Parameter adjustment method, device and system and storage medium

Abstract

The invention discloses a parameter adjustment method, device and system and a storage medium, which are used for improving the parameter adjustment precision and reducing the cost. The method comprises the following steps: collecting historical operation data of the desulfurization wastewater treatment system; according to the historical operation data, boundary conditions of the next moment are predicted, and the boundary conditions comprise the boundary condition of desulfurization wastewater turbidity, the boundary condition of desulfurization wastewater inflow and the boundary condition of unit load; screening a target working condition from an operation target working condition library of the desulfurization wastewater treatment system according to the boundary condition of the next moment; obtaining an optimal working condition which enables the operation cost to be the lowest from the target working conditions; and guiding the parameter adjustment of the wastewater treatment system according to the optimal working condition. According to the scheme, by predicting the boundary condition of the next moment, the target working condition of the next moment is accurately determined, the precision of parameter adjustment is improved, and then the optimal working condition with the lowest operation cost is screened from the target working conditions, so that the operation cost is reduced.

Description

Technical Field

[0001] This application relates to the field of wastewater treatment technology, and in particular to a parameter adjustment method, apparatus, system and storage medium. Background Technology

[0002] Currently, coal-fired power plants mainly use limestone-gypsum wet flue gas desulfurization technology, which generates acidic, complex, and difficult-to-treat desulfurization wastewater during operation. Existing technologies often rely on experience, pre-determined adjustment cycles, or the experience of staff to add chemicals to adjust the parameters of the wastewater treatment system. If the chemicals are not added in a timely manner, the adjustment effect cannot be achieved; if the chemicals are added for filtration, it will affect the wastewater treatment effect without increasing operating costs.

[0003] Therefore, how to provide a parameter adjustment method to improve the accuracy of parameter adjustment and reduce costs has become an urgent technical problem to be solved. Summary of the Invention

[0004] This application provides a parameter adjustment method, apparatus, system, and storage medium to improve parameter adjustment accuracy and reduce costs.

[0005] This application provides a parameter adjustment method, including:

[0006] Collect historical operating data of the desulfurization wastewater treatment system;

[0007] The boundary conditions for the next moment are predicted based on the historical operating data, wherein the boundary conditions include the boundary conditions of desulfurization wastewater turbidity, the boundary conditions of desulfurization wastewater influent flow rate, and the boundary conditions of unit load.

[0008] Based on the boundary conditions of the next moment, target operating conditions are selected from the target operating condition library of the desulfurization wastewater treatment system.

[0009] Obtain the optimal operating condition that minimizes operating costs from the target operating conditions;

[0010] The optimal operating conditions guide the adjustment of wastewater treatment system parameters.

[0011] The beneficial effects of this application are as follows: It collects historical operating data of the desulfurization wastewater treatment system; predicts the boundary conditions for the next moment based on the historical operating data, wherein the boundary conditions include boundary conditions for desulfurization wastewater turbidity, desulfurization wastewater influent flow rate, and unit load; selects target operating conditions from the target operating condition database of the desulfurization wastewater treatment system based on the boundary conditions for the next moment; obtains the optimal operating condition that minimizes operating costs from the target operating conditions; and guides the adjustment of wastewater treatment system parameters based on this optimal operating condition. Because the boundary conditions for the next moment are predicted based on historical operating data, and target operating conditions that meet the conditions are selected from the target operating condition database based on the boundary conditions, the target operating conditions for the next moment are accurately determined, improving the accuracy of parameter adjustment. Furthermore, by selecting the optimal operating condition that minimizes operating costs from the target operating conditions and using the optimal operating condition to guide parameter adjustment, the effect of cost reduction is achieved.

[0012] In one embodiment, predicting the boundary conditions for the next moment based on the historical operating data includes:

[0013] Discretize the historical operation data;

[0014] The discretized historical running data is input into the VMD-CNN-LSTM model to predict the boundary conditions at the next time step.

[0015] In one embodiment, the process of establishing the target operating condition database for the desulfurization wastewater treatment system is as follows:

[0016] The FCM algorithm is used to cluster the sample data to obtain the membership matrix and cluster centers of the samples;

[0017] The sample data is classified according to the membership matrix. Taking the operating cost of the desulfurization wastewater treatment system as the decision objective, the class with the lowest relative cost at the class center is found.

[0018] Calculate the distance from all sample points in the class to the class center, and return the closest sample as the operating baseline value under the current working conditions;

[0019] Based on the aforementioned operating baseline values, the sampling sliding window method is used to determine whether the data within the sliding window represents a stable operating condition, thereby obtaining a database of target operating conditions for the desulfurization wastewater treatment system.

[0020] In one embodiment, the step of clustering sample data using the FCM algorithm to obtain the membership matrix and cluster centers of the samples includes:

[0021] Calculate the membership degree of each sample point using the following membership degree calculation formula:

[0022]

[0023] Among them, u ij Let d be the membership degree of the j-th sample point belonging to cluster i. ij is the Euclidean distance from the j-th sample to the i-th cluster center, c is the number of clusters, and m is the fuzzification coefficient;

[0024] Cluster centers are determined using the following formula:

[0025]

[0026] Among them, V i u is the weighted average sample of cluster center i. ij Let m be the membership degree of the j-th sample point belonging to cluster i, m be the fuzzification coefficient, and n be the total number of samples.

[0027] In one embodiment, after obtaining the membership matrix and cluster centers of the samples, the method further includes:

[0028] Calculate the clustering effectiveness evaluation function values ​​corresponding to different numbers of clusters;

[0029] The optimal number of fuzzy clusters is determined based on the clustering effectiveness evaluation function value corresponding to different numbers of clusters.

[0030] In one embodiment, calculating the clustering effectiveness evaluation function value corresponding to different cluster numbers includes:

[0031] The clustering effectiveness evaluation function value is calculated using the following formula:

[0032]

[0033] Where XB(U,V) is the clustering effectiveness evaluation function value, u ij Let d(x) be the membership degree of the j-th sample point belonging to cluster i. j v i ) is the sample x j With cluster center v i Euclidean distance, d(v) i v j ) is the cluster center v i To the cluster center v j The Euclidean distance, where c is the number of clusters, m is the fuzziness coefficient, and n is the total number of samples;

[0034] The step of determining the optimal fuzzy cluster number based on the clustering effectiveness evaluation function values ​​corresponding to different cluster numbers includes:

[0035] The cluster number with the smallest clustering effectiveness evaluation function value is selected as the optimal fuzzy cluster number.

[0036] In one embodiment, obtaining the optimal operating condition that minimizes operating costs from the target operating conditions includes:

[0037] Define the power consumption cost function, define the sewage discharge cost function, and determine the cost of chemical dosage;

[0038] The total operating cost function is determined based on the sewage discharge cost function, the electricity consumption cost function, and the chemical dosage cost.

[0039] The operating cost parameters corresponding to the target operating condition are substituted into the total operating cost function to determine the optimal operating condition that minimizes the operating cost.

[0040] In one embodiment, defining the power consumption cost function includes:

[0041] The electricity cost function is defined as follows:

[0042] W E =W t +W m +W s +W d ;

[0043]

[0044] Among them, W E The total power consumption (kWh) of the desulfurization wastewater treatment system, W t For the energy consumption of the pump system (kWh), W m For the energy consumption of the hybrid system (kWh), W s Energy consumption (kWh) for sludge treatment system, W d For the energy consumption (kWh) of the dosing system, C dh The power consumption cost (yuan / kWh) of the desulfurization wastewater treatment system, P dj The on-grid electricity price (yuan / kWh) for the region where the power plant is located, and h represents the power generation (kWh).

[0045] In one embodiment, defining the sewage discharge cost function includes:

[0046] The sewage discharge cost function is defined as follows:

[0047]

[0048] Where λ is the environmental protection tax reduction coefficient, Qf is the discharge volume of Class I water pollutants (L), and P i Qs represents the equivalent value of Class I water pollutants (t), Qs represents the discharge amount of Class II water pollutants (L), and P represents the discharge amount of Class II water pollutants (L). j Let τ be the equivalent value of Class II water pollutants (t), τ be the regional environmental tax rate (yuan / t), and c be the equivalent value of Class II water pollutants (t). dsFor emission standard limits (mg / L), c v The concentration of suspended solids in desulfurization wastewater is the hourly average (mg / L).

