.

[0311] Step 1.3: Randomly allocate the water quality parameter samples of each chemical wastewater to the training sample set and the test sample set to generate the training sample set and the test sample set, and construct the sample dataset for training the ATP and BOD prediction model of the wastewater treatment process.

[0312] In this embodiment, a total of 350 sets of chemical wastewater quality parameter samples were obtained, of which the training sample set included 200 sets of chemical wastewater quality parameter samples and the test sample set included 150 sets of test samples.

[0313] Step 2: Establish ATP and BOD prediction models for the wastewater treatment process based on a multi-input multi-output fuzzy neural network.

[0314] In this embodiment, the multi-input multi-output fuzzy neural network includes an input layer, an RBF layer, a rule layer, and an output layer, such as... Figure 1 As shown.

[0315] The input layer of the multi-input multi-output fuzzy neural network is configured as follows:

[0316] x i =u i (1)

[0317] x = [x1, x2, ..., x k (2)

[0318] In the formula, i is the index of the input variable, i = 1, 2, ..., k, and k is the total number of input variables. In this embodiment, k = 14; x iLet u be the value of the i-th input variable. i Let x be the input value of the i-th input neuron, and let x be the input vector.

[0319] The RBF layer contains multiple neurons, and the output function value of each neuron in the RBF layer is:

[0320]

[0321] In the formula, φ j c is the output value of the j-th neuron in the RBF layer, where j is the neuron number (j = 1, 2, ..., Q) and Q is the total number of neurons in the RBF layer (Q = 8 in this embodiment). ij σ is the center of the j-th neuron. ij Let be the width of the j-th neuron.

[0322] The number of neurons in the rule layer and the RBF layer are equal, and the output function value of the neurons in the rule layer is:

[0323]

[0324] In the formula, l is the inner layer number of the rule layer, l = 1, 2, ..., Q; v l φ is the output value of the l-th layer in the rule layer. l This is the output value of the l-th layer in the RBF layer.

[0325] The output value of the output layer is:

[0326] W = [w 1 ,w 2 ,…,w M ] T (5)

[0327] y = Wv (6)

[0328] In the formula, W is the weight matrix, w M Let be the output value of the Mth neuron in the output layer, where M is the total number of neurons in the output layer, and T is the transpose matrix; y is the abnormal condition discrimination matrix, y = [y1, y2], where y1 is the predicted value of adenosine triphosphate (ATP) concentration, and y2 is the predicted value of biochemical oxygen demand (BOD) concentration; v is the output matrix of the regular layer, v = [v1, v2, ..., v Q ] T v Q This is the output value of the Q-th layer in the rule layer.

[0329] In this embodiment, a multi-input multi-output fuzzy neural network for predicting ATP and BOD in wastewater treatment processes is trained using the Levenberg-Marquardt algorithm based on adaptive learning rate. By combining the second-order algorithm with the adaptive learning rate, the convergence speed of the Levenberg-Marquardt algorithm is effectively improved by utilizing the adaptive learning rate.

[0330] The update rule for the Levenberg-Marquardt algorithm based on adaptive learning rate is set as follows:

[0331] Θ(t+1)=Θ(t)+(Ψ(t)+λ(t)×I) -1 Ω(t) (7)

[0332] In the formula, t is time, Θ(t+1) is the variable vector at time t+1, Θ(t) is the variable vector at time t, Ψ(t) is the quasi-Hessian matrix at time t, λ(t) is the adaptive learning rate at time t, 0<λ(t)<1, I is the identity matrix, and Ω(t) is the gradient vector at time t.

[0333] In the update rule of the Levenberg-Marquardt algorithm based on adaptive learning rate, the formula for calculating the adaptive learning rate λ(t) is as follows:

[0334] λ(t)=μ(t)λ(t-1) (8)

[0335]

[0336] In the formula, μ(t) is the adaptive factor at time t, and λ(t-1) is the adaptive learning rate at time t-1; τ min (t) is the smallest eigenvalue of the quasi-Hessian matrix at time t, τ max (t) is the largest eigenvalue of the quasi-Hessian matrix at time t, 0 < τ min (t)<τ max (t).

