A flood prevention intelligent control method, system and monitoring manhole cover thereof

Abstract

The present invention relates to the technical field of flood control and drainage, and discloses a flood control intelligent control method, system and monitoring manhole cover thereof, wherein the flood control intelligent control method comprises the following steps: collecting drainage network data; generating two-dimensional structure data from the collected drainage network data, the two-dimensional structure data comprising a data matrix and a relationship matrix; setting monitoring manhole covers for all manhole covers, generating two-dimensional structure data and inputting the data into a layout model, the layout model outputting a layout strategy; executing the layout strategy, collecting drainage network data, generating new two-dimensional structure data and inputting the data into a control model, the control model outputting a control strategy; the present invention collects drainage network data at regular intervals, fully considers the connection status between the manhole cover and the sewer pipe and considers the changes in the environment around the manhole cover. Compared with achieving flood control and drainage through a single manhole cover, the system can balance multiple flood control goals and improve the timeliness and safety of flood control and drainage.

Description

Technical Field

[0001] The present invention relates to the technical field of flood control and drainage, and more specifically, to a flood control intelligent control method, system and monitoring manhole cover thereof. Background Art

[0002] As an important part of the urban pipeline network, the urban drainage network not only plays a role in sewage discharge, but also has the function of flood prevention and drainage when urban waterlogging occurs. Using deep learning technology, the layout strategy of monitoring and controlling manhole covers is output according to information such as geographical location and altitude. According to the water level changes of the sewer network, the status of the manhole cover, rainfall, water depth and other information, the angle of the manhole cover is controlled to achieve intelligent flood prevention and drainage of the drainage network. However, the dynamic changes in the urban environment may affect the layout strategy and the control of the manhole cover angle. New buildings may change the surface runoff, and changes in the surrounding land type will also affect the hydrological response characteristics of the area, including the infiltration capacity of the accumulated water. Summary of the invention

[0003] The purpose of the present invention is to provide a flood prevention intelligent control method, system and monitoring manhole cover in order to solve the above problems.

[0004] The present invention provides a flood prevention intelligent control method, comprising the following steps:

[0005] Step 100, collecting drainage network data;

[0006] Step 200, generating two-dimensional structure data from the collected drainage network data, the two-dimensional structure data including a data matrix and a relationship matrix, one unit of the data matrix represents one-dimensional structure data of an independent object, the independent object includes a monitoring manhole cover, a control manhole cover and a sewer pipe, and one unit of the data matrix only contains the drainage network data of the independent object it represents;

[0007] The element in the xth row and yth column of the relationship matrix represents the association between the xth unit and the independent object represented by the yth unit of the data matrix. If there is an association, the value of this element of the relationship matrix is ​​1, otherwise it is 0;

[0008] The one-dimensional structured data includes n data items sorted by time, and the tth data item represents the independent object data collected at the tth moment;

[0009] Step 300, all manhole covers are set as monitoring manhole covers, two-dimensional structure data is generated and input into a layout model, the layout model includes a first intermediate layer, a second intermediate layer and a first output layer;

[0010] Input the one-dimensional structure data into the first intermediate layer, output the first intermediate representation data to the second intermediate layer, the second intermediate layer further inputs the relationship matrix, the second intermediate layer outputs the second intermediate representation data to the first output layer, the first output layer outputs a layout strategy, the layout strategy includes the manhole cover types of all manhole covers, and the manhole cover types include control manhole covers and monitoring manhole covers;

[0011] Step 400, executing the deployment strategy.

[0012] Furthermore, the drainage network data is collected once at each time sampling point, including monitoring manhole cover data, control manhole cover data and sewer pipe data. The monitoring manhole cover data includes the monitoring manhole cover geographical location, altitude, rainfall, water accumulation depth and the monitoring manhole cover surrounding environment information;

[0013] The control manhole cover data includes the control manhole cover's geographical location, altitude, rainfall, water depth, and the surrounding environment information of the control manhole cover;

[0014] The sewer data includes the water level ratio of the sewer, the starting and ending point coordinates, and the vertex coordinates. The water level ratio is the ratio of the water level of the sewer at the current moment to the designed maximum water level.

[0015] Furthermore, the time sampling point refers to the moment when data is collected at a predetermined time interval, and the default value of the predetermined time interval is 15 minutes.

