Water supply scheme determination method and device

Abstract

The invention provides a water supply scheme determination method and device. The method comprises the following steps: acquiring multiple groups of real-time water quality information data of multiple water supply sources; preprocessing the multiple groups of real-time water quality information data to obtain multiple groups of target water quality information feature data; respectively inputting the multiple groups of target water quality information feature data into a target water quality information prediction model to obtain multiple groups of water quality information prediction results; and determining a target water supply scheme according to the multiple groups of water quality information prediction results and the multiple water quality information preset ranges. According to the scheme, the water supply scheme is adjusted in time by predicting the change trend of the water quality information, the operation efficiency and safety of the water supply system can be improved, and the water supply cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of water supply systems, and in particular to a method and device for determining a water supply scheme. Background Art

[0002] When supplying water, multiple water sources are often used to supply water at the same time, and the water quality information of different water sources is different. The traditional water supply system management method lacks real-time and accuracy, and cannot adjust the water supply plan in time according to the water quality information of the water supply source, resulting in water resource waste and water supply safety issues. Summary of the invention

[0003] The present invention provides a method and device for determining a water supply plan, which can improve the operating efficiency and safety of a water supply system and reduce water supply costs by predicting the changing trend of water quality information and adjusting the water supply plan in a timely manner.

[0004] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0005] A method for determining a water supply plan, comprising:

[0006] Obtain multiple groups of real-time water quality information data from multiple water supply sources;

[0007] Preprocessing the multiple groups of real-time water quality information data to obtain multiple groups of target water quality information characteristic data;

[0008] Inputting the target water quality information characteristic data of the plurality of groups into the decomposition layer of the target water quality information prediction model for decomposition processing respectively, to obtain a plurality of characteristic data sequences of the plurality of groups;

[0009] Inputting the plurality of characteristic data sequences of the plurality of groups into the intermediate layer of the target water quality information prediction model for processing respectively, and obtaining the plurality of intermediate layer processing results;

[0010] Inputting the processing results of the intermediate layers of the plurality of groups into the prediction layer of the target water quality information prediction model for water quality information prediction processing, thereby obtaining the water quality information prediction results of the plurality of groups; wherein the decomposition layer of the target water quality information prediction model is obtained by training according to the first preset network model, the intermediate layer is obtained by training according to the second preset network model and the third preset network model, and the prediction layer is obtained by training according to the fourth preset network model;

[0011] A target water supply plan is determined based on multiple groups of water quality information prediction results and multiple water quality information preset ranges.

[0012] Optionally, preprocessing is performed on the multiple groups of real-time water quality information data to obtain multiple groups of target water quality information characteristic data, including:

[0013] Cleaning the multiple groups of real-time water quality information data to obtain multiple groups of first processed data;

[0014] Performing transformation processing on multiple groups of the first processed data to obtain multiple groups of second processed data;

[0015] The plurality of groups of the second processed data are subjected to reduction processing to obtain a plurality of groups of the third processed data as the plurality of groups of target water quality information characteristic data.

[0016] Optionally, multiple groups of target water quality information feature data are respectively input into the decomposition layer of the target water quality information prediction model for decomposition processing to obtain multiple groups of multiple feature data sequences, including:

[0017] The target water quality information characteristic data of multiple groups are respectively input into the decomposition layer of the target water quality information prediction model, according to Y v =T v +S v +R v ,v=1,2,...,N for decomposition processing to obtain multiple groups of trend component characteristic data series, seasonal component characteristic data series and residual component characteristic data series;

[0018] Among them, Y v Represents the target water quality information characteristic data at time v; T v represents the trend component; S v represents seasonal component; R v Represents the residual component.

[0019] Optionally, the plurality of characteristic data sequences of the plurality of groups are respectively input into the intermediate layer of the target water quality information prediction model for processing to obtain the plurality of intermediate layer processing results, including:

[0020] Inputting the plurality of groups of trend component characteristic data sequences into the first processing module of the intermediate layer of the target water quality information prediction model for processing respectively, and obtaining the plurality of groups of first intermediate layer processing results;

[0021] Inputting the plurality of groups of seasonal component characteristic data sequences into the second processing module of the intermediate layer of the target water quality information prediction model for processing respectively, and obtaining the plurality of groups of second intermediate layer processing results;

[0022] The plurality of groups of residual component characteristic data sequences are respectively input into the third processing module of the intermediate layer of the target water quality information prediction model for processing to obtain a plurality of groups of third intermediate layer processing results.

[0023] Optionally, multiple groups of trend component characteristic data sequences are respectively input into the first processing module of the intermediate layer of the target water quality information prediction model for processing to obtain multiple groups of first intermediate layer processing results, including:

[0024] Inputting the plurality of groups of trend component characteristic data sequences into the convolutional neural network unit of the first processing module of the middle layer of the target water quality information prediction model respectively for convolution processing to obtain a plurality of groups of first convolution processing results;

[0025] The multiple groups of the first convolution processing results are respectively input into the recurrent neural network unit of the first processing module for processing to obtain multiple groups of first intermediate layer processing results.