[0337] The variable vector Θ(t) includes the connection weight vector w(t), the center vector c(t), and the width vector σ(t). The formula for calculating the variable vector is as follows:

[0338] Θ(t)=[w1(t),...,w J (t),c1(t),...,c J (t),...,σ1(t),...,σ J (t)] (10)

[0339] In the formula, w J (t) represents the J-th connection weight vector value at time t, cJ (t) represents the J-th center vector value at time t, σ J (t) represents the J-th width vector value at time t.

[0340] The formulas for calculating the quasi-Hessian matrix Ψ(t) and the gradient vector Ω(t) are as follows:

[0341] Ψ(t)=j T (t)j(t) (10)

[0342] Ω(t)=j T (t)e(t) (11)

[0343] e(t) = y d (t)-y(t) (12)

[0344] in,

[0345] The Jacobian vector j(t) is:

[0346]

[0347] In the formula, j(t) is the Jacobian vector at time t, e(t) is the error value of the multi-input multi-output fuzzy neural network at time t, and y d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

[0348] Step 3: Train the ATP and BOD prediction models for the wastewater treatment process using the training set to obtain the trained ATP and BOD prediction models for the wastewater treatment process. This specifically includes the following steps:

[0349] Step 3.1: Set the training precision value.

[0350] Step 3.2: Input the water quality parameter samples of chemical wastewater from the training set into the ATP and BOD prediction model of the wastewater treatment process constructed in Step 2. Use the ATP and BOD prediction model of the wastewater treatment process to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration of the chemical wastewater.

[0351] Step 3.3: Compare the predicted values ​​of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration from the ATP and BOD prediction model for the wastewater treatment process with the ATP and BOD concentration values ​​from the water quality parameter samples of chemical wastewater. Calculate the accuracy value of the ATP and BOD prediction model for the wastewater treatment process. If the accuracy value of the ATP and BOD prediction model for the wastewater treatment process is less than the preset accuracy value, proceed to step 3.4; otherwise, proceed to step 3.5.

[0352] Step 3.4: Update the adaptive learning rate of the ATP and BOD prediction model for the wastewater treatment process based on the Levenberg-Marquardt algorithm to obtain the updated ATP and BOD prediction model. Then, randomly select water quality parameter samples of chemical wastewater from the training set and input them into the updated ATP and BOD prediction model. Use the updated ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) of the chemical wastewater samples. After obtaining the predicted values ​​of ATP and BOD concentrations of the chemical wastewater samples, return to step 3.3.

[0353] Step 3.5: Complete the training of the ATP and BOD prediction models for the wastewater treatment process, and obtain the trained ATP and BOD prediction models for the wastewater treatment process.

[0354] Step 4: Validate the prediction performance of the trained ATP and BOD prediction models for wastewater treatment processes using the test set. If the prediction performance of the trained ATP and BOD prediction models for wastewater treatment processes has reached its optimal level, proceed to Step 5. Otherwise, return to Step 3 and continue training the ATP and BOD prediction models for wastewater treatment processes using the training set. This specifically includes the following steps:

[0355] Step 4.1: Input the water quality parameter samples of chemical wastewater from the test set into the trained wastewater treatment process ATP and BOD prediction model. Use the wastewater treatment process ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration in the chemical wastewater.

[0356] Step 4.2: Compare the predicted values ​​of ATP and BOD concentrations from the wastewater treatment process prediction model with the ATP and BOD concentrations in the chemical wastewater quality parameter samples. Calculate the accuracy of the wastewater treatment process prediction model. If the accuracy is not less than the preset accuracy, proceed to step 5; otherwise, return to step 3 and continue training the wastewater treatment process prediction model using the training set.

[0357] In this embodiment, the error values ​​of the ATP and BOD prediction models for the wastewater treatment process are:

[0358]

[0359] The accuracy values ​​of the ATP and BOD prediction models for wastewater treatment processes are:

[0360]

[0361] in,

[0362] e(t) = y d (t)-y(t) (16)

[0363] In the formula, RMSE(t) represents the error value of the ATP and BOD prediction models for the wastewater treatment process, t represents time (t = 1, 2, ..., N), N represents the total number of times, e(t) represents the error value of the multi-input multi-output fuzzy neural network at time t, and y represents the error value of the model. d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

[0364] Step 5: Using the ATP and BOD prediction model of the wastewater treatment process after testing, predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) in real time based on the water quality parameters of the chemical wastewater, and generate an abnormal operating condition discrimination matrix, such as... Figure 2 As shown, the abnormal operating condition discrimination matrix is ​​used to identify the type of abnormal operating condition, and the sludge condition and biochemical treatment potential of the chemical wastewater treatment plant are evaluated based on the concentration values ​​of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD), so as to adjust the load operation plan of the chemical wastewater treatment plant.