[0016] Furthermore, the monitoring manhole cover and the sewer pipe are associated with each other, which means that: the monitoring manhole cover is connected to the sewer pipe;

[0017] The control manhole cover and the sewer pipe are associated with each other, which means: the control manhole cover is connected to the sewer pipe;

[0018] The association between a sewer pipe and a sewer pipe means that the sewer pipe and the sewer pipe are connected.

[0019] Furthermore, the formula for the first intermediate layer is as follows:

[0020] ;

[0021] in Represents the tth first intermediate representation data output by the first intermediate layer, represents the t-1th first intermediate representation data, Represents the tth data item of one-dimensional structured data, and are the first and second weight parameters, is the first bias parameter, tanh is the hyperbolic tangent function;

[0022] The formula for the second intermediate layer is as follows:

[0023] ;

[0024] in The second intermediate representation data representing the vth cell of the data matrix, is the set of cells of the data matrix that are associated with the vth cell of the data matrix, The nth first intermediate representation data output when the one-dimensional structure data of the vth unit of the data matrix is ​​input into the first intermediate layer, is a normalizing constant, represents the third weight parameter, represents the second bias parameter, is the sigmoid function;

[0025] The formula for the first output layer is as follows:

[0026] ;

[0027] in Represents a layout vector. The cth component value of the layout vector represents the probability value of the cth layout strategy. The layout strategy with the largest probability value is selected as the output. The layout strategy group contains all executable layout strategies. A layout strategy includes the manhole cover types of all manhole covers. The manhole cover types include control manhole covers and monitoring manhole covers. represents the set of all cells of the data matrix, is the fourth weight parameter, is the third bias parameter.

[0028] Further, the steps of arranging the training of the model include:

[0029] Step 301, initializing the parameters of the layout model;

[0030] Step 302: Observe the drainage network data at time a , Deployment strategy executed at all times , Drainage network data at all times , execute the placement strategy Rewards ;

[0031] ;

[0032] in, express The number of control manhole covers deployed at all times, Represents the total number of all manhole covers in the drainage network. and Indicates execution layout strategy and The flood control effect after the flood control is obtained by collecting data in the simulation environment. , , are the first, second and third weight coefficients, express Moment A suitability evaluation value for controlling the surrounding environment information of the manhole cover, The calculation formula is as follows:

[0033] ;

[0034] in, , , , , are the fourth, fifth, sixth, seventh and eighth weight coefficients, Indicates The proportion of impermeable materials around the manhole cover is controlled by Indicates The height above sea level of the control manhole cover, Indicates Control the building density around the manhole cover. Indicates Control the population density around the manhole cover. Indicates Control the flow of traffic around the manhole cover;

[0035] Step 303, then calculate the placement strategy error:

[0036] ;

[0037] in express The layout strategy error at the moment, represents the discount factor, , Represents layout model input The maximum probability value in the arrangement vector output when , Represents layout model input The placement vector outputted when the placement strategy The probability value of

[0038] Step 304: Update the layout model. The updated formula is as follows:

[0039] ;

[0040] , represents the step size of deep learning, Indicates delivery update;

[0041] Step 305, iterate steps 302-304 until the layout model converges or the number of iterations reaches a set value; the default value of this value is 30.

[0042] Further, collecting drainage network data, generating new two-dimensional structure data and inputting it into a control model, the control model comprising a third intermediate layer, a fourth intermediate layer and a second output layer;

[0043] The one-dimensional structure data in the newly generated two-dimensional structure data is input into the third intermediate layer, and the third intermediate representation data is output to the fourth intermediate layer. The fourth intermediate layer also inputs the relationship matrix, and the fourth intermediate layer outputs the fourth intermediate representation data to the second output layer. The second output layer outputs a control strategy, which includes all the angles of the control manhole covers. The formula of the third intermediate layer is as follows:

[0044] ;

[0045] ;

[0046] ;

[0047] ;

[0048] in, , , represents the fifth, sixth, and seventh weight parameters, , , represents the fourth, fifth and sixth bias parameters, represents the dot product, , and Respectively represent the first, second, and third intermediate states, Represents the tth data item of one-dimensional structured data, and They represent the t-th and t-1-th third intermediate representation data respectively, n≥t≥1, n represents the total number of data items of the one-dimensional structure data, and when t=1 ;