[0026] Optionally, multiple groups of seasonal component characteristic data sequences are respectively input into the second processing module of the intermediate layer of the target water quality information prediction model for processing to obtain multiple groups of second intermediate layer processing results, including:

[0027] Inputting the plurality of groups of seasonal component characteristic data sequences into the convolutional neural network unit of the second processing module of the middle layer of the target water quality information prediction model respectively for convolution processing, and obtaining a plurality of groups of second convolution processing results;

[0028] Multiple groups of the second convolution processing results are respectively input into the recurrent neural network unit of the second processing module for processing to obtain multiple groups of second intermediate layer processing results.

[0029] Optionally, multiple groups of residual component feature data sequences are respectively input into the third processing module of the intermediate layer of the target water quality information prediction model for processing to obtain multiple groups of third intermediate layer processing results, including:

[0030] Inputting the plurality of groups of residual component feature data sequences into the convolutional neural network units of the third processing module of the middle layer of the target water quality information prediction model respectively for convolution processing, to obtain a plurality of groups of third convolution processing results;

[0031] The multiple groups of the third convolution processing results are respectively input into the recurrent neural network unit of the third processing module for processing to obtain multiple groups of third intermediate layer processing results.

[0032] Optionally, the plurality of groups of intermediate layer processing results are respectively input into the prediction layer of the target water quality information prediction model for water quality information prediction processing to obtain a plurality of groups of water quality information prediction results, including:

[0033] The multiple groups of intermediate layer processing results are respectively input into the shallow fully connected neural network units of the prediction layer of the target water quality information prediction model for feature fitting processing to obtain multiple groups of water quality information prediction results.

[0034] Optionally, determining a target water supply scheme according to multiple groups of water quality information prediction results and multiple water quality information preset ranges includes:

[0035] Comparing the water quality information prediction results of the plurality of groups with a plurality of preset ranges of water quality information to obtain comparison information;

[0036] A target water supply plan is determined according to the comparison information.

[0037] The present invention also provides a water supply scheme determination device, comprising:

[0038] An acquisition module, used to acquire multiple groups of real-time water quality information data from multiple water supply sources;

[0039] A processing module is used to pre-process multiple groups of real-time water quality information data to obtain multiple groups of target water quality information feature data; input the multiple groups of target water quality information feature data into the decomposition layer of the target water quality information prediction model for decomposition processing to obtain multiple groups of multiple feature data sequences; input the multiple groups of multiple feature data sequences into the intermediate layer of the target water quality information prediction model for processing to obtain multiple groups of intermediate layer processing results; input the multiple groups of intermediate layer processing results into the prediction layer of the target water quality information prediction model for water quality information prediction processing to obtain multiple groups of water quality information prediction results; wherein the decomposition layer of the target water quality information prediction model is trained according to the first preset network model, the intermediate layer is trained according to the second preset network model and the third preset network model, and the prediction layer is trained according to the fourth preset network model; determine the target water supply plan according to the multiple groups of water quality information prediction results and multiple water quality information preset ranges.

[0040] The above solution of the present invention includes at least the following beneficial effects:

[0041] The above-mentioned scheme of the present invention obtains multiple groups of real-time water quality information data of multiple water supply sources; pre-processes the multiple groups of real-time water quality information data to obtain multiple groups of target water quality information characteristic data; respectively inputs the multiple groups of target water quality information characteristic data into the decomposition layer of the target water quality information prediction model for decomposition processing to obtain multiple groups of multiple characteristic data sequences; respectively inputs the multiple groups of multiple characteristic data sequences into the intermediate layer of the target water quality information prediction model for processing to obtain multiple groups of intermediate layer processing results; respectively inputs the multiple groups of intermediate layer processing results into the prediction layer of the target water quality information prediction model for water quality information prediction processing to obtain multiple groups of water quality information prediction results; determines the target water supply plan based on the multiple groups of water quality information prediction results and multiple water quality information preset ranges; thereby, the water supply plan can be adjusted in time by predicting the changing trend of the water quality information, which can improve the operating efficiency and safety of the water supply system and reduce the water supply cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1is a flow chart of a method for determining a water supply scheme provided by an embodiment of the present invention;

[0043] Figure 2 is a neural network structure diagram of a target water quality information prediction model provided by an embodiment of the present invention;

[0044] Figure 3 is a neural network structure diagram of a recurrent neural network unit provided in an embodiment of the present invention;

[0045] Figure 4 A module diagram of a device for determining a water supply plan provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0047] like Figure 1 As shown, an embodiment of the present invention provides a method for determining a water supply scheme, comprising:

[0048] Step 11, obtaining multiple groups of real-time water quality information data of multiple water supply sources;

[0049] Step 12, preprocessing the multiple groups of real-time water quality information data to obtain multiple groups of target water quality information characteristic data;

[0050] Step 13, inputting the multiple groups of target water quality information characteristic data into the decomposition layer of the target water quality information prediction model for decomposition processing, and obtaining multiple groups of multiple characteristic data sequences;

[0051] Step 14, inputting the plurality of characteristic data sequences of the plurality of groups into the intermediate layer of the target water quality information prediction model for processing respectively, and obtaining the plurality of intermediate layer processing results;

[0052] Step 15, respectively inputting the multiple groups of intermediate layer processing results into the prediction layer of the target water quality information prediction model for water quality information prediction processing to obtain multiple groups of water quality information prediction results; wherein the decomposition layer of the target water quality information prediction model is trained according to the first preset network model, the intermediate layer is trained according to the second preset network model and the third preset network model, and the prediction layer is trained according to the fourth preset network model;

[0053] Step 16: Determine a target water supply plan based on multiple groups of water quality information prediction results and multiple water quality information preset ranges.