[0049] This application also provides a parameter adjustment device, including:

[0050] The data acquisition module is used to collect historical operating data of the desulfurization wastewater treatment system;

[0051] The prediction module is used to predict the boundary conditions for the next moment based on the historical operating data, wherein the boundary conditions include the boundary conditions of the turbidity of the desulfurization wastewater, the boundary conditions of the influent flow rate of the desulfurization wastewater, and the boundary conditions of the unit load.

[0052] The filtering module is used to filter target operating conditions from the target operating condition library of the desulfurization wastewater treatment system according to the boundary conditions at the next moment.

[0053] The acquisition module is used to acquire the optimal operating condition that minimizes the operating cost from the target operating condition;

[0054] The adjustment module is used to guide the adjustment of wastewater treatment system parameters based on the optimal operating conditions.

[0055] In one embodiment, the prediction module includes:

[0056] The discrete submodule is used to discretize the historical running data;

[0057] The prediction submodule is used to input the discretized historical running data into the VMD-CNN-LSTM model so as to predict the boundary conditions at the next time step.

[0058] In one embodiment, the establishment module is used to establish a database of target operating conditions for the desulfurization wastewater treatment system, and the establishment module includes:

[0059] The clustering submodule is used to cluster sample data using the FCM algorithm to obtain the membership matrix and cluster centers of the samples.

[0060] The classification submodule is used to classify sample data according to the membership matrix, and with the operating cost of the desulfurization wastewater treatment system as the decision objective, finds the class with the lowest relative cost at the class center.

[0061] The first calculation submodule is used to calculate the distance from all sample points in the class to the class center and return the nearest sample as the operating baseline value under the current working conditions.

[0062] The determination submodule is used to determine whether the data in the sliding window is a stable operating condition based on the operating benchmark value using the sampling sliding window method, so as to obtain the operating target condition library of the desulfurization wastewater treatment system.

[0063] In one embodiment, the clustering submodule is further configured to:

[0064] Calculate the membership degree of each sample point using the following membership degree calculation formula:

[0065]

[0066] Among them, u ij Let d be the membership degree of the j-th sample point belonging to cluster i. ij is the Euclidean distance from the j-th sample to the i-th cluster center, c is the number of clusters, and m is the fuzzification coefficient;

[0067] Cluster centers are determined using the following formula:

[0068]

[0069] Among them, V i u is the weighted average sample of cluster center i. ij Let m be the membership degree of the j-th sample point belonging to cluster i, m be the fuzzification coefficient, and n be the total number of samples.

[0070] In one embodiment, the establishment module includes, further comprising:

[0071] The second calculation submodule is used to calculate the clustering effectiveness evaluation function value corresponding to different numbers of clusters;

[0072] The first determination submodule is used to determine the optimal number of fuzzy clusters based on the clustering effectiveness evaluation function value corresponding to different numbers of clusters.

[0073] In one embodiment, the second computing submodule is further configured to:

[0074] The clustering effectiveness evaluation function value is calculated using the following formula:

[0075]

[0076] Where XB(U,V) is the clustering effectiveness evaluation function value, u ij Let d(x) be the membership degree of the j-th sample point belonging to cluster i. j v i ) is the sample x j With cluster center v i Euclidean distance, d(v) i v j ) is the cluster center v i To the cluster center v j The Euclidean distance, where c is the number of clusters, m is the fuzziness coefficient, and n is the total number of samples;

[0077] The step of determining the optimal fuzzy cluster number based on the clustering effectiveness evaluation function values ​​corresponding to different cluster numbers includes:

[0078] The cluster number with the smallest clustering effectiveness evaluation function value is selected as the optimal fuzzy cluster number.

[0079] In one embodiment, the acquisition module includes:

[0080] Define a submodule for defining the power consumption cost function, defining the sewage discharge cost function, and determining the chemical dosage cost;

[0081] The second determining submodule is used to determine the total operating cost function based on the sewage discharge cost function, the electricity consumption cost function, and the chemical dosage cost.

[0082] The third determining submodule is used to substitute the operating cost parameters corresponding to the target operating condition into the total operating cost function in order to determine the optimal operating condition that minimizes the operating cost.

[0083] In one embodiment, defining the power consumption cost function includes:

[0084] The electricity cost function is defined as follows:

[0085]

[0086] Among them, W E The total power consumption (kWh) of the desulfurization wastewater treatment system, W t For the energy consumption of the pump system (kWh), W m For the energy consumption of the hybrid system (kWh), W s Energy consumption (kWh) for sludge treatment system, W d For the energy consumption (kWh) of the dosing system, C dh The power consumption cost of the desulfurization wastewater treatment system (yuan / kWh), P dj The on-grid electricity price for the region where the power plant is located (yuan / kWh), and h represents the power generation (kWh).

[0087] In one embodiment, defining the sewage discharge cost function includes:

[0088] The sewage discharge cost function is defined as follows:

[0089]

[0090] Where λ is the environmental protection tax reduction coefficient, Qf is the discharge volume of Class I water pollutants (L), and P i Qs represents the equivalent value of Class I water pollutants (t), Qs represents the discharge amount of Class II water pollutants (L), and P represents the equivalent value of Class I water pollutants (t). jLet τ be the equivalent value of Class II water pollutants (t), τ be the regional environmental tax rate (yuan / t), and c be the equivalent value of Class II water pollutants (t). ds For emission standard limits (mg / L), c v The concentration of suspended solids in desulfurization wastewater is the hourly average (mg / L).

[0091] This application also provides a parameter adjustment system, including:

[0092] At least one processor; and,

[0093] A memory communicatively connected to the at least one processor; wherein,

[0094] The memory stores instructions that can be executed by the at least one processor to implement the parameter adjustment method described in any of the above embodiments.

[0095] This application also provides a computer-readable storage medium, which, when the instructions in the storage medium are executed by a processor corresponding to a parameter adjustment system, enables the parameter adjustment system to implement the parameter adjustment method described in any of the above embodiments.

[0096] Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings.

[0097] The technical solution of this application will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0098] The accompanying drawings are provided to further illustrate the present application and form part of the specification. They are used together with the embodiments of the present application to explain the application and do not constitute a limitation thereof. In the drawings:

[0099] Figure 1 This is a flowchart of a parameter adjustment method in one embodiment of this application;

[0100] Figure 2 This is a flowchart illustrating the prediction of boundary conditions in one embodiment of this application;

[0101] Figure 3 This is a scatter plot of historical running data after clustering in one embodiment of this application;

[0102] Figure 4 This is a comparison chart of the predicted and actual values ​​of suspended solids concentration by different algorithms in one embodiment of this application;

[0103] Figure 5This is a schematic diagram of the structure of a parameter adjustment device according to an embodiment of this application;

[0104] Figure 6 This is a schematic diagram of the hardware structure of a parameter adjustment system according to an embodiment of this application. Detailed Implementation

[0105] The preferred embodiments of this application are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit this application.

[0106] Figure 1 This is a flowchart of a parameter adjustment method in one embodiment of this application, as shown below. Figure 1 As shown, the method can be implemented as follows: S101-S105:

[0107] In step S101, historical operating data of the desulfurization wastewater treatment system is collected;

[0108] In step S102, the boundary conditions for the next moment are predicted based on the historical operating data, wherein the boundary conditions include the boundary conditions of desulfurization wastewater turbidity, the boundary conditions of desulfurization wastewater influent flow rate, and the boundary conditions of unit load.

[0109] In step S103, target operating conditions are selected from the target operating condition library of the desulfurization wastewater treatment system according to the boundary conditions at the next moment.