[0049] The formula for the fourth intermediate layer is as follows:

[0050] ;

[0051] in Represents the data matrix The fourth intermediate representation data of the units, Represents the data matrix The aggregation coefficient of each unit, Represents the first The set of cells of the data matrix with associated cells, represents the activation function, represents the eighth weight parameter, Represents the data matrix The output of the A third intermediate representation data;

[0052] ;

[0053] ;

[0054] ;

[0055] in Represents the data matrix The output of the A third intermediate representation data, and Represents the first The linear transformation features of the jth and jth units; represents the aggregation weight coefficient, represents the splicing weight coefficient, T represents transposition, represents an exponential function with a natural constant as the base, and LeakyRelu represents a modified linear unit function;

[0056] The formula for the second output layer is as follows:

[0057] ;

[0058] in Represents a control vector. The dth component value of the control vector represents the probability value of the dth strategy. The control strategy with the largest probability value is selected as the output. The control strategy group contains all executable control strategies. A control strategy includes all control manhole cover angles. represents the set of all cells of the data matrix, is the ninth weight parameter, is the seventh bias parameter.

[0059] Furthermore, the steps of controlling the training of the model include:

[0060] Step 401, initializing the parameters of the control model;

[0061] Step 402, observe Drainage network data at all times , Control strategies executed at all times , Drainage network data at all times , execute control strategy Rewards ;

[0062] ;

[0063] in, express The number of manhole covers under control at all times, Indicates Control the depth of water accumulation in the manhole cover. Indicates Control the rainfall around the manhole cover. Indicates The proportion of water level in the sewer pipe connected to the control manhole cover, Indicates Control the opening angle of the manhole cover. , , , , are the ninth, tenth, eleventh, twelfth and thirteenth weight coefficients;

[0064] Step 403, then calculate the control strategy error:

[0065] ;

[0066] in express The control strategy error at time represents the discount factor, , Represents the control model input The maximum probability value in the control vector output when Represents the control model input The control vector outputted at this time corresponds to the control strategy The probability value of

[0067] Step 404, update the control model, the updated formula is as follows:

[0068] ;

[0069] , represents the step size of deep learning, Indicates delivery update;

[0070] Step 405, iterate steps 402-404 until the control model converges or the number of iterations reaches a set value; the default value of this value is 30.

[0071] The present invention provides a flood control intelligent control system, which is used to store computer-readable instructions. When the computer-readable instructions are read, the aforementioned flood control intelligent control method can be executed.

[0072] The present invention provides a monitoring manhole cover, which is used for communication connection with a controller, and the controller is used for executing the aforementioned intelligent flood prevention control method.

[0073] The beneficial effects of the present invention are as follows: the present invention collects drainage network data at regular intervals, fully considers the connection status between the manhole cover and the sewer pipe, and considers changes in the environment around the manhole cover. Compared with achieving flood control and drainage through a single manhole cover, the system can balance multiple flood control goals and improve the timeliness and safety of flood control and drainage. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 It is a flow chart of a flood prevention intelligent control method of the present invention. DETAILED DESCRIPTION

[0075] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the contents of this specification. Each example may omit, replace or add various processes or components as needed. In addition, the features described relative to some examples may also be combined in other examples.

[0076] At least one embodiment of the present invention discloses a flood prevention intelligent control method, such as Figure 1 As shown, the following steps are included:

[0077] Step 100, collecting drainage network data, the drainage network data is collected once at each time sampling point, including monitoring manhole cover data, control manhole cover data and sewer pipe data;

[0078] The monitoring manhole cover data includes the monitoring manhole cover's geographical location, altitude, rainfall, water depth, and the surrounding environment information of the monitoring manhole cover;

[0079] The control manhole cover data includes the control manhole cover's geographical location, altitude, rainfall, water depth, and the surrounding environment information of the control manhole cover;

[0080] The sewer pipe data includes the water level ratio of the sewer pipe, the starting and ending point coordinates, and the vertex coordinates. The water level ratio is the ratio of the water level of the sewer pipe at the current moment to the designed maximum water level;