[0054] In this embodiment, multiple groups of real-time water quality information data of multiple water supply sources are obtained; the multiple groups of real-time water quality information data are preprocessed to obtain multiple groups of target water quality information characteristic data; the multiple groups of target water quality information characteristic data are respectively input into the decomposition layer of the target water quality information prediction model for decomposition processing to obtain multiple groups of multiple characteristic data sequences; the multiple groups of multiple characteristic data sequences are respectively input into the intermediate layer of the target water quality information prediction model for processing to obtain multiple groups of intermediate layer processing results; the multiple groups of intermediate layer processing results are respectively input into the prediction layer of the target water quality information prediction model for water quality information prediction processing to obtain multiple groups of water quality information prediction results; the target water supply plan is determined based on the multiple groups of water quality information prediction results and multiple water quality information preset ranges; thereby, the water supply plan can be adjusted in time by predicting the changing trend of the water quality information, the operating efficiency and safety of the water supply system can be improved, and the water supply cost can be reduced.

[0055] In the solution of the present invention, multiple groups of real-time water quality information data, such as temperature, pH value, turbidity, dissolved oxygen, etc., of multiple water supply sources are collected in real time through sensors, meters and other equipment.

[0056] In an optional embodiment of the present invention, step 12 includes:

[0057] Step 121, cleaning the multiple groups of real-time water quality information data to obtain multiple groups of first processed data;

[0058] Step 122, transforming the plurality of groups of the first processed data to obtain a plurality of groups of second processed data;

[0059] Step 123, performing reduction processing on the multiple groups of the second processed data to obtain multiple groups of third processed data as multiple groups of target water quality information characteristic data.

[0060] In this embodiment, multiple groups of real-time water quality information data are cleaned to remove noise, abnormal values, and duplicate values ​​in the data, so as to improve the quality of the data;

[0061] By performing transformation processing on the plurality of groups of the first processed data, specifically, performing normalization processing and smoothing processing on the plurality of groups of the first processed data, the data is converted into a form more suitable for subsequent analysis, so as to facilitate subsequent analysis;

[0062] By performing reduction processing on multiple groups of the second processed data, specifically, performing dimensionality reduction processing and clustering processing on multiple groups of the second processed data, the complexity of the data is reduced and the efficiency of subsequent analysis is improved;

[0063] By performing the above-mentioned processing on multiple groups of the real-time water quality information data, it is possible to improve data quality, reduce data complexity, facilitate subsequent analysis, and help improve the efficiency of subsequent data analysis.

[0064] In an optional embodiment of the present invention, step 13 includes:

[0065] The target water quality information characteristic data of multiple groups are respectively input into the decomposition layer of the target water quality information prediction model, according to Y v =T v +S v +R v ,v=1,2,...,N for decomposition processing to obtain multiple groups of trend component characteristic data series, seasonal component characteristic data series and residual component characteristic data series;

[0066] Among them, Y v Represents the target water quality information characteristic data at time v; T v represents the trend component; S v represents seasonal component; R v Represents the residual component.

[0067] In this embodiment, in practice, the real-time water quality information data will change with the seasons. At the same time, the change of the real-time water quality information data has a trend of movement. In addition, the real-time water quality information data is affected by sudden pollution. In order to improve the accuracy of the water quality information prediction results, in this embodiment, according to Y v =T v +S v +R v , v = 1, 2, ..., N decomposes the multiple groups of target water quality information characteristic data to obtain multiple groups of trend component characteristic data sequences, seasonal component characteristic data sequences and residual component characteristic data sequences;

[0068] A cycle has n p Sample points, the sample points at the same position in different periods form a subsequence, that is, there are n p subsequences; is the trend component obtained after the kth cycle; is the seasonal component obtained after the kth cycle; where

[0069] Specific process:

[0070] Detrending,

[0071] Subsequence smoothing: Each subsequence is subjected to the LOESS process, with one time point before and after each, and the length of the combined sequence is (N+2*n p ) time series

[0072] Low pass filtering: Perform 3 times with length n p 、n p , 3. Perform LOESS process once to obtain a time series of length N. Remove periodic differences;

[0073] Trend removal:

[0074] Seasonal items removed,

[0075] Trend smoothing: Perform the LOESS process and get

[0076]

[0077] In an optional embodiment of the present invention, step 14 includes:

[0078] Step 141, inputting the plurality of groups of trend component characteristic data sequences into the first processing module of the intermediate layer of the target water quality information prediction model for processing, and obtaining the plurality of groups of first intermediate layer processing results;

[0079] Step 142, inputting the plurality of groups of seasonal component characteristic data sequences into the second processing module of the intermediate layer of the target water quality information prediction model for processing, and obtaining a plurality of groups of second intermediate layer processing results;

[0080] Step 143, inputting the plurality of groups of residual component characteristic data sequences into the third processing module of the intermediate layer of the target water quality information prediction model for processing, to obtain a plurality of groups of third intermediate layer processing results.