[0110] In step S104, the optimal operating condition that minimizes operating costs is obtained from the target operating conditions;

[0111] In step S105, the parameters of the wastewater treatment system are adjusted according to the optimal operating conditions.

[0112] In this application, historical operating data of the desulfurization wastewater treatment system was collected. This historical operating data includes key parameters such as desulfurization wastewater turbidity, influent flow rate, three-tank level, unit load, limestone silo level, coagulant level, wastewater cyclone tank level, effluent tank level, sludge level in the clarifier, lime slurry dosage, organic sulfur dosage, flocculant dosage, coagulant dosage, dewatering machine dosage, and pH value. To ensure the integrity and accuracy of the data during collection, the historical data was preprocessed to remove abnormal data (such as data exceeding turbidity emission concentration standards, measured parameters outside the normal range, and abrupt data points) to guarantee the reliability of subsequent analysis.

[0113] Based on the historical operating data, the boundary conditions for the next moment are predicted. The boundary conditions include the boundary conditions of desulfurization wastewater turbidity, the boundary conditions of desulfurization wastewater influent flow rate, and the boundary conditions of unit load. That is, the boundary prediction values ​​of desulfurization wastewater turbidity, desulfurization wastewater influent flow rate, and unit load are obtained.

[0114] In this application, the historical operating data is first discretized to facilitate subsequent model input and processing. The discretization method can be selected according to the data characteristics and analysis requirements, such as equal-width discretization, equal-frequency discretization, etc. Then, the discretized historical operating data is input into the VMD-CNN-LSTM model to determine the predicted values ​​of the boundary conditions through the model. Figure 2 This is a flowchart illustrating the prediction of boundary conditions in one embodiment of this application, as shown below. Figure 2 As shown, the specific prediction process is as follows:

[0115] (1) Perform variational mode decomposition (VMD) on the historical operating data of the boundary conditions to determine the modal components and center frequencies corresponding to the historical operating data.

[0116] VMD (Variational Mode Decomposition) is an adaptive signal processing method. Its core lies in decomposing the original historical running data f(t) into K mode components u with finite bandwidth through a mathematical variational framework. k (t) (i.e., IMF), each component has a unique center frequency ω k The completeness of the decomposition is ensured by constraining the bandwidth of each mode through the objective function.

[0117] In this application, the non-stationary, high-noise original time-series data of historical operation data (such as the turbidity of desulfurization wastewater) is adaptively decomposed into multiple sub-signals (inherent modal components, IMFs) with different frequency characteristics. Then, the Hilbert transform of each modal component is calculated, the bandwidth of each component is estimated using Gaussian smoothing, and the problem of minimizing the sum of modal bandwidth estimates is solved to obtain the corresponding modal component signals and related parameters such as center frequency.

[0118] Taking desulfurization wastewater turbidity data as an example, the parameter values ​​of VMD parameters are obtained, such as the number of modes K (decomposition dimension) and the penalty factor α (bandwidth controller). Then, VMD decomposition is performed on historical operating data (desulfurization wastewater turbidity) based on the parameter values ​​to reduce the volatility of wastewater data.

[0119] The parameter values ​​of VMD (number of modes K and penalty factor α) can be predetermined empirically. In one embodiment of this application, the VMD parameters are automatically determined using a particle swarm optimization algorithm.

[0120] Initialize the number of modes and the penalty factor. For example, set the search space for the number of modes K∈[3,10] and the penalty factor α∈[1000,5000]. Perform VMD decomposition based on the initialized number of modes and penalty factor to obtain the modal components corresponding to the current VMD parameters. Substitute the modal components corresponding to the current VMD parameters into the fitness function to calculate the current fitness value corresponding to the current VMD parameters. For example, the fitness function can be the joint index of modal component kurtosis-entropy.

[0121]

[0122] Among them, Fitness(x) i The value represents the fitness score; a higher value indicates a better decomposition effect. k For the k-th modal component; Kurtosis(u k ) represents the kurtosis of the modal components; Entropy(u k ) represents the Shannon entropy of the modal components;

[0123] μ k σ is the mean of the modal components. k The standard deviation of the modal components;

[0124] Entropy(u k )=-∑p(u k )logp(u k ), p(u k ) is the probability mass function.

[0125] The fitness function described above is used to combine the parameters (K) of each particle. i α i The parameters are quantitatively scored, and the score results directly determine the quality of the parameter combination. A combination with a high kurtosis-entropy joint index indicates that the decomposed modal components have both significant impact characteristics and orderliness, which is more in line with engineering requirements. High-scoring parameters will be reserved first for subsequent iterations.

[0126] The particle dynamically adjusts its movement direction based on a comparison between its current fitness value and its individual best (pbest) and the group's best (gbest).

[0127]

[0128] in, The updated velocity of the particle determines the direction and distance of its next movement; w(t) is the dynamic inertia weight, which controls the influence of historical velocities. For example, a larger value (such as 0.9) enhances global exploration capabilities, while a smaller value (such as 0.4) enhances local exploration capabilities. It can decrease linearly with iteration. c1 is the particle's current velocity, and c1 is the individual learning factor, which controls the intensity of the particle's movement towards its historical best position (pbest). Typically, c1 = 2. id is the best historical position of particle i, the highest fitness position that particle i has experienced in dimension d. The particle's current position, i.e., the VMD parameter combination (K, α) to be optimized; c2 is the social learning factor, which controls the intensity of the particle's movement towards the group's historical best position (gbest), usually taken as c2 = 2; r1 and r2 are random factors, which are independent and belong to the [0,1] random number range; gbest d The highest fitness position found by the entire particle swarm in dimension d is the historical optimal position of the swarm.

[0129] The positions of pbest and gbest are both determined by their fitness values. High-fit particles guide the population to converge toward regions with better parameters, avoiding blind searching.

[0130] When the rate of change of the fitness value is less than the preset threshold or the maximum number of iterations is reached, the corresponding VMD parameter (K, α) is determined to be the optimal VMD decomposition parameter after fitness function verification, which can be directly applied to actual signal processing tasks.

[0131] In one embodiment, historical operational data of the boundary conditions are input as a one-dimensional time series signal to be decomposed into the following model to determine the modal components and center frequencies corresponding to the historical operational data:

[0132]

[0133] Where f(t) represents historical running data; u k ω represents the modal component of the k-th mode after decomposition. k δ is the center frequency of the k-th mode; K is the number of modes after decomposition; ||·|| is the L2 norm; δ t The Dirac distribution function; Let be the partial derivative at time t.

[0134] In the above model, the objective function is essentially to minimize the sum of the bandwidths of all modal components, where the Hilbert transform term... Used to extract u k The analytic signal (t) has a gradient magnitude that reflects the degree of frequency spread. Without bandwidth constraints, the decomposition may result in infinitely narrow or overlapping invalid components; constraints are used to ensure signal integrity. This is achieved by alternately updating the modal components u. k (t) and center frequency ω kThe process continues until convergence (residual < 1e-6), finally outputting K IMF components (such as low-frequency trends, periodic fluctuations, and noise) and their corresponding center frequencies. This decomposes the turbidity data into K modal components, each carrying different frequency band characteristics.

[0135] Therefore, this application independently performs VMD decomposition on the historical operating data for each boundary condition to generate corresponding modal components. For example, the turbidity of desulfurization wastewater is decomposed into K turbidity modal components; the unit load signal is decomposed into K load modal components.

[0136] (2) Figure 2 As shown, from multiple historical process parameters, the mutual information method is used to select target process parameters that are highly correlated with historical operating data. The modal components corresponding to the historical operating data are combined with the target process parameters to reconstruct a new feature vector.