[0081] In one embodiment of the present invention, the entire drainage pipe network is regarded as a mesh structure composed of multiple sewer pipes, and each sewer pipe can be regarded as an independent object, and its boundary is the connection point of the sewer pipes;

[0082] In one embodiment of the present invention, the surrounding environment information of the manhole cover includes urban environment data within a predetermined range around the manhole cover, and the urban environment data includes building information (latitude and longitude coordinates, height and floor area), road information (latitude and longitude coordinates of the starting and ending points of the road), population density information, traffic flow information, ground material (binary coding according to permeability and impermeability, impermeable material coding is 0, including ordinary asphalt, ordinary concrete, marble, rubber floor tiles, and wooden boards; permeable material coding is 1, including porous asphalt, permeable concrete, grass bricks, permeable bricks, soil, and green space) and proportion information;

[0083] The time sampling point refers to the moment when data is collected at a predetermined time interval;

[0084] In one embodiment of the present invention, the default value of the predetermined time interval is 15 minutes, and the default value of the predetermined range is 50 meters by 50 meters.

[0085] Step 200, generating two-dimensional structure data from the collected drainage network data, the two-dimensional structure data including a data matrix and a relationship matrix, one unit of the data matrix represents one-dimensional structure data of an independent object, the independent objects include a monitoring manhole cover, a control manhole cover and a sewer pipe, one unit of the data matrix representing the monitoring manhole cover only contains the drainage network data of the monitoring manhole cover, one unit of the data matrix representing the control manhole cover only contains the drainage network data of the control manhole cover, and one unit of the data matrix representing the sewer pipe only contains the drainage network data of the sewer pipe;

[0086] The element in the xth row and yth column of the relationship matrix represents the association between the xth unit and the independent object represented by the yth unit of the data matrix. If there is an association, the value of this element of the relationship matrix is ​​1, otherwise it is 0;

[0087] The monitoring manhole cover and the sewer pipe are associated with each other, which means: the monitoring manhole cover is connected to the sewer pipe;

[0088] The control manhole cover and the sewer pipe are associated with each other, which means: the control manhole cover is connected to the sewer pipe;

[0089] The existence of a connection between a sewer pipe and a sewer pipe means that: the sewer pipe is connected to the sewer pipe;

[0090] The one-dimensional structured data includes n data items sorted by time, and the tth data item represents the independent object data collected at the tth moment;

[0091] Step 300, all manhole covers are set as monitoring manhole covers, two-dimensional structure data is generated and input into a layout model, the layout model includes a first intermediate layer, a second intermediate layer and a first output layer;

[0092] Input the one-dimensional structure data into the first intermediate layer, output the first intermediate representation data to the second intermediate layer, the second intermediate layer further inputs the relationship matrix, the second intermediate layer outputs the second intermediate representation data to the first output layer, the first output layer outputs a layout strategy, the layout strategy includes the manhole cover types of all manhole covers, and the manhole cover types include control manhole covers and monitoring manhole covers;

[0093] In one embodiment of the present invention, the formula of the first intermediate layer is as follows:

[0094]

[0095] in Represents the tth first intermediate representation data output by the first intermediate layer, represents the t-1th first intermediate representation data, Represents the tth data item of one-dimensional structured data, and are the first and second weight parameters, is the first bias parameter and tanh is the hyperbolic tangent function.

[0096] In one embodiment of the present invention, the formula of the second intermediate layer is as follows:

[0097]

[0098] in The second intermediate representation data representing the vth cell of the data matrix, is the set of cells of the data matrix that are associated with the vth cell of the data matrix, The nth first intermediate representation data output when the one-dimensional structure data of the vth unit of the data matrix is ​​input into the first intermediate layer, is a normalizing constant, represents the third weight parameter, represents the second bias parameter, is the sigmoid function.

[0099] In one embodiment of the present invention, the formula of the first output layer is as follows:

[0100]

[0101] in Represents a layout vector. The cth component value of the layout vector represents the probability value of the cth layout strategy. The layout strategy with the largest probability value is selected as the output. The layout strategy group contains all executable layout strategies. A layout strategy includes the manhole cover types of all manhole covers. The manhole cover types include control manhole covers and monitoring manhole covers. represents the set of all cells of the data matrix, is the fourth weight parameter, is the third bias parameter.