[0365] The abnormal operating condition discrimination matrix is ​​as follows:

[0366]

[0367] In the formula, ATP is the concentration of adenosine triphosphate (ATP) predicted by the ATP and BOD prediction model for the wastewater treatment process, and BOD is the concentration of biochemical oxygen demand (BOD) predicted by the ATP and BOD prediction model for the wastewater treatment process; C1 is the first type of abnormal operating condition, C2 is the second type of abnormal operating condition, C3 is the third type of abnormal operating condition, and C4 is the fourth type of abnormal operating condition.

[0368] In this embodiment, when the ATP concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is greater than 1.1, it is characterized as high ATP; when the ATP concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is less than 1.1, it is characterized as low ATP. Similarly, when the BOD concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is greater than 1100, it is characterized as high BOD; when the BOD concentration predicted by the ATP and BOD prediction model for the wastewater treatment process is less than 1100, it is characterized as low BOD.

[0369] When determining the type of abnormal operating condition based on the abnormal operating condition discrimination matrix, according to the concentration values ​​of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD):

[0370] When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are high, the current operating condition of the chemical wastewater treatment plant is determined to be the first type of abnormal operating condition, C1. The influent concentration of the chemical wastewater treatment plant exceeds its treatment capacity, the BOD concentration is abnormal, but the sludge activity is relatively high, and the biochemical treatment capacity is high.

[0371] When adenosine triphosphate (ATP) is high and biochemical oxygen demand (BOD) is low, the current operating condition of the chemical wastewater treatment plant is determined to be the second type of abnormal operating condition, C2. The operating condition is good, the biochemical treatment capacity of the chemical wastewater treatment plant is sufficient, and the aeration rate can be appropriately reduced or the treatment load increased.

[0372] When adenosine triphosphate (ATP) is low and biochemical oxygen demand (BOD) is high, the current operating condition of the chemical wastewater treatment plant is determined to be the third type of abnormal operating condition, C3. This indicates that the sludge in the chemical wastewater treatment plant is being impacted by water flow. It is necessary to measure the ATP concentration in the sludge and analyze the abnormal operating condition in conjunction with the ATP concentration value of the sludge to determine the possible abnormal state of the sludge.

[0373] When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are low, the current operating condition of the chemical wastewater treatment plant is determined to be the fourth abnormal operating condition type C4, indicating low biochemical treatment potential and low biochemical treatment capacity. The sludge treatment capacity of the chemical wastewater treatment plant can be improved by reducing the wastewater treatment volume, increasing the aeration volume, or supplementing carbon sources.

[0374] Therefore, the biochemical potential assessment method for abnormal operating conditions in the chemical wastewater treatment process in this embodiment establishes an abnormal operating condition discrimination matrix by using the real-time predicted values ​​of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration. This enables accurate identification of abnormal operating conditions such as sludge poisoning and carbon source depletion. By automatically pushing process adjustment plans and emergency remedial measures, it effectively ensures that the effluent quality of chemical wastewater treatment meets the standards under abnormal operating conditions.

[0375] In the description of this invention, it should be noted that, unless otherwise expressly specified and limited, the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0376] In this invention, terms such as "upper," "lower," "bottom," and "top" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are merely relational terms determined for the convenience of describing the structural relationship of the various components or elements of this invention, and do not specifically refer to any component or element in this invention, and should not be construed as limiting this invention.

[0377] In this invention, terms such as "connected" and "linked" should be interpreted broadly, indicating a fixed connection, an integral connection, or a detachable connection; a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can determine the specific meaning of these terms in this invention based on the specific circumstances, and they should not be construed as limitations on the invention.