[0137] Taking desulfurization wastewater turbidity as an example, to measure the interdependence between desulfurization wastewater turbidity and related process parameters, historical process parameters related to desulfurization wastewater turbidity are obtained. The mutual information values ​​between desulfurization wastewater turbidity and each process parameter are calculated. Historical process parameters with mutual information values ​​higher than preset values ​​are identified as target process parameters highly correlated with desulfurization wastewater turbidity, such as pH, reagent level, and flow rate. Then, the modal components corresponding to the predicted values ​​of boundary conditions are combined and reconstructed with the process parameters highly correlated with each boundary prediction value. For example, the desulfurization wastewater turbidity modal components and the target process parameters highly correlated with wastewater turbidity are aligned by time and horizontally concatenated to form a new feature vector. Assuming the wastewater turbidity VMD is decomposed into 7 turbidity IMF components, and horizontally concatenated with lime slurry dosage and pH value, a new 9-dimensional feature vector is formed. Similarly, assuming the unit load VMD is decomposed into 3 conforming IMF components and concatenated with boiler pressure and steam flow rate, a new 5-dimensional feature vector is formed.

[0138] (3) Input the new feature vector into the CNN part of the VMD-CNN-LSTM model to extract the spatial coupling features between the target process parameters and historical operation data (such as the nonlinear relationship between the amount of lime milk added and the turbidity).

[0139] In CNN network design, to preserve the temporal information of time series, a one-dimensional convolutional structure without pooling is used. By setting multiple convolutional kernels with different receptive fields, multi-scale local features of water quality data can be extracted. Typically, CNN networks consist of alternating convolution and pooling components. Convolutional kernels extract local correlation features of water quality sequences through sliding window operations, while traditional pooling layers reduce feature dimensionality through downsampling operations, thereby reducing the number of network parameters and computational load.

[0140] For the new feature vectors reconstructed after VMD decomposition, this application uses non-pooling convolutional layers to extract the spatial coupling features and local dependencies of water quality data. Each node of the Flatten layer is connected to all nodes of the previous layer to summarize and output the previously extracted features. Finally, the historical running data of the boundary conditions is refined by the CNN network into the input format of the LSTM layer, which is more advantageous for time series information.

[0141] (4) The spatial coupling features are input into the LSTM part of the VMD-CNN-LSTM model to extract temporal features (such as the fluctuation pattern of unit load). Specifically, after the LSTM layer inputs data, it outputs the hidden state as a temporal feature expression, which includes the time evolution pattern (such as load fluctuation and water quality change trend). For example, each LSTM unit outputs the temporal feature vector corresponding to the modal component.

[0142] Each unit in an LSTM neural network consists of an input gate, a forget gate, an output gate, and a memory unit. By controlling the input, forget, and output gates, LSTM adds or deletes current and past time state information into the memory unit. The cell state vector Ct is the core memory unit for the current time step, responsible for conveying long-term dependencies in the time series. t-1 and h t Let z be the hidden state vectors at time t-1 and the current time t, respectively, and let σ be the Sigmoid activation function, as shown in the following formula. f z i z o These are the three gate states inside the LSTM, corresponding to the forget gate, input gate, and output gate, respectively. W f W i W o These are the weight coefficients for the forget gate, input gate, and output gate, respectively, and z is the value after the result is transformed by the sigmoid function.

[0143] z f =σ(W f ×[x t h t-1 ])

[0144] z i =σ(W i ×[x t h t-1 ])

[0145] z o =σ(W o ×[x t h t-1 ])

[0146] (5) Finally, the time-series features output by multiple LSTM units are reconstructed, for example, by weighted fusion, to generate predicted values ​​for boundary conditions, such as the boundary predicted values ​​for desulfurization wastewater turbidity, desulfurization wastewater influent flow rate, and unit load. The weights in the weighted fusion process can be obtained through an attention mechanism.

[0147] Target operating conditions are selected from the operating target condition database of the desulfurization wastewater treatment system based on the boundary conditions at the next time step. In this application, a corresponding operating target condition database for different operating conditions is pre-established based on the FCM algorithm, so that target operating conditions that meet the boundary conditions at the next time step can be selected from the operating target condition database. The method for establishing the operating target condition database of the desulfurization wastewater treatment system is as follows:

[0148] First, historical sample data is acquired, which should comprehensively cover various normal operating conditions. Then, an improved FCM algorithm (introducing a kernel function) is used to cluster the sample data, obtaining the membership matrix and cluster centers. The specific clustering process is as follows:

[0149] ① Predetermine the range of cluster sizes. Specifically, determine the range [kmin, kmax] of the number of clusters k based on the data characteristics of historical sample data from the desulfurization wastewater treatment system and the analysis requirements. Data characteristics refer to the distribution of the data, data dimensions, and the correlation between different parameters. For example, if the data shows a clear grouping trend across different parameter dimensions, the range of cluster sizes can be appropriately expanded; if the data distribution is relatively uniform, the range of cluster sizes should be relatively smaller. Analysis requirements consider the level of detail required for classifying operating conditions in actual business operations. If a more detailed division of operating conditions is needed for more precise optimization and control, the range of cluster sizes can be set larger; conversely, if only a general distinction between different operating conditions is needed, the range of cluster sizes should be smaller. For example, if preliminary analysis of historical operating data reveals 3-5 relatively obvious distribution patterns in key parameters such as unit load and desulfurization wastewater turbidity, then the range of cluster sizes k can be determined. min =3,k max =5.

[0150] ② Initialize the preset parameters: Given a fuzziness coefficient m and the number of clusters c, randomly select c samples as a cluster center set. The fuzziness coefficient m controls the degree of fuzziness in the clustering; the larger m is, the more fuzzy the clustering result, and the smaller the difference in membership degree between sample points belonging to different classes; the smaller m is, the clearer the clustering result, and the greater the difference in membership degree between sample points belonging to different classes. In practical applications, based on extensive experiments and experience, m=2 is a commonly used value, achieving a good balance between cluster clarity and fuzziness. Number of clusters c and cluster center set: Given an initial number of clusters c=k... min Then, c samples are randomly selected from the sample data to form a cluster center set. For example, if c min If the number of cluster centers is 3, then 3 samples are randomly selected from the historical data as the initial 3 cluster centers.

[0151] ③ Calculate the membership degree of each group of samples and the cluster center of each cluster in the historical sample data according to the preset parameters:

[0152] For each sample point x i (i = 1, 2, ..., n, where n is the total number of samples) and each cluster center v j (j=1,2,...,c), calculate the sample point x according to the following membership degree calculation formula. i The membership degree u of cluster j ij :

[0153] Calculate the membership degree of each sample point using the following membership degree calculation formula:

[0154]

[0155] Among them, u ij Let d be the membership degree of the j-th sample point belonging to cluster i. ij is the Euclidean distance from the j-th sample to the i-th cluster center, c is the number of clusters, and m is the fuzzification coefficient;

[0156] The membership degree u ij Substitute the values ​​into the following formula to update the cluster centers:

[0157]

[0158] Among them, V i The weighted average sample of cluster center i is obtained by weighting all samples according to their membership degree; u ij denoted as the membership degree of the j-th sample point belonging to cluster i; m is the fuzzification coefficient, usually taken as 2; n is the total number of samples.

[0159] Repeat the membership calculation and cluster center update steps above until the convergence condition is met (such as the change in cluster centers being less than a very small threshold, or the number of iterations reaching a preset maximum value).

[0160] In one embodiment of this application, the clustering effectiveness evaluation function value corresponding to different numbers of clusters is calculated based on the membership degree of each group of samples and the cluster centers of each cluster. The optimal number of fuzzy clusters is then determined based on the clustering effectiveness evaluation function value corresponding to different numbers of clusters. Commonly used clustering effectiveness evaluation functions include the partition coefficient (PC), partition entropy (PE), and the Xie-Beni index. For example, the membership degree of each group of samples corresponding to different numbers of clusters and the cluster centers of each cluster are substituted into the following effectiveness evaluation function to calculate the clustering effectiveness evaluation function value corresponding to different numbers of clusters:

[0161]

[0162] Where XB(U,V) is the clustering effectiveness evaluation function value; u ij d(x) represents the membership degree of the j-th sample point to cluster i; j v i ) is the sample x j With cluster center v i Euclidean distance; d(v i v j ) is the cluster center v i To the cluster center v j The Euclidean distance; c is the number of clusters; m is the fuzziness coefficient; n is the total number of samples.