[0102] In one embodiment of the present invention, the step of arranging the training of the model includes:

[0103] Step 301, initializing the parameters of the layout model;

[0104] Step 302: Observe the drainage network data at time a , Deployment strategy executed at all times , Drainage network data at all times , execute the placement strategy Rewards ;

[0105]

[0106] in, express The number of control manhole covers deployed at all times, Represents the total number of all manhole covers in the drainage network. and Indicates execution layout strategy and The flood control effect after the flood control is obtained by collecting data in the simulation environment. , , are the first, second and third weight coefficients, express Moment A suitability evaluation value for controlling the surrounding environment information of the manhole cover, The calculation formula is as follows:

[0107]

[0108] in, , , , , are the fourth, fifth, sixth, seventh and eighth weight coefficients, Indicates The proportion of impermeable materials around the manhole cover is controlled by Indicates The height above sea level of the control manhole cover, Indicates Control the building density around the manhole cover. Indicates Control the population density around the manhole cover. Indicates Control the flow of traffic around the manhole cover;

[0109] Step 303, then calculate the placement strategy error:

[0110]

[0111] in express The layout strategy error at the moment, represents the discount factor, , Represents layout model input The maximum probability value in the arrangement vector output when , Represents layout model input The placement vector outputted when the placement strategy The probability value of

[0112] Step 304: Update the layout model. The updated formula is as follows:

[0113]

[0114] , represents the step size of deep learning, Indicates delivery update;

[0115] Step 305, iterate steps 302-304 until the layout model converges or the number of iterations reaches a set value; the default value of this value is 30.

[0116] In one embodiment of the present invention, a complete drainage network system is simulated using physical simulation software according to current drainage network data, a layout strategy is implemented in the drainage network system, and then the control strategy is executed through the output of a trained control model, and the rate of change of water depth is collected as flood control effect data.

[0117] Step 400, executing the layout strategy, collecting drainage network data, generating new two-dimensional structure data and inputting it into the control model, the control model including the third intermediate layer, the fourth intermediate layer and the second output layer;

[0118] Input one-dimensional structure data in the newly generated two-dimensional structure data into the third intermediate layer, output the third intermediate representation data to the fourth intermediate layer, the fourth intermediate layer also inputs the relationship matrix, the fourth intermediate layer outputs the fourth intermediate representation data to the second output layer, and the second output layer outputs a control strategy, which includes all control angles of the manhole cover;

[0119] In one embodiment of the present invention, the formula of the third intermediate layer is as follows:

[0120]

[0121]

[0122]

[0123]

[0124] in, , , represents the fifth, sixth, and seventh weight parameters, , , represents the fourth, fifth and sixth bias parameters, represents the dot product, , and Respectively represent the first, second, and third intermediate states, Represents the tth data item of one-dimensional structured data, and They represent the t-th and t-1-th third intermediate representation data respectively, n≥t≥1, n represents the total number of data items of the one-dimensional structure data, and when t=1 .

[0125] In one embodiment of the present invention, the formula of the fourth intermediate layer is as follows:

[0126] in Represents the data matrix The fourth intermediate representation data of the units, Represents the data matrix The aggregation coefficient of each unit, Represents the first The set of cells of the data matrix with associated cells, represents the activation function, represents the eighth weight parameter, Represents the data matrix The output of the A third intermediate representation data;

[0127]

[0128]

[0129]

[0130] in Represents the data matrix The output of the A third intermediate representation data, and Represents the first The linear transformation features of the jth and jth units; represents the aggregation weight coefficient, represents the splicing weight coefficient, T represents transposition, represents an exponential function with a natural constant as the base, and LeakyRelu represents the rectified linear unit function.

[0131] In one embodiment of the present invention, the formula of the second output layer is as follows:

[0132]

[0133] in Represents a control vector. The dth component value of the control vector represents the probability value of the dth strategy. The control strategy with the largest probability value is selected as the output. The control strategy group contains all executable control strategies. A control strategy includes all control manhole cover angles. represents the set of all cells of the data matrix, is the ninth weight parameter, is the seventh bias parameter.