[0081] In this embodiment, if Figure 2 As shown, the middle layer of the target water quality information prediction model is horizontally divided into three processing modules: a first processing module, a second processing module and a third processing module;

[0082] The first processing module, the second processing module and the third processing module all include a convolutional neural network unit and a recurrent neural network unit; the spatial features of the sequence are learned by the convolutional neural network unit, and trend features (collectively referred to as spatial features here) are extracted;

[0083] Specifically, the convolution kernel size is k, and the one-dimensional convolution at time point t is:

[0084]

[0085] Among them, x t is the sample value at time t; w i is the decay rate of k-1 sample points before this moment;

[0086] The temporal features of the sequence are learned through recurrent neural network units, which tend to extract long-term features and gradual features;

[0087] Among them, the recurrent neural network unit adopts the GRU (gated recurrent unit) structure, as shown in the following figure: Figure 3 As shown in the figure, the GRU structure is simpler, with fewer parameters calculated during forward and backward propagation, and faster convergence;

[0088] The recurrent neural network unit includes: input layer, hidden layer and output layer; X t is the input unit; S t is a hidden unit; O t is the output unit; L t is the loss function; Y t is the true value; the weight parameter W of the hidden layer s It will be output to the next layer along with the input cycle; the input of the recurrent neural network unit is a time series: X = [X 1 ,X 2 ,...,X n ]; the output is the prediction result; W represents the connection weight matrix between each unit; at time t, the state S of the hidden layer satisfies the following update method: S t =σ(X t ·W x +S t-1 ·W s +b s );

[0089] Among them, b s is the bias term; σ is the activation function, usually sigmoid or tanh;

[0090] The output at time t is: Q t =S t ·W o +b s ;

[0091] The predicted value output is:

[0092] The total loss function is:

[0093] The above is the forward propagation process of the recurrent neural network unit;

[0094] The calculation of each parameter is done by back propagation; where Wo and b s Calculated by chain rule; where W s and W x Calculated by the derivative rule of multivariate functions;

[0095]

[0096]

[0097] The GRU structure includes: reset gate and update gate;

[0098] The reset gate is used to control how much previous data is retained, using r t Represents the reset gate, when r t =0, it means forgetting all previous data and resetting to the state of the current input data;

[0099] r t =σ(W r x t +U r h t-1 +b r );

[0100] Where σ is the sigmoid function; x t is the input data; h t-1 is the candidate data of the previous moment; W r and U r is to reset the weight of the gate; b r To reset the gate bias;

[0101] The update gate is used to control the update and retention of data. t represents the update gate, when z t =0, it means to keep all previous data;

[0102] z t =σ(W z x t +U z h t-1 +b z );

[0103] Where σ is the sigmoid function; x t is the input data; h t-1 is the candidate data of the previous moment; W z and U z is the weight of the update gate; b r is the bias of the update gate;

[0104] Current candidate values:

[0105] rt is the data for resetting the gate; x t is the input data; h t-1 is the candidate data of the previous moment; W and U are learnable parameters; b is the bias; ⊙ is the element product;

[0106] Output data:

[0107] (1-z t )⊙h t-1 For previously retained data; For updated data.

[0108] In an optional embodiment of the present invention, step 141 includes:

[0109] Step 1411, inputting the plurality of groups of trend component characteristic data sequences into the convolutional neural network unit of the first processing module of the middle layer of the target water quality information prediction model for convolution processing, and obtaining a plurality of groups of first convolution processing results;

[0110] Step 1412, input multiple groups of the first convolution processing results into the recurrent neural network units of the first processing module for processing respectively, to obtain multiple groups of first intermediate layer processing results.

[0111] In this embodiment, multiple groups of trend component feature data sequences are respectively input into the convolution layer of the convolutional neural network unit of the first processing module of the intermediate layer of the target water quality information prediction model to perform filter operations, extract spatial feature information in the trend component feature data sequence, and obtain multiple groups of first operation results; multiple groups of first operation results are respectively input into the pooling layer of the convolutional neural network unit of the first processing module to perform pooling operations, reduce the spatial size of the data, reduce the number of network parameters, speed up the calculation, and obtain multiple groups of first pooling results; multiple groups of first pooling results are input into the fully connected layer of the convolutional neural network unit of the first processing module for regression processing, and the learned implicit features are mapped to the sample space to obtain multiple groups of prediction task results as multiple groups of first convolution processing results;

[0112] Input multiple groups of the first convolution processing results into the input layer of the recurrent neural network unit of the first processing module as the first input layer output results; input multiple groups of the first input layer output results and the first input layer target parameters into the hidden layer of the recurrent neural network unit of the first processing module for processing, and obtain multiple groups of first hidden layer output results; input multiple groups of the first hidden layer output results and the first hidden layer target parameters into the output layer of the recurrent neural network unit of the first processing module for processing, and obtain multiple groups of output results as multiple groups of first intermediate layer processing results.