[0163] Calculate the clustering effectiveness evaluation function values ​​for different cluster numbers k. Based on the properties of the evaluation function, select the cluster number k that optimizes the evaluation function as the optimal fuzzy cluster number. For example, the smaller the XB value, the better the clustering effect; therefore, select the cluster number k with the smallest XB value as the optimal fuzzy cluster number. Similarly, the cluster number k with the largest PC or smallest PE can be selected as the optimal fuzzy cluster number.

[0164] After obtaining the membership matrix and cluster centers of the samples, the sample data are classified according to the membership matrix. Taking the operating cost of the desulfurization wastewater treatment system as the decision objective, the target class with the minimum operating cost at the cluster center is found.

[0165] Based on the calculated membership matrix, each sample point is assigned to the cluster with the highest membership degree. For example, for sample point x... i , if u ij It is all u il If the largest value is found in (l = 1, 2, ..., c), then the sample point x will be selected.i Assign it to cluster j.

[0166] For each cluster, the operating cost of the desulfurization wastewater treatment system at the cluster center is calculated, with the operating cost as the decision objective. The operating cost can be calculated by establishing a total operating cost function. The cost model can consider the impact of factors such as lime slurry dosage, organic sulfur dosage, flocculant dosage, coagulant aid dosage, and dewatering machine dosage on the cost. The relative costs at the cluster centers of different clusters are compared to find the target cluster with the minimum relative cost.

[0167] In one embodiment of this application, the total operating cost function is determined based on the sewage discharge cost function, the electricity consumption cost function, and the chemical dosage cost.

[0168] (1) Electricity cost function: The electricity cost function is defined as the total energy consumption (including the energy consumption of the pump system, mixing system, sludge treatment system and dosing system) multiplied by the on-grid electricity price of the region where the power plant is located.

[0169] This application simplifies the energy consumption of wastewater treatment, considering only controllable energy consumption with significant energy-saving potential. The energy consumption model for the desulfurization wastewater treatment system is defined as follows:

[0170] W E =W t +W m +W s +W d ;

[0171]

[0172] Among them, W E The total power consumption (kWh) of the desulfurization wastewater treatment system, W t For the energy consumption of the pump system (kWh), W m For the energy consumption of the hybrid system (kWh), W s Energy consumption (kWh) for sludge treatment system, W d For the energy consumption (kWh) of the dosing system, C dh The power consumption cost of the desulfurization wastewater treatment system (yuan / kWh), P dj The on-grid electricity price for the region where the power plant is located (yuan / kWh), and h represents the power generation (kWh).

[0173] (2) Discharge cost function: The discharge cost function is defined as a function of the amount of taxable water pollutants discharged, the equivalent value, the regional environmental tax rate, and the hourly average concentration of suspended solids in desulfurization wastewater.

[0174] The Environmental Protection Tax Law of the People's Republic of China establishes an environmental tax collection and management system based on pollution equivalent accounting. This law clearly defines a progressive tax incentive mechanism: when the concentration of taxable water pollutants discharged by an enterprise is 30% lower than the national and local emission standards, it can enjoy a 75% reduction or exemption of environmental protection tax; if the concentration is further lower than the standard by 50%, the reduction or exemption ratio is adjusted to 50%. The definition of pollution discharge costs is as follows:

[0175]

[0176] Where λ is the environmental protection tax reduction coefficient; Qf is the discharge volume of Class I water pollutants (L), and P i Qs represents the equivalent value of Class I water pollutants (t), Qs represents the discharge amount of Class II water pollutants (L), and P represents the discharge amount of Class II water pollutants (L). j Let τ be the equivalent value of Class II water pollutants (t), τ be the regional environmental tax rate (yuan / t), and c be the equivalent value of Class II water pollutants (t). ds For emission standard limits (mg / L), c v The concentration of suspended solids in desulfurization wastewater is the hourly average (mg / L).

[0177] From an environmental management perspective, water pollutants are divided into two categories. Category I pollutants include highly toxic substances that accumulate easily in the environment and cause irreversible ecological damage (such as heavy metals and radioactive elements). Typical examples include total mercury and alkyl mercury. The discharge volume for Category I pollutants refers to the total amount of pollutants discharged from the outlet (e.g., kg / hour). The control requirement for Category I pollutants is that they must meet standards at the workshop or production facility discharge outlet, and dilution discharge is prohibited. Category II pollutants are relatively less harmful and biodegradable conventional pollutants. Typical examples include COD (Chemical Oxygen Demand), BOD (Biochemical Oxygen Demand), and SS (Suspended Solids). The discharge volume for Category II pollutants refers to the total amount of pollutants discharged into the external environment from the enterprise's total discharge outlet. The control requirement for Category II pollutants is to control the concentration limits at the total discharge outlet.

[0178] (3) The cost function of dosage can be obtained by multiplying the dosage by the unit price.

[0179] The total operating cost function is obtained by adding the costs obtained from the electricity consumption cost function, the sewage discharge cost function, and the chemical dosage cost function.

[0180] Next, calculate the Euclidean distance from all sample points in the target class to the class center, and return the closest sample as the operating baseline value under the current operating conditions. For example, under the operating conditions of a unit load of 300MW and a turbidity of 50NTU in the desulfurization wastewater outlet, the corresponding operating baseline values ​​are obtained, including the dosage of lime slurry, organic sulfur, flocculant, coagulant aid, dewatering machine, pH value, etc.

[0181] To determine whether a condition is in a steady-state operating condition, a sampling sliding window is set. The size of the sliding window can be determined based on the actual data acquisition frequency and analysis requirements. For example, if the data acquisition frequency is once per minute, the sliding window size can be set to 10 minutes, meaning the window contains 10 data points. For the data within the sliding window, the variation range of key parameters (such as desulfurization wastewater turbidity, desulfurization wastewater influent flow rate, unit load, etc.) is calculated. If the variation range of the key parameters is less than a preset threshold, the data within the sliding window is determined to be in a steady-state operating condition. For example, if the variation threshold for desulfurization wastewater turbidity is set to ±5 NTU, and the difference between the maximum and minimum values ​​of desulfurization wastewater turbidity within the sliding window is less than 5 NTU, then the operating condition corresponding to the data within the sliding window is considered a steady-state operating condition. The data determined to be in a steady-state operating condition and its corresponding operating baseline values ​​are stored in the steady-state operating condition database.

[0182] Through the above steps, the operating baseline values ​​of key parameter combinations under typical different operating conditions can be obtained. Then, the operating baseline values ​​under all operating conditions can be obtained by interpolation, and the operating target operating condition library of the desulfurization wastewater treatment system can be established.

[0183] The optimal operating condition that minimizes operating costs is determined from the target operating conditions. The operating costs under different operating conditions are then compared. Operating costs can be calculated using the previously established cost model. The operating condition with the lowest operating cost is selected as the optimal operating condition. The operating parameters corresponding to the optimal operating condition (such as lime slurry dosage, organic sulfur dosage, flocculant dosage, coagulant aid dosage, dewatering machine dosage, pH value, etc.) can serve as a reference for optimizing the operation of the desulfurization wastewater treatment system.

[0184] The optimal operating condition guides the adjustment of wastewater treatment system parameters. Based on the parameters corresponding to the optimal operating condition, the parameters in actual operation are adjusted. The specific adjustment process needs to combine real-time monitoring data and prediction results to ensure that the wastewater treatment system maintains a highly efficient and stable operating state under different operating conditions. Simultaneously, historical data needs to be updated and re-analyzed regularly to adapt to actual conditions such as unit load and coal type changes, continuously optimizing the operating parameters of the wastewater treatment system. For example, the adjustment of operating parameters is guided by the predicted results of desulfurization wastewater discharge constraints and boundary conditions. Turbidity, which serves as a wastewater discharge constraint, is predicted in advance to ensure it is controlled within 70 mg / L. Simultaneously, the predicted results of the desulfurization wastewater treatment system boundary conditions are compared with the steady-state operating condition database. The predicted results include the desulfurization wastewater influent flow rate and unit load. The influent flow rate and unit load with the closest values ​​are selected, and a set of target parameter values ​​is found in the corresponding steady-state operating condition database. These target parameter values ​​are then used to adjust the parameters to be adjusted. For example, the turbidity of wastewater outlet in the steady-state operating condition pool is divided into N1 segments, and the load in the steady-state operating condition pool is divided into N2 segments. Each segment represents a range of values, and each range of values ​​corresponds to a set of target values ​​for parameters to be adjusted.