[0134] In one embodiment of the present invention, the step of controlling the training of the model includes:

[0135] Step 401, initializing the parameters of the control model;

[0136] Step 402, observe Drainage network data at all times , Control strategies executed at all times , Drainage network data at all times , execute control strategy Rewards ;

[0137]

[0138] in, express The number of manhole covers under control at all times, Indicates Control the depth of water accumulation in the manhole cover. Indicates Control the rainfall around the manhole cover. Indicates The proportion of water level in the sewer pipe connected to the control manhole cover, Indicates Control the opening angle of the manhole cover. , , , , are the ninth, tenth, eleventh, twelfth and thirteenth weight coefficients;

[0139] Step 403, then calculate the control strategy error:

[0140]

[0141] in express The control strategy error at time represents the discount factor, , Represents the control model input The maximum probability value in the control vector output when Represents the control model input The control vector outputted at this time corresponds to the control strategy The probability value of

[0142] Step 404, update the control model, the updated formula is as follows:

[0143]

[0144] , represents the step size of deep learning, Indicates delivery update;

[0145] Step 405, iterate steps 402-404 until the control model converges or the number of iterations reaches a set value; the default value of this value is 30.

[0146] In at least one embodiment of the present invention, a flood control intelligent control system is provided, which is used to store computer-readable instructions, and when the computer-readable instructions are read, the aforementioned flood control intelligent control method can be executed.

[0147] In at least one embodiment of the present invention, a monitoring manhole cover is provided, which is used to communicate with a controller, and the controller is used to execute the aforementioned intelligent flood control method.

[0148] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation mode. The above-mentioned specific implementation mode is merely illustrative and not restrictive. Under the guidance of this embodiment, ordinary technicians in this field can also make more forms of equivalent embodiments, all of which are within the protection of this embodiment.

Claims

1. A flood control intelligent control method, characterized in that: include: Collect drainage network data; Generate two-dimensional structure data from the collected drainage network data. The two-dimensional structure data includes a data matrix and a relationship matrix. One unit of the data matrix represents one-dimensional structure data of an independent object. The independent object includes a monitoring manhole cover, a control manhole cover and a sewer pipe. It means that one unit of the data matrix only contains the drainage network data of the independent object it represents. The element in the xth row and yth column of the relationship matrix represents the association between the xth unit and the independent object represented by the yth unit of the data matrix. If there is an association, the value of this element of the relationship matrix is ​​1, otherwise it is 0; The one-dimensional structured data includes n data items sorted by time, and the tth data item represents the independent object data collected at the tth moment; All manhole covers are set as monitoring manhole covers, two-dimensional structure data is generated and input into a layout model, the layout model includes a first intermediate layer, a second intermediate layer and a first output layer; Input the one-dimensional structure data into the first intermediate layer, output the first intermediate representation data to the second intermediate layer, the second intermediate layer further inputs the relationship matrix, the second intermediate layer outputs the second intermediate representation data to the first output layer, the first output layer outputs a layout strategy, the layout strategy includes the manhole cover types of all manhole covers, and the manhole cover types include control manhole covers and monitoring manhole covers; The formula for the first intermediate layer is as follows: ； in Represents the tth first intermediate representation data output by the first intermediate layer, represents the t-1th first intermediate representation data, Represents the tth data item of one-dimensional structured data, and are the first and second weight parameters, is the first bias parameter, tanh is the hyperbolic tangent function; The formula for the second intermediate layer is as follows: ； in The second intermediate representation data representing the vth cell of the data matrix, is the set of cells of the data matrix that are associated with the vth cell of the data matrix, The nth first intermediate representation data output when the one-dimensional structure data of the vth unit of the data matrix is ​​input into the first intermediate layer, is a normalizing constant, represents the third weight parameter, represents the second bias parameter, is the sigmoid function; The formula for the first output layer is as follows: ； in Represents a layout vector. The cth component value of the layout vector represents the probability value of the cth layout strategy. The layout strategy with the largest probability value is selected as the output. The layout strategy group contains all executable layout strategies. A layout strategy includes the manhole cover types of all manhole covers. The manhole cover types include control manhole covers and monitoring manhole covers. represents the set of all cells of the data matrix, is the fourth weight parameter, is the third bias parameter; Execute the placement strategy.