[0378] Of course, the above description is not intended to limit the present invention, and the present invention is not limited to the examples given above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A method for assessing the biochemical potential under abnormal operating conditions in a chemical wastewater treatment process, characterized in that, Assessing the biochemical treatment potential of abnormal chemical wastewater conditions by real-time prediction of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration includes the following steps: Step 1: Obtain water quality parameters from the chemical wastewater treatment plant. After preprocessing, normalizing, and principal component analysis of the water quality parameters, multiple chemical wastewater water quality parameter samples are obtained. Each chemical wastewater water quality parameter sample is randomly assigned to the training sample set and the test sample set. The training sample set and the test sample set together constitute the sample dataset used to train the ATP and BOD prediction models for the wastewater treatment process. Step 2: Establish ATP and BOD prediction models for the wastewater treatment process based on a multi-input multi-output fuzzy neural network; Step 3: Use the training set to train the ATP and BOD prediction models for the wastewater treatment process to obtain the trained ATP and BOD prediction models for the wastewater treatment process. Step 4: Use the test set to verify the prediction effect of the ATP and BOD prediction model for the wastewater treatment process after training. If the prediction effect of the ATP and BOD prediction model for the wastewater treatment process after training has reached the best, proceed to step 5. Otherwise, return to step 3 and continue to train the ATP and BOD prediction model for the wastewater treatment process using the training set. Step 5: Using the ATP and BOD prediction model of the wastewater treatment process after testing, predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) in real time based on the water quality parameters of the chemical wastewater, generate an abnormal operating condition discrimination matrix, use the abnormal operating condition discrimination matrix to identify the type of abnormal operating condition, and evaluate the sludge status and biochemical treatment potential of the chemical wastewater treatment plant based on the ATP and BOD concentration values, and adjust the load operation plan of the chemical wastewater treatment plant.

2. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 1, characterized in that, Step 1 specifically includes the following steps: Step 1.1: Obtain the chemical wastewater from the chemical wastewater treatment plant and measure the influent flow rate, water quality parameters, adenosine triphosphate (ATP), and biochemical oxygen demand (BOD). Step 1.2: Preprocess the water quality parameters of the chemical wastewater, remove abnormal data from each water quality parameter, normalize the preprocessed water quality parameters, and analyze the obtained water quality parameters using principal component analysis. Based on the correlation coefficient and contribution rate of each water quality parameter to the prediction of ATP and BOD in the wastewater treatment process, screen the water quality parameters and influent flow rate as input variables, and determine the output variables by combining the measured ATP concentration and BOD concentration of the chemical wastewater. Use the input and output variables to form a sample of water quality parameters for the chemical wastewater. Step 1.3: Randomly allocate the water quality parameter samples of each chemical wastewater to the training sample set and the test sample set to generate the training sample set and the test sample set, and construct the sample dataset for training the ATP and BOD prediction model of the wastewater treatment process.

3. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 2, characterized in that, When the chemical wastewater is wastewater from a propylene oxide plant and ethylene alkali residue wastewater, the water quality parameters include: influent chemical oxygen demand (COD) concentration, effluent COD concentration, formaldehyde concentration, cumene concentration, methanol concentration, propylene oxide concentration, propylene concentration, formic acid concentration, influent flow rate, sludge retention time, sludge MLSS concentration in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sulfide concentration in the biological treatment tank, volatile phenol concentration in the biological treatment tank, sludge settling ratio, sludge volume index, effluent biochemical oxygen demand (BOD) concentration, adenosine triphosphate (ATP) concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP). When the chemical wastewater is purified terephthalic acid wastewater and ethylene alkali residue wastewater, the water quality parameters include influent chemical oxygen demand (COD) concentration, effluent COD concentration, formaldehyde concentration, terephthalic acid concentration, isophthalic acid concentration, acetic acid concentration, influent flow rate, sludge retention time, sludge MLSS concentration in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sludge settling ratio, sludge volume index, effluent biochemical oxygen demand (BOD) concentration, adenosine triphosphate (ATP) concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP). When the chemical wastewater is propylene oxide wastewater, purified terephthalic acid wastewater, and ethylene alkali residue wastewater, the water quality parameters include influent chemical oxygen demand (COD) concentration, effluent COD concentration, formaldehyde concentration, cumene concentration, methanol concentration, propylene oxide concentration, propylene concentration, formic acid concentration, terephthalic acid concentration, isophthalic acid concentration, acetic acid concentration, influent flow rate, sludge retention time, sludge MLSS concentration in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sludge settling ratio, sludge volume index, effluent biochemical oxygen demand (BOD) concentration, adenosine triphosphate (ATP) concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP). When the chemical wastewater is propylene oxide wastewater, aromatic hydrocarbon wastewater, purified terephthalic acid wastewater, and ethylene alkali residue wastewater, the water quality parameters include influent chemical oxygen demand (COD) concentration, effluent COD concentration, cyanide concentration, polycyclic aromatic hydrocarbons (PAHs), volatile phenol concentration, sulfide concentration, formaldehyde concentration, cumene concentration, methanol concentration, propylene oxide concentration, formic acid concentration, terephthalic acid concentration, isophthalic acid concentration, acetic acid concentration, influent flow rate, sludge retention time, MLSS concentration of sludge in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, influent ammonia nitrogen concentration, effluent ammonia nitrogen concentration, sludge settling ratio, sludge volume index, effluent biochemical oxygen demand (BOD) concentration, adenosine triphosphate (ATP) concentration in the biological treatment tank, temperature, pH, and oxidation-reduction potential (ORP).

4. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 3, characterized in that, When the chemical wastewater is propylene oxide plant wastewater and ethylene alkali residue wastewater, the input variables include influent chemical oxygen demand (COD) concentration, MLSS concentration of sludge in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, sludge volume index, temperature, influent ammonia nitrogen concentration, propylene oxide concentration, formaldehyde concentration, and influent flow rate; the output variables include adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration. When the chemical wastewater is purified terephthalic acid wastewater and ethylene alkali residue wastewater, the input variables include influent chemical oxygen demand (COD) concentration, MLSS concentration of sludge in the biological treatment tank, dissolved oxygen (DO) concentration in the biological treatment tank, sludge volume index, temperature, influent ammonia nitrogen concentration, formaldehyde concentration, terephthalic acid concentration, influent flow rate, and isophthalic acid concentration. The output variables include adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration. When the chemical wastewater is propylene oxide wastewater, purified terephthalic acid wastewater, and ethylene alkali residue wastewater, the input variables include the influent chemical oxygen demand (COD) concentration, the sludge MLSS concentration in the biological treatment tank, the dissolved oxygen (DO) concentration in the biological treatment tank, the sludge volume index, the temperature, the influent ammonia nitrogen concentration, the formaldehyde concentration, the cumene concentration, the methanol concentration, the propylene oxide concentration, the terephthalic acid concentration, and the influent flow rate; the output variables include the adenosine triphosphate (ATP) concentration and the biochemical oxygen demand (BOD) concentration. When the chemical wastewater is propylene oxide wastewater, aromatic hydrocarbon wastewater, purified terephthalic acid wastewater, and ethylene alkali residue wastewater, the input variables include the influent chemical oxygen demand (COD) concentration, the MLSS concentration of the sludge in the biological treatment tank, the concentration of polycyclic aromatic hydrocarbons, volatile phenols, sulfides, the dissolved oxygen (DO) concentration in the biological treatment tank, the sludge volume index, temperature, the influent ammonia nitrogen concentration, formaldehyde concentration, cumene concentration, propylene oxide concentration, terephthalic acid concentration, and influent flow rate. The output variables include the adenosine triphosphate (ATP) concentration and the biochemical oxygen demand (BOD) concentration.

5. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 2, characterized in that, In step 2, the multi-input multi-output fuzzy neural network includes an input layer, an RBF layer, a rule layer, and an output layer; The input layer of the multi-input multi-output fuzzy neural network is configured as follows: x i =u i (1) x=[x1,x2,...,x k ] (2) In the formula, i is the index of the input variable, i = 1, 2, ..., k, and k is the total number of input variables; x i Let u be the value of the i-th input variable. i Let x be the input value of the i-th input neuron, and let x be the input vector. The RBF layer contains multiple neurons, and the output function value of each neuron in the RBF layer is: In the formula, φ j c is the output value of the j-th neuron in the RBF layer, where j is the neuron number (j = 1, 2, ..., Q) and Q is the total number of neurons in the RBF layer; ij σ is the center of the j-th neuron. ij The width of the j-th neuron; The rule layer contains multiple neurons, and the output function value of each neuron in the rule layer is: In the formula, l is the inner layer number of the rule layer, l = 1, 2, ..., Q; v l φ is the output value of the l-th layer in the rule layer. l This is the output value of the l-th layer in the RBF layer; The output value of the output layer is: In=[in 1 ,In 2 ,…,In M ] T (5) y = Wv (6) In the formula, W is the weight matrix, w M Let be the output value of the Mth neuron in the output layer, where M is the total number of neurons in the output layer, and T is the transpose matrix; y is the abnormal condition discrimination matrix, y = [y1, y2], where y1 is the predicted value of adenosine triphosphate (ATP) concentration, and y2 is the predicted value of biochemical oxygen demand (BOD) concentration; v is the output matrix of the regular layer, v = [v1, v2, ..., v Q ] T v Q This is the output value of the Q-th layer in the rule layer.

6. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 5, characterized in that, The number of neurons in the rule layer is equal to that in the RBF layer.

7. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 6, characterized in that, In step 2, a multi-input multi-output fuzzy neural network for predicting ATP and BOD in the wastewater treatment process is trained and constructed using the Levenberg-Marquardt algorithm based on adaptive learning rate. The update rule for the Levenberg-Marquardt algorithm based on adaptive learning rate is set as follows: Θ(t+1)=Θ(t)+(Ψ(t)+λ(t)×I) -1 Ω(t) (7) In the formula, t is time, Θ(t+1) is the variable vector at time t+1, Θ(t) is the variable vector at time t, Ψ(t) is the quasi-Hessian matrix at time t, λ(t) is the adaptive learning rate at time t, 0<λ(t)<1, I is the identity matrix, and Ω(t) is the gradient vector at time t. In the update rule of the Levenberg-Marquardt algorithm based on adaptive learning rate, the formula for calculating the adaptive learning rate λ(t) is as follows: λ(t)=μ(t)λ(t-1) (8) In the formula, μ(t) is the adaptive factor at time t, and λ(t-1) is the adaptive learning rate at time t-1; τ min (t) is the smallest eigenvalue of the quasi-Hessian matrix at time t, τ max (t) is the largest eigenvalue of the quasi-Hessian matrix at time t, 0 < τ min (t)<τ max (t); The variable vector Θ(t) includes the connection weight vector w(t), the center vector c(t), and the width vector σ(t). The formula for calculating the variable vector is as follows: Θ(t)=[w1(t),...,w J (t),c1(t),...,c J (t),...,σ1(t),...,σ J (t)] (10) In the formula, w J (t) represents the J-th connection weight vector value at time t, c J (t) represents the J-th center vector value at time t, σ J (t) represents the J-th width vector value at time t; The formulas for calculating the quasi-Hessian matrix Ψ(t) and the gradient vector Ω(t) are as follows: Ψ(t)=j T (t)j(t) (10) Ω(t)=j T (t)e(t) (11) e(t)=y d (t)-y(t) (12) in, The Jacobian vector j(t) is: In the formula, j(t) is the Jacobian vector at time t, e(t) is the error value of the multi-input multi-output fuzzy neural network at time t, and y d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

8. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 1, characterized in that, Step 3 specifically includes the following steps: Step 3.1, set the training precision value; Step 3.2: Input the water quality parameter samples of chemical wastewater in the training set into the ATP and BOD prediction model of the wastewater treatment process constructed in Step 2. Use the ATP and BOD prediction model of the wastewater treatment process to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration of the chemical wastewater. Step 3.3: Compare the predicted values ​​of adenosine triphosphate (ATP) concentration and biochemical oxygen demand (BOD) concentration from the ATP and BOD prediction model for the wastewater treatment process with the ATP and BOD concentration values ​​from the water quality parameter samples of chemical wastewater. Calculate the accuracy value of the ATP and BOD prediction model for the wastewater treatment process. If the accuracy value of the ATP and BOD prediction model for the wastewater treatment process is less than the preset accuracy value, proceed to step 3.4; otherwise, proceed to step 3.

5. Step 3.4: Update the adaptive learning rate of the ATP and BOD prediction model for the wastewater treatment process based on the Levenberg-Marquardt algorithm to obtain the updated ATP and BOD prediction model. Then, randomly select water quality parameter samples of chemical wastewater from the training set and input them into the updated ATP and BOD prediction model. Use the updated ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) of the chemical wastewater samples. After obtaining the predicted values ​​of ATP and BOD concentrations of the chemical wastewater samples, return to step 3.