[0185] Through the above steps, intelligent optimization of the desulfurization wastewater treatment system of coal-fired power plants can be achieved, reducing operating costs and improving wastewater treatment efficiency and quality.

[0186] Taking a large thermal power plant as an example, the plant has two 1000MW ultra-supercritical coal-fired units, both of which use a coagulation-sedimentation process (commonly known as the triple-tank sedimentation method) to treat desulfurization wastewater. The entire desulfurization wastewater treatment system of the power plant includes: an effective volume V = 500m³. 3 The system includes a wastewater buffer tank, a desulfurization wastewater treatment unit, a wastewater lime slurry preparation and dosing unit, a wastewater dosing unit, and a product water collection and transportation unit.

[0187] When establishing the steady-state operating condition database, desulfurization wastewater data from January to February 2024 of the power plant were selected as the test sample. Based on the operating time of each load range of the unit, the unit's operating load was mostly concentrated at around 50-60%, and the unit operating load was taken as 500.131–599.43 MW. Cluster analysis was performed on 653 sets of sample data within the operating condition range of inlet desulfurization wastewater flow rate of 15.446–35.952 mg·L⁻¹.

[0188] (1) Import the operating data of the power plant's desulfurization wastewater treatment system and perform data cleaning to remove abnormal data. Data points with excessive turbidity emission concentration, measurement parameters outside the normal range, and abrupt changes are identified and deleted.

[0189] (2) Using the above-processed healthy 653 unit operation sample data, as shown in Table 1, select the adjustable parameter as the amount of lime milk added.

[0190] Table 1 Relevant parameters under typical loads

[0191]

[0192] (3) The FCM algorithm was used to cluster the sample data, and the energy consumption of desulfurization wastewater treatment was divided into 5 categories from low to high: A, B, C, D, and E. The classification of the optimization decision samples is shown in Table 2. The sample of category A with the lowest operating energy consumption was selected as the optimization decision sample of the desulfurization wastewater treatment system.

[0193] Table 2 Classification of Optimization Decision Samples

[0194]

[0195]

[0196] (4) When the fuzzification constant m is 2, the adaptive function L(1) = 0 when the number of clusters is 1. Figure 3 This is a scatter plot of historical operational data clustered in one embodiment of this application, such as... Figure 3 As shown, the data is divided into three clusters through clustering, with different clusters represented by dots of different colors, and triangles serving as cluster centers.

[0197] (5) Samples classified into the same category are defined as having the same operating mode. Based on this, the average value of the operating parameters under the decision sample A category is selected as the operation optimization strategy under this operating condition, which is also the actual achievable baseline state of the desulfurization wastewater treatment system under this operating condition. It is compared and analyzed with the average value of each parameter in the actual operation under this operating condition, as shown in Table 3.

[0198] Table 3 Comparison of baseline and actual values ​​of desulfurization wastewater treatment system

[0199]

[0200] When using the VMD-CNN-LSTM model to predict the constraints and boundary conditions of wastewater treatment, 150 historical load operation data points at the current time are taken, with a sampling interval of 1 minute.

[0201] (1) Variational Mode Decomposition. Import historical time series data. The desulfurization wastewater quality data has strong fluctuations. The variational mode decomposition method is used to decompose the data to smooth the series, thereby improving the prediction accuracy.

[0202] (2) CNN Neural Network. The CNN method is used to extract a new dataset composed of the main influencing parameters and each group of water quality modal components, and the LSTM method is used to extract the load time series features.

[0203] (3) Prediction. Finally, the water quality prediction results of each modal component are integrated and reconstructed using LSTM, and single-point prediction is made using the new dataset. The idea is, for example, starting from 6, use the first 6 data to predict the 7th; then use the first 7 to predict the 8th... and use the first 159 to predict the 160th. Figure 4 This is a comparison chart of the predicted values ​​and actual values ​​of suspended solids concentration by different algorithms in one embodiment of this application. Through the comparison of multiple algorithms, it can be seen that the method provided in this application has a better prediction effect.

[0204] Finally, guidance is provided for adjusting the operating parameters of the desulfurization wastewater treatment system. Based on the predicted values ​​of constraints and boundary conditions, the corresponding target values ​​for each parameter are found from the target operating condition database, and adjustments are made to the actual operation.

[0205] In one embodiment, step S102 above can be implemented as steps A1-A2 as follows:

[0206] In step A1, the historical operation data is discretized;

[0207] In step A2, the discretized historical running data is input into the VMD-CNN-LSTM model to predict the boundary conditions at the next time step.

[0208] In one embodiment, the process of establishing the target operating condition database for the desulfurization wastewater treatment system can be implemented as follows: steps B1-B4:

[0209] In step B1, the FCM algorithm is used to cluster the sample data to obtain the membership matrix and cluster centers of the samples;

[0210] In step B2, the sample data is classified according to the membership matrix, and the class with the lowest relative cost at the class center is found with the operating cost of the desulfurization wastewater treatment system as the decision objective.

[0211] In step B3, the distance from all sample points in the class to the class center is calculated, and the sample with the closest distance is returned as the operating baseline value under the current working condition;

[0212] In step B4, the sampling sliding window method is used to determine whether the data in the sliding window represents a stable operating condition based on the operating baseline value, so as to obtain the operating target condition library of the desulfurization wastewater treatment system.

[0213] In one embodiment, step B1 above can be implemented as steps B11-B12:

[0214] In step B11, the membership degree of each sample point is calculated according to the following membership degree calculation formula:

[0215]

[0216] Among them, u ij Let d be the membership degree of the j-th sample point belonging to cluster i. ij is the Euclidean distance from the j-th sample to the i-th cluster center, c is the number of clusters, and m is the fuzzification coefficient;

[0217] In step B12, cluster centers are determined according to the following cluster center formula:

[0218]

[0219] Among them, V i u is the weighted average sample of cluster center i. ij Let m be the membership degree of the j-th sample point belonging to cluster i, m be the fuzzification coefficient, and n be the total number of samples.

[0220] In one embodiment, after step B1 above, the method may also be implemented as steps C1-C2:

[0221] In step C1, the clustering effectiveness evaluation function values ​​corresponding to different numbers of clusters are calculated;

[0222] In step C2, the optimal number of fuzzy clusters is determined based on the clustering effectiveness evaluation function values ​​corresponding to different numbers of clusters.

[0223] In one embodiment, step C1 above can be implemented as follows:

[0224] The clustering effectiveness evaluation function value is calculated using the following formula:

[0225]

[0226] Where XB(U,V) is the clustering effectiveness evaluation function value, u ij Let d(x) be the membership degree of sample point xi to cluster j. j v i ) is the sample x j With cluster center v i Euclidean distance, d(v) i v j ) is the cluster center v i To the cluster center v j The Euclidean distance, where c is the number of clusters, m is the fuzziness coefficient, and n is the total number of samples;

[0227] Step C2 above can be implemented as follows:

[0228] The cluster number with the smallest clustering effectiveness evaluation function value is selected as the optimal fuzzy cluster number.

[0229] In one embodiment, step S104 above can be implemented as steps D1-D3 as follows:

[0230] In step D1, the power consumption cost function, the sewage discharge cost function, and the chemical dosage cost are defined;

[0231] In step D2, the total operating cost function is determined based on the sewage discharge cost function, the electricity consumption cost function, and the chemical dosage cost.

[0232] In step D3, the operating cost parameters corresponding to the target operating condition are substituted into the total operating cost function to determine the optimal operating condition that minimizes the operating cost.