[0113] In an optional embodiment of the present invention, step 142 includes:

[0114] Step 1421, inputting the plurality of groups of seasonal component characteristic data sequences into the convolutional neural network units of the second processing module of the middle layer of the target water quality information prediction model for convolution processing, and obtaining a plurality of groups of second convolution processing results;

[0115] Step 1422, input multiple groups of the second convolution processing results into the recurrent neural network unit of the second processing module for processing, to obtain multiple groups of second intermediate layer processing results.

[0116] In this embodiment, multiple groups of seasonal component characteristic data sequences are respectively input into the convolution layer of the convolutional neural network unit of the second processing module of the intermediate layer of the target water quality information prediction model to perform filter operations, extract spatial feature information in the seasonal component characteristic data sequences, and obtain multiple groups of second operation results; multiple groups of second operation results are respectively input into the pooling layer of the convolutional neural network unit of the second processing module to perform pooling operations, reduce the spatial size of the data, reduce the number of network parameters, speed up the calculation, and obtain multiple groups of second pooling results; multiple groups of second pooling results are respectively input into the fully connected layer of the convolutional neural network unit of the second processing module for regression processing, and the learned implicit features are mapped to the sample space to obtain multiple groups of prediction task results as multiple groups of second convolution processing results;

[0117] Input multiple groups of the second convolution processing results respectively into the input layer of the recurrent neural network unit of the second processing module as the second input layer output results; input multiple groups of the second input layer output results and the second input layer target parameters respectively into the hidden layer of the recurrent neural network unit of the second processing module for processing to obtain multiple groups of second hidden layer output results; input multiple groups of the second hidden layer output results and the second hidden layer target parameters into the output layer of the recurrent neural network unit of the second processing module for processing, and obtain multiple groups of output results as multiple groups of second intermediate layer processing results.

[0118] In an optional embodiment of the present invention, step 143 includes:

[0119] Step 1431, inputting the plurality of groups of residual component feature data sequences into the convolutional neural network units of the third processing module of the middle layer of the target water quality information prediction model respectively for convolution processing, to obtain a plurality of groups of third convolution processing results;

[0120] Step 1432, input the multiple groups of the third convolution processing results into the recurrent neural network units of the third processing module for processing respectively, to obtain multiple groups of third intermediate layer processing results.

[0121] In this embodiment, multiple groups of residual component feature data sequences are input into the convolution layer of the convolutional neural network unit of the third processing module of the intermediate layer of the target water quality information prediction model to perform filter operations, extract spatial feature information in the residual component feature data sequence, and obtain multiple groups of third operation results; multiple groups of third operation results are respectively input into the pooling layer of the convolutional neural network unit of the third processing module for pooling operations, thereby reducing the spatial size of the data, reducing the number of network parameters, and speeding up the calculation to obtain multiple groups of third pooling results; multiple groups of third pooling results are respectively input into the fully connected layer of the convolutional neural network unit of the third processing module for regression processing, and the learned implicit features are mapped to the sample space to obtain multiple groups of prediction task results as multiple groups of third convolution processing results;

[0122] Input the multiple groups of third convolution processing results into the input layer of the recurrent neural network unit of the third processing module as the third input layer output results; input the multiple groups of third input layer output results and the third input layer target parameters into the hidden layer of the recurrent neural network unit of the third processing module for processing, and obtain multiple groups of third hidden layer output results; input the multiple groups of third hidden layer output results and the third hidden layer target parameters into the output layer of the recurrent neural network unit of the third processing module for processing, and obtain multiple groups of output results as multiple groups of third intermediate layer processing results.

[0123] In an optional embodiment of the present invention, step 15 includes:

[0124] The multiple groups of intermediate layer processing results are respectively input into the shallow fully connected neural network units of the prediction layer of the target water quality information prediction model for feature fitting processing to obtain multiple groups of water quality information prediction results.

[0125] In this embodiment, multiple groups of first intermediate layer processing results, second intermediate layer processing results and third intermediate layer processing results are respectively input into the input layer of the shallow fully connected neural network unit of the prediction layer of the target water quality information prediction model as input layer results, multiple groups of input layer results and first target parameters are respectively input into the hidden layer of the shallow fully connected neural network unit of the prediction layer for processing to obtain multiple groups of hidden layer results, multiple groups of hidden layer results and second target parameters are respectively input into the output layer of the shallow fully connected neural network unit of the prediction layer for regression processing, and the obtained multiple groups of output results are used as multiple groups of water quality information prediction results.

[0126] In an optional embodiment of the present invention, step 16 includes:

[0127] Step 161, comparing the water quality information prediction results of the plurality of groups with a plurality of preset ranges of water quality information to obtain comparison information;

[0128] Step 162: Determine a target water supply plan based on the comparison information.