3. Step 3.5: Complete the training of the ATP and BOD prediction models for the wastewater treatment process, and obtain the trained ATP and BOD prediction models for the wastewater treatment process.

9. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 1, characterized in that, Step 4 specifically includes the following steps: Step 4.1: Input the water quality parameter samples of chemical wastewater from the test set into the trained wastewater treatment process ATP and BOD prediction model. Use the wastewater treatment process ATP and BOD prediction model to predict the concentration of adenosine triphosphate (ATP) and the concentration of biochemical oxygen demand (BOD) in the chemical wastewater, and obtain the predicted values ​​of ATP concentration and BOD concentration in the chemical wastewater. Step 4.2: Compare the predicted values ​​of ATP and BOD concentrations from the wastewater treatment process ATP and BOD prediction model with the ATP and BOD concentrations in the chemical wastewater quality parameter samples. Calculate the accuracy of the wastewater treatment process ATP and BOD prediction model. If the accuracy is not less than the preset accuracy, proceed to step 5; otherwise, return to step 3 and continue training the wastewater treatment process ATP and BOD prediction model using the training set.

10. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 8 or 9, characterized in that, The error values ​​of the ATP and BOD prediction models for the wastewater treatment process are: The accuracy values ​​of the ATP and BOD prediction models for wastewater treatment processes are: in, e(t)=y d (t)-y(t) (16) In the formula, RMSE(t) represents the error value of the ATP and BOD prediction models for the wastewater treatment process, t represents time (t = 1, 2, ..., N), N represents the total number of times, e(t) represents the error value of the multi-input multi-output fuzzy neural network at time t, and y represents the error value of the model. d y(t) represents the predicted value of the multi-input multi-output fuzzy neural network at time t, and y(t) represents the output variable value of the chemical wastewater quality parameter sample at time t.

11. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 1, characterized in that, In step 5, the abnormal operating condition discrimination matrix is: In the formula, ATP is the concentration of adenosine triphosphate (ATP) predicted by the ATP and BOD prediction model for the wastewater treatment process, and BOD is the concentration of biochemical oxygen demand (BOD) predicted by the ATP and BOD prediction model for the wastewater treatment process; C1 is the first type of abnormal operating condition, C2 is the second type of abnormal operating condition, C3 is the third type of abnormal operating condition, and C4 is the fourth type of abnormal operating condition.

12. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 11, characterized in that, In step 5, when determining the type of abnormal operating condition based on the abnormal operating condition discrimination matrix and according to the concentration values ​​of adenosine triphosphate (ATP) and biochemical oxygen demand (BOD): When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are high, the current operating condition of the chemical wastewater treatment plant is determined to be the first type of abnormal operating condition, C1. This means that the influent concentration of the chemical wastewater treatment plant exceeds its treatment capacity, the BOD concentration is abnormal, and the biochemical treatment capacity is high. When adenosine triphosphate (ATP) is high and biochemical oxygen demand (BOD) is low, the current operating condition of the chemical wastewater treatment plant is determined to be the second type of abnormal operating condition, C2, indicating that the biochemical treatment capacity of the chemical wastewater treatment plant is excessive. When adenosine triphosphate (ATP) is low and biochemical oxygen demand (BOD) is high, the current operating condition of the chemical wastewater treatment plant is determined to be the third type of abnormal operating condition, C3. It is determined that the sludge in the current chemical wastewater treatment plant is being impacted by water flow. It is necessary to measure the adenosine triphosphate (ATP) in the sludge and analyze the abnormal operating condition in combination with the concentration value of adenosine triphosphate (ATP) in the sludge. When both adenosine triphosphate (ATP) and biochemical oxygen demand (BOD) are low, the current operating condition of the chemical wastewater treatment plant is determined to be the fourth abnormal operating condition type C4, indicating that the biochemical treatment capacity of the chemical wastewater treatment plant is low.

13. The method for assessing the biochemical potential of abnormal operating conditions in a chemical wastewater treatment process according to claim 11, characterized in that, When the operating condition of a chemical wastewater treatment plant is determined to be the fourth abnormal operating condition type C4, the sludge disposal capacity of the chemical wastewater treatment plant can be improved by reducing the wastewater treatment volume, increasing the aeration volume, or supplementing the carbon source.

Images