[0233] In one embodiment, defining the power consumption cost function in step D1 above can be implemented as follows:

[0234] The electricity cost function is defined as follows:

[0235] W E =W t +W m +W s +W d ;

[0236]

[0237] Among them, W E The total power consumption (kWh) of the desulfurization wastewater treatment system, W t For the energy consumption of the pump system (kWh), W m For the energy consumption of the hybrid system (kWh), W s Energy consumption (kWh) for sludge treatment system, W d For the energy consumption (kWh) of the dosing system, C dh The power consumption cost of the desulfurization wastewater treatment system (yuan / kWh), P dj The on-grid electricity price for the region where the power plant is located (yuan / kWh), and h represents the power generation (kWh).

[0238] In one embodiment, defining the sewage discharge cost function in step D1 above can be implemented as follows:

[0239] The sewage discharge cost function is defined as follows:

[0240]

[0241] Where λ is the environmental protection tax reduction coefficient, Qf is the discharge volume of Class I water pollutants (L), and P iQs represents the equivalent value of Class I water pollutants (t), Qs represents the discharge amount of Class II water pollutants (L), and P represents the equivalent value of Class I water pollutants (t). j Let τ be the equivalent value of Class II water pollutants (t), τ be the regional environmental tax rate (yuan / t), and c be the equivalent value of Class II water pollutants (t). ds For emission standard limits (mg / L), c v The concentration of suspended solids in desulfurization wastewater is the hourly average (mg / L).

[0242] Figure 5 This is a schematic diagram of the structure of a parameter adjustment device according to an embodiment of this application, as shown below. Figure 5 As shown, it includes:

[0243] The data acquisition module 501 is used to collect historical operating data of the desulfurization wastewater treatment system;

[0244] The prediction module 502 is used to predict the boundary conditions for the next moment based on the historical operating data, wherein the boundary conditions include the boundary conditions of the turbidity of the desulfurization wastewater, the boundary conditions of the influent flow rate of the desulfurization wastewater, and the boundary conditions of the unit load.

[0245] The filtering module 503 is used to filter target operating conditions from the operating target operating condition library of the desulfurization wastewater treatment system according to the boundary conditions at the next moment.

[0246] The acquisition module 504 is used to acquire the optimal operating condition that minimizes the operating cost from the target operating condition;

[0247] The adjustment module 505 is used to guide the adjustment of wastewater treatment system parameters according to the optimal operating conditions.

[0248] In one embodiment, the prediction module includes:

[0249] The discrete submodule is used to discretize the historical running data;

[0250] The prediction submodule is used to input the discretized historical running data into the VMD-CNN-LSTM model so as to predict the boundary conditions at the next time step.

[0251] In one embodiment, the establishment module is used to establish a database of target operating conditions for the desulfurization wastewater treatment system, and the establishment module includes:

[0252] The clustering submodule is used to cluster sample data using the FCM algorithm to obtain the membership matrix and cluster centers of the samples.

[0253] The classification submodule is used to classify sample data according to the membership matrix, and with the operating cost of the desulfurization wastewater treatment system as the decision objective, finds the class with the lowest relative cost at the class center.

[0254] The first calculation submodule is used to calculate the distance from all sample points in the class to the class center and return the nearest sample as the operating baseline value under the current working conditions.

[0255] The determination submodule is used to determine whether the data in the sliding window is a stable operating condition based on the operating benchmark value using the sampling sliding window method, so as to obtain the operating target condition library of the desulfurization wastewater treatment system.

[0256] In one embodiment, the clustering submodule is further configured to:

[0257] Calculate the membership degree of each sample point using the following membership degree calculation formula:

[0258]

[0259] Among them, u ij Let d be the membership degree of the j-th sample point belonging to cluster i. ij is the Euclidean distance from the j-th sample to the i-th cluster center, c is the number of clusters, and m is the fuzzification coefficient;

[0260] Cluster centers are determined using the following formula:

[0261]

[0262] Among them, V i u is the weighted average sample of cluster center i. ij Let m be the membership degree of the j-th sample point belonging to cluster i, m be the fuzzification coefficient, and n be the total number of samples.

[0263] In one embodiment, the establishment module includes, further comprising:

[0264] The second calculation submodule is used to calculate the clustering effectiveness evaluation function value corresponding to different numbers of clusters;

[0265] The first determination submodule is used to determine the optimal number of fuzzy clusters based on the clustering effectiveness evaluation function value corresponding to different numbers of clusters.

[0266] In one embodiment, the second computing submodule is further configured to:

[0267] The clustering effectiveness evaluation function value is calculated using the following formula:

[0268]

[0269] Where XB(U,V) is the clustering effectiveness evaluation function value, u ij Let d(x) be the membership degree of sample point xi to cluster j. j v i ) is the sample x j With cluster center vi Euclidean distance, d(v) i v j ) is the cluster center v i To the cluster center v j The Euclidean distance, where c is the number of clusters, m is the fuzziness coefficient, and n is the total number of samples;

[0270] The step of determining the optimal fuzzy cluster number based on the clustering effectiveness evaluation function values ​​corresponding to different cluster numbers includes:

[0271] The cluster number with the smallest clustering effectiveness evaluation function value is selected as the optimal fuzzy cluster number.

[0272] In one embodiment, the acquisition module includes:

[0273] Define a submodule for defining the power consumption cost function, defining the sewage discharge cost function, and determining the chemical dosage cost;

[0274] The second determining submodule is used to determine the total operating cost function based on the sewage discharge cost function, the electricity consumption cost function, and the chemical dosage cost.

[0275] The third determining submodule is used to substitute the operating cost parameters corresponding to the target operating condition into the total operating cost function in order to determine the optimal operating condition that minimizes the operating cost.

[0276] In one embodiment, defining the power consumption cost function includes:

[0277] The electricity cost function is defined as follows:

[0278] W E =W t +W m +W s +W d ;

[0279]

[0280] Among them, W E The total power consumption (kWh) of the desulfurization wastewater treatment system, W t For the energy consumption of the pump system (kWh), W m For the energy consumption of the hybrid system (kWh), W s Energy consumption (kWh) for sludge treatment system, W d For the energy consumption (kWh) of the dosing system, C dh The power consumption cost (yuan / kWh) of the desulfurization wastewater treatment system, P dj The on-grid electricity price (yuan / kWh) for the region where the power plant is located, and h represents the power generation (kWh).

[0281] In one embodiment, defining the sewage discharge cost function includes:

[0282] The sewage discharge cost function is defined as follows:

[0283]

[0284] Where λ is the environmental protection tax reduction coefficient, Qf is the discharge volume of Class I water pollutants (L), and P i Qs represents the equivalent value of Class I water pollutants (t), Qs represents the discharge amount of Class II water pollutants (L), and P represents the equivalent value of Class I water pollutants (t). j Let τ be the equivalent value of Class II water pollutants (t), τ be the regional environmental tax rate (yuan / t), and c be the equivalent value of Class II water pollutants (t). ds For emission standard limits (mg / L), c v The concentration of suspended solids in desulfurization wastewater is the hourly average (mg / L).

[0285] Figure 6 This is a schematic diagram of the hardware structure of a parameter adjustment system according to an embodiment of this application, as shown below. Figure 6 As shown, the parameter adjustment system includes:

[0286] At least one processor 620; and,

[0287] Memory 604 communicatively connected to the at least one processor 620; wherein,

[0288] The memory 604 stores instructions that can be executed by the at least one processor 620 to implement the parameter adjustment method described in any of the above embodiments.

[0289] Reference Figure 6 The parameter adjustment system 600 may include one or more of the following components: processing component 602, memory 604, power supply component 606, multimedia component 608, audio component 610, input / output (I / O) interface 612, sensor component 614, and communication component 616.