[0129] In this embodiment, according to the water supply requirements, multiple water quality information preset ranges are set, and the multiple water quality information preset ranges are: an unqualified water supply preset range, a qualified water supply preset range, and an excellent water supply preset range;

[0130] Compare the water quality information prediction results of the plurality of groups with the plurality of preset water quality information ranges, determine the preset water quality information range to which the water quality information prediction results of each group belong, and obtain the belonging range information as comparison information;

[0131] If the water quality information prediction result falls within the preset range of unqualified water supply, the water supply source corresponding to the water quality information prediction result needs to stop supplying water and conduct pollution investigation;

[0132] If the water quality information prediction result falls within the preset range of qualified water supply, the water supply source corresponding to the water quality information prediction result can be used for water supply;

[0133] If the water quality information prediction result belongs to the preset range of excellent water supply, the water supply source corresponding to the water quality information prediction result can be used for water supply;

[0134] When the water quality information prediction results of multiple groups all fall within the preset range of qualified water supply, water supply is carried out based on the principle of proximity, using a nearby water supply source to determine the target water supply plan;

[0135] When the water quality information prediction results of multiple groups all belong to the preset range of excellent water supply, the water supply is carried out according to the principle of proximity, using a nearby water supply source to determine the target water supply plan;

[0136] When the water quality information prediction results of multiple groups belong to the qualified water supply preset range and the excellent water supply preset range respectively, the water supply source whose water quality information prediction results belong to the excellent water supply preset range is used for water supply. At the same time, in accordance with the proximity principle, a nearby water supply source is used for water supply to determine the target water supply plan.

[0137] like Figure 4 As shown, the embodiment of the present invention further provides a water supply scheme determining device 40, comprising:

[0138] An acquisition module 41 is used to acquire multiple groups of real-time water quality information data of multiple water supply sources;

[0139] The processing module 42 is used to pre-process the multiple groups of real-time water quality information data to obtain multiple groups of target water quality information characteristic data; input the multiple groups of target water quality information characteristic data into the decomposition layer of the target water quality information prediction model for decomposition processing to obtain multiple groups of multiple characteristic data sequences; input the multiple groups of multiple characteristic data sequences into the intermediate layer of the target water quality information prediction model for processing to obtain multiple groups of intermediate layer processing results; input the multiple groups of intermediate layer processing results into the prediction layer of the target water quality information prediction model for water quality information prediction processing to obtain multiple groups of water quality information prediction results; wherein the decomposition layer of the target water quality information prediction model is trained according to the first preset network model, the intermediate layer is trained according to the second preset network model and the third preset network model, and the prediction layer is trained according to the fourth preset network model; determine the target water supply plan according to the multiple groups of water quality information prediction results and multiple water quality information preset ranges.

[0140] Optionally, preprocessing is performed on the multiple groups of real-time water quality information data to obtain multiple groups of target water quality information characteristic data, including:

[0141] Cleaning the multiple groups of real-time water quality information data to obtain multiple groups of first processed data;

[0142] Performing transformation processing on multiple groups of the first processed data to obtain multiple groups of second processed data;

[0143] The plurality of groups of the second processed data are subjected to reduction processing to obtain a plurality of groups of the third processed data as the plurality of groups of target water quality information characteristic data.

[0144] Optionally, multiple groups of target water quality information feature data are respectively input into the decomposition layer of the target water quality information prediction model for decomposition processing to obtain multiple groups of multiple feature data sequences, including:

[0145] The target water quality information characteristic data of multiple groups are respectively input into the decomposition layer of the target water quality information prediction model, according to Y v =T v +S v +R v ,v=1,2,...,N for decomposition processing to obtain multiple groups of trend component characteristic data series, seasonal component characteristic data series and residual component characteristic data series;

[0146] Among them, Y v Represents the target water quality information characteristic data at time v; T v represents the trend component; S v represents seasonal component; R v Represents the residual component.

[0147] Optionally, the plurality of characteristic data sequences of the plurality of groups are respectively input into the intermediate layer of the target water quality information prediction model for processing to obtain the plurality of intermediate layer processing results, including:

[0148] Inputting the plurality of groups of trend component characteristic data sequences into the first processing module of the intermediate layer of the target water quality information prediction model for processing respectively, and obtaining the plurality of groups of first intermediate layer processing results;

[0149] Inputting the plurality of groups of seasonal component characteristic data sequences into the second processing module of the intermediate layer of the target water quality information prediction model for processing respectively, and obtaining the plurality of groups of second intermediate layer processing results;

[0150] The plurality of groups of residual component characteristic data sequences are respectively input into the third processing module of the intermediate layer of the target water quality information prediction model for processing to obtain a plurality of groups of third intermediate layer processing results.

[0151] Optionally, multiple groups of trend component characteristic data sequences are respectively input into the first processing module of the intermediate layer of the target water quality information prediction model for processing to obtain multiple groups of first intermediate layer processing results, including:

[0152] Inputting the plurality of groups of trend component characteristic data sequences into the convolutional neural network unit of the first processing module of the middle layer of the target water quality information prediction model respectively for convolution processing to obtain a plurality of groups of first convolution processing results;

[0153] The multiple groups of the first convolution processing results are respectively input into the recurrent neural network unit of the first processing module for processing to obtain multiple groups of first intermediate layer processing results.