2. A flood control intelligent control method according to claim 1, characterized in that: The drainage network data is collected once at each time sampling point, including monitoring manhole cover data, control manhole cover data and sewer pipe data. The monitoring manhole cover data includes the monitoring manhole cover geographical location, altitude, rainfall, water accumulation depth and the surrounding environment information of the monitoring manhole cover; The control manhole cover data includes the control manhole cover's geographical location, altitude, rainfall, water depth, and the surrounding environment information of the control manhole cover; The sewer data includes the water level ratio of the sewer, the starting and ending point coordinates, and the vertex coordinates. The water level ratio is the ratio of the water level of the sewer at the current moment to the designed maximum water level.

3. A flood control intelligent control method according to claim 2, characterized in that: The time sampling point refers to the moment when data is collected at a predetermined time interval. The default value of the predetermined time interval is 15 minutes.

4. The intelligent flood control method according to claim 1, characterized in that: The monitoring manhole cover and the sewer pipe are associated with each other, which means: the monitoring manhole cover is connected to the sewer pipe; The control manhole cover and the sewer pipe are associated with each other, which means: the control manhole cover is connected to the sewer pipe; The association between a sewer pipe and a sewer pipe means that the sewer pipe and the sewer pipe are connected.

5. The intelligent flood control method according to claim 1, characterized in that: The steps to arrange the training of the model include: Step 301, initializing the parameters of the layout model; Step 302: Observe the drainage network data at time a , Deployment strategy executed at all times , Drainage network data at all times , execute the placement strategy Rewards ; ； in, express The number of control manhole covers deployed at all times, Represents the total number of all manhole covers in the drainage network. and Indicates execution layout strategy and The flood control effect after the flood control is obtained by collecting data in the simulation environment. , , are the first, second and third weight coefficients, express Moment A suitability evaluation value for controlling the surrounding environment information of the manhole cover, The calculation formula is as follows: ； in, , , , , are the fourth, fifth, sixth, seventh and eighth weight coefficients, Indicates The proportion of impermeable materials around the manhole cover is controlled by Indicates The height above sea level of the control manhole cover, Indicates Control the building density around the manhole cover. Indicates Control the population density around the manhole cover. Indicates Control the flow of traffic around the manhole cover; Step 303, then calculate the placement strategy error: ； in express The layout strategy error at the moment, represents the discount factor, , Represents layout model input The maximum probability value in the arrangement vector output when , Represents layout model input The placement vector outputted when the placement strategy The probability value of Step 304: Update the layout model. The updated formula is as follows: ； , represents the step size of deep learning, Indicates delivery update; Step 305, iterate steps 302-304 until the layout model converges or the number of iterations reaches a set value; the default value of this value is 30.

6. The intelligent flood control method according to claim 1, characterized in that: Collect drainage network data, generate new two-dimensional structure data and input into the control model, the control model includes a third intermediate layer, a fourth intermediate layer and a second output layer; The one-dimensional structure data in the newly generated two-dimensional structure data is input into the third intermediate layer, and the third intermediate representation data is output to the fourth intermediate layer. The fourth intermediate layer also inputs the relationship matrix, and the fourth intermediate layer outputs the fourth intermediate representation data to the second output layer. The second output layer outputs a control strategy, which includes all the angles of the control manhole covers. The formula of the third intermediate layer is as follows: ； ； ； ； in, , , represents the fifth, sixth, and seventh weight parameters, , , represents the fourth, fifth and sixth bias parameters, represents the dot product, , and Respectively represent the first, second, and third intermediate states, Represents the tth data item of one-dimensional structured data, and They represent the t-th and t-1-th third intermediate representation data respectively, n≥t≥1, n represents the total number of data items of the one-dimensional structure data, and when t=1 ; The formula for the fourth intermediate layer is as follows: ； in Represents the data matrix The fourth intermediate representation data of the units, Represents the data matrix The aggregation coefficient of each unit, Represents the first The set of cells of the data matrix with associated cells, represents the activation function, represents the eighth weight parameter, Represents the data matrix The output of the A third intermediate representation data; ； ； ； in Represents the data matrix The output of the A third intermediate representation data, and Represents the first The linear transformation features of the jth and jth units; represents the aggregation weight coefficient, represents the splicing weight coefficient, T represents transposition, represents an exponential function with a natural constant as the base, and LeakyRelu represents a modified linear unit function; The formula for the second output layer is as follows: ； in