[0290] Processing component 602 typically controls parameters to regulate the overall operation of system 600. Processing component 602 may include one or more processors 620 to execute instructions to complete all or part of the steps of the method described above. Furthermore, processing component 602 may include one or more modules to facilitate interaction between processing component 602 and other components. For example, processing component 602 may include a multimedia module to facilitate interaction between multimedia component 608 and processing component 602.

[0291] Memory 604 is configured to store various types of data to support the operation of parameter adjustment system 600. Examples of this data include instructions for any application or method operating on parameter adjustment system 600, such as text, images, videos, etc. Memory 604 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0292] Power supply component 606 provides power to various components of parameter regulation system 600. Power supply component 606 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to parameter regulation system 600.

[0293] The multimedia component 608 includes a screen that provides an output interface between the parameter adjustment system 600 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may sense not only the boundaries of touch or swipe actions but also the duration and pressure associated with the touch or swipe operation. In some embodiments, the multimedia component 608 may also include a front-facing camera and / or a rear-facing camera. When the parameter adjustment system 600 is in an operating mode, such as a shooting mode or a video mode, the front-facing camera and / or the rear-facing camera may receive external multimedia data. Each front-facing camera and rear-facing camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

[0294] Audio component 610 is configured to output and / or input audio signals. For example, audio component 610 includes a microphone (MIC) configured to receive external audio signals when the parameter adjustment system 600 is in an operating mode, such as alarm mode, recording mode, voice recognition mode, and voice output mode. The received audio signals may be further stored in memory 604 or transmitted via communication component 616. In some embodiments, audio component 610 also includes a speaker for outputting audio signals.

[0295] I / O interface 612 provides an interface between processing component 602 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, power buttons, and lock buttons.

[0296] Sensor assembly 614 includes one or more sensors for providing status assessments of various aspects of the parameter adjustment system 600. For example, sensor assembly 614 may include a sound sensor. Additionally, sensor assembly 614 can detect the on / off state of the parameter adjustment system 600, the relative positioning of components (e.g., the display and keypad of the parameter adjustment system 600), and the operating state of the parameter adjustment system 600 or one of its components (e.g., the operating state of the air distribution plate, structural state, discharge scraper, etc.), the orientation or acceleration / deceleration of the parameter adjustment system 600, and temperature changes of the parameter adjustment system 600. Sensor assembly 614 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 614 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, sensor assembly 614 may also include an accelerometer, a gyroscope, a magnetometer, a pressure sensor, a material buildup thickness sensor, or a temperature sensor.

[0297] Communication component 616 is configured to enable parameter adjustment system 600 to provide wired or wireless communication capabilities with other devices and cloud platforms. Parameter adjustment system 600 can access wireless networks based on communication standards, such as WiFi, 2G, or 3G, or combinations thereof. In one exemplary embodiment, communication component 616 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, communication component 616 also includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0298] In an exemplary embodiment, the parameter adjustment system 600 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the parameter adjustment method described in any of the above embodiments.

[0299] This application also provides a computer-readable storage medium, which, when the instructions in the storage medium are executed by a processor corresponding to a parameter adjustment system, enables the parameter adjustment system to implement the parameter adjustment method described in any of the above embodiments.

[0300] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0301] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0302] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0303] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0304] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A parameter adjustment method, characterized in that, include: Collect historical operating data of the desulfurization wastewater treatment system; The boundary conditions for the next moment are predicted based on the historical operating data, wherein the boundary conditions include the boundary conditions of desulfurization wastewater turbidity, the boundary conditions of desulfurization wastewater influent flow rate, and the boundary conditions of unit load. Based on the boundary conditions of the next moment, target operating conditions are selected from the target operating condition library of the desulfurization wastewater treatment system. Obtain the optimal operating condition that minimizes operating costs from the target operating conditions; The optimal operating conditions guide the adjustment of wastewater treatment system parameters.

2. The method as described in claim 1, characterized in that, The step of predicting the boundary conditions for the next moment based on the historical operating data includes: Discretize the historical operation data; The discretized historical running data is input into the VMD-CNN-LSTM model to predict the boundary conditions at the next time step.

3. The method as described in claim 1, characterized in that, The process of establishing the target operating condition database for the desulfurization wastewater treatment system is as follows: The FCM algorithm is used to cluster the sample data to obtain the membership matrix and cluster centers of the samples; The sample data is classified according to the membership matrix. Taking the operating cost of the desulfurization wastewater treatment system as the decision objective, the class with the lowest relative cost at the class center is found. Calculate the distance from all sample points in the class to the class center, and return the closest sample as the operating baseline value under the current working conditions; Based on the aforementioned operating baseline values, the sampling sliding window method is used to determine whether the data within the sliding window represents a stable operating condition, thereby obtaining a database of target operating conditions for the desulfurization wastewater treatment system.

4. The method as described in claim 3, characterized in that, The FCM algorithm is used to cluster the sample data to obtain the membership matrix and cluster centers of the samples, including: Calculate the membership degree of each sample point using the following membership degree calculation formula: Among them, u ij Let d be the membership degree of the j-th sample point belonging to cluster i. ij is the Euclidean distance from the j-th sample to the i-th cluster center, c is the number of clusters, and m is the fuzzification coefficient; Cluster centers are determined using the following formula: Among them, V i u is the weighted average sample of cluster center i. ij Let m be the membership degree of the j-th sample point belonging to cluster i, m be the fuzzification coefficient, and n be the total number of samples.

5. The method as described in claim 1, characterized in that, The step of obtaining the optimal operating condition that minimizes operating costs from the target operating conditions includes: Define the power consumption cost function, define the sewage discharge cost function, and determine the cost of chemical dosage; The total operating cost function is determined based on the sewage discharge cost function, the electricity consumption cost function, and the chemical dosage cost. The operating cost parameters corresponding to the target operating condition are substituted into the total operating cost function to determine the optimal operating condition that minimizes the operating cost.

6. The method as described in claim 1, characterized in that, The defined electricity consumption cost function includes: The electricity cost function is defined as follows: IN E =In t +W m +W s +W d ; Among them, W E The total power consumption (kWh) of the desulfurization wastewater treatment system, W t For the energy consumption of the pump system (kWh), W m For the energy consumption of the hybrid system (kWh), W s Energy consumption (kWh) for sludge treatment system, W d For the energy consumption (kWh) of the dosing system, C dh The power consumption cost of the desulfurization wastewater treatment system (yuan / kWh), P dj The on-grid electricity price for the region where the power plant is located (yuan / kWh), and h represents the power generation (kWh).

7. The method as described in claim 1, characterized in that, The defined sewage discharge cost function includes: The sewage discharge cost function is defined as follows: Where λ is the environmental protection tax reduction coefficient, Qf is the discharge volume of Class I water pollutants (L), and P i Qs represents the equivalent value of Class I water pollutants (t), Qs represents the discharge amount of Class II water pollutants (L), and P represents the discharge amount of Class II water pollutants (L). j Let τ be the equivalent value of Class II water pollutants (t), τ be the regional environmental tax rate (yuan / t), and c be the equivalent value of Class II water pollutants (t). ds For emission standard limits (mg / L), c v The concentration of suspended solids in desulfurization wastewater is the hourly average (mg / L).

8. A parameter adjustment device, characterized in that, include: The data acquisition module is used to collect historical operating data of the desulfurization wastewater treatment system; The prediction module is used to predict the boundary conditions for the next moment based on the historical operating data, wherein the boundary conditions include the boundary conditions of the turbidity of the desulfurization wastewater, the boundary conditions of the influent flow rate of the desulfurization wastewater, and the boundary conditions of the unit load. The filtering module is used to filter target operating conditions from the target operating condition library of the desulfurization wastewater treatment system according to the boundary conditions at the next moment. The acquisition module is used to acquire the optimal operating condition that minimizes the operating cost from the target operating condition; The adjustment module is used to guide the adjustment of wastewater treatment system parameters based on the optimal operating conditions.

9. A parameter adjustment system, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to implement the parameter adjustment method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor corresponding to the parameter adjustment system, the parameter adjustment system is able to implement the parameter adjustment method as described in any one of claims 1-7.

Images