[0154] Optionally, multiple groups of seasonal component characteristic data sequences are respectively input into the second processing module of the intermediate layer of the target water quality information prediction model for processing to obtain multiple groups of second intermediate layer processing results, including:

[0155] Inputting the plurality of groups of seasonal component characteristic data sequences into the convolutional neural network unit of the second processing module of the middle layer of the target water quality information prediction model respectively for convolution processing, and obtaining a plurality of groups of second convolution processing results;

[0156] Multiple groups of the second convolution processing results are respectively input into the recurrent neural network unit of the second processing module for processing to obtain multiple groups of second intermediate layer processing results.

[0157] Optionally, multiple groups of residual component feature data sequences are respectively input into the third processing module of the intermediate layer of the target water quality information prediction model for processing to obtain multiple groups of third intermediate layer processing results, including:

[0158] Inputting the plurality of groups of residual component feature data sequences into the convolutional neural network units of the third processing module of the middle layer of the target water quality information prediction model respectively for convolution processing, to obtain a plurality of groups of third convolution processing results;

[0159] The multiple groups of the third convolution processing results are respectively input into the recurrent neural network unit of the third processing module for processing to obtain multiple groups of third intermediate layer processing results.

[0160] Optionally, the plurality of groups of intermediate layer processing results are respectively input into the prediction layer of the target water quality information prediction model for water quality information prediction processing to obtain a plurality of groups of water quality information prediction results, including:

[0161] The multiple groups of intermediate layer processing results are respectively input into the shallow fully connected neural network units of the prediction layer of the target water quality information prediction model for feature fitting processing to obtain multiple groups of water quality information prediction results.

[0162] Optionally, determining a target water supply scheme according to multiple groups of water quality information prediction results and multiple water quality information preset ranges includes:

[0163] Comparing the water quality information prediction results of the plurality of groups with a plurality of preset ranges of water quality information to obtain comparison information;

[0164] A target water supply plan is determined according to the comparison information.

[0165] It should be noted that the device is a device corresponding to the above method, and all implementation methods in the above method embodiments are applicable to the embodiments of the device and can achieve the same technical effects.

[0166] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0167] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0168] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0169] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0170] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0171] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical disks.

[0172] In addition, it should be noted that in the apparatus and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. Moreover, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order, and some steps can be performed in parallel or independently of each other. For those of ordinary skill in the art, it is understood that all or any steps or components of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or a network of computing devices in hardware, firmware, software or a combination thereof, which can be achieved by those of ordinary skill in the art using their basic programming skills after reading the description of the present invention.

[0173] Therefore, the purpose of the present invention can also be achieved by running a program or a group of programs on any computing device. The computing device can be a well-known general device. Therefore, the purpose of the present invention can also be achieved by simply providing a program product containing a program code that implements the method or device. That is to say, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any well-known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. In addition, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order. Some steps can be performed in parallel or independently of each other.

[0174] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for determining a water supply plan, characterized in that: include: Obtain multiple groups of real-time water quality information data from multiple water supply sources; Preprocessing the multiple groups of real-time water quality information data to obtain multiple groups of target water quality information characteristic data; Inputting the target water quality information characteristic data of the plurality of groups into the decomposition layer of the target water quality information prediction model for decomposition processing respectively, to obtain a plurality of characteristic data sequences of the plurality of groups; Inputting the plurality of characteristic data sequences of the plurality of groups into the intermediate layer of the target water quality information prediction model for processing respectively, and obtaining the plurality of intermediate layer processing results; Inputting the processing results of the intermediate layers of the plurality of groups into the prediction layer of the target water quality information prediction model for water quality information prediction processing, thereby obtaining the water quality information prediction results of the plurality of groups; wherein the decomposition layer of the target water quality information prediction model is obtained by training according to the first preset network model, the intermediate layer is obtained by training according to the second preset network model and the third preset network model, and the prediction layer is obtained by training according to the fourth preset network model; A target water supply plan is determined based on multiple groups of water quality information prediction results and multiple water quality information preset ranges.

2. The method for determining a water supply plan according to claim 1, characterized in that: Preprocessing the multiple groups of real-time water quality information data to obtain multiple groups of target water quality information characteristic data, including: Cleaning the multiple groups of real-time water quality information data to obtain multiple groups of first processed data; Performing transformation processing on multiple groups of the first processed data to obtain multiple groups of second processed data; The plurality of groups of the second processed data are subjected to reduction processing to obtain a plurality of groups of the third processed data as the plurality of groups of target water quality information characteristic data.

3. The method for determining a water supply plan according to claim 1, characterized in that: Inputting the target water quality information characteristic data of the plurality of groups into the decomposition layer of the target water quality information prediction model for decomposition processing respectively, a plurality of characteristic data sequences of the plurality of groups are obtained, including: The target water quality information characteristic data of multiple groups are respectively input into the decomposition layer of the target water quality information prediction model, according to Y v =T v +S v +R v ,v=1,2,...,N for decomposition processing to obtain multiple groups of trend component characteristic data series, seasonal component characteristic data series and residual component characteristic data series; Among them, Y v Represents the target water quality information characteristic data at time v; T v represents the trend component; S v represents seasonal component; R v represents the residual component.

4. The method for determining a water supply plan according to claim 3, characterized in that: Inputting the plurality of characteristic data sequences of the plurality of groups into the intermediate layer of the target water quality information prediction model for processing respectively, and obtaining the plurality of intermediate layer processing results, including: Inputting the plurality of groups of trend component characteristic data sequences into the first processing module of the intermediate layer of the target water quality information prediction model for processing respectively, and obtaining the plurality of groups of first intermediate layer processing results; Inputting the plurality of groups of seasonal component characteristic data sequences into the second processing module of the intermediate layer of the target water quality information prediction model for processing respectively, and obtaining the plurality of groups of second intermediate layer processing results; The plurality of groups of residual component characteristic data sequences are respectively input into the third processing module of the intermediate layer of the target water quality information prediction model for processing to obtain a plurality of groups of third intermediate layer processing results.

5. The method for determining a water supply plan according to claim 4, characterized in that: Inputting multiple groups of trend component characteristic data sequences into the first processing module of the intermediate layer of the target water quality information prediction model for processing, respectively, to obtain multiple groups of first intermediate layer processing results, including: Inputting the plurality of groups of trend component characteristic data sequences into the convolutional neural network unit of the first processing module of the middle layer of the target water quality information prediction model respectively for convolution processing to obtain a plurality of groups of first convolution processing results; The multiple groups of the first convolution processing results are respectively input into the recurrent neural network unit of the first processing module for processing to obtain multiple groups of first intermediate layer processing results.

6. The method for determining a water supply plan according to claim 4, characterized in that: Inputting multiple groups of seasonal component characteristic data sequences into the second processing module of the intermediate layer of the target water quality information prediction model for processing, respectively, to obtain multiple groups of second intermediate layer processing results, including: Inputting the plurality of groups of seasonal component characteristic data sequences into the convolutional neural network unit of the second processing module of the middle layer of the target water quality information prediction model respectively for convolution processing, and obtaining a plurality of groups of second convolution processing results; Multiple groups of the second convolution processing results are respectively input into the recurrent neural network unit of the second processing module for processing to obtain multiple groups of second intermediate layer processing results.

7. The method for determining a water supply plan according to claim 4, characterized in that: Inputting the plurality of groups of residual component characteristic data sequences into the third processing module of the intermediate layer of the target water quality information prediction model for processing respectively, and obtaining the plurality of groups of third intermediate layer processing results, including: Inputting the plurality of groups of residual component feature data sequences into the convolutional neural network units of the third processing module of the middle layer of the target water quality information prediction model respectively for convolution processing, to obtain a plurality of groups of third convolution processing results; The multiple groups of the third convolution processing results are respectively input into the recurrent neural network unit of the third processing module for processing to obtain multiple groups of third intermediate layer processing results.

8. The method for determining a water supply plan according to claim 1, characterized in that: Inputting the plurality of groups of intermediate layer processing results into the prediction layer of the target water quality information prediction model for water quality information prediction processing, and obtaining the plurality of groups of water quality information prediction results, including: The multiple groups of intermediate layer processing results are respectively input into the shallow fully connected neural network units of the prediction layer of the target water quality information prediction model for feature fitting processing to obtain multiple groups of water quality information prediction results.

9. The method for determining a water supply plan according to claim 1, characterized in that: Determining a target water supply plan according to the plurality of groups of water quality information prediction results and a plurality of water quality information preset ranges includes: Comparing the water quality information prediction results of the plurality of groups with a plurality of preset ranges of water quality information to obtain comparison information; A target water supply plan is determined according to the comparison information.

10. A water supply scheme determination device, characterized in that: include: An acquisition module, used to acquire multiple groups of real-time water quality information data from multiple water supply sources; A processing module, used for preprocessing the multiple groups of real-time water quality information data to obtain multiple groups of target water quality information characteristic data; Inputting multiple groups of target water quality information feature data into the decomposition layer of the target water quality information prediction model for decomposition processing to obtain multiple groups of feature data sequences; inputting multiple groups of feature data sequences into the intermediate layer of the target water quality information prediction model for processing to obtain multiple groups of intermediate layer processing results; Inputting the processing results of the intermediate layers of the plurality of groups into the prediction layer of the target water quality information prediction model for water quality information prediction processing, thereby obtaining the water quality information prediction results of the plurality of groups; wherein the decomposition layer of the target water quality information prediction model is obtained by training according to the first preset network model, the intermediate layer is obtained by training according to the second preset network model and the third preset network model, and the prediction layer is obtained by training according to the fourth preset network model; A target water supply plan is determined based on multiple groups of water quality information prediction results and multiple water quality information preset ranges.

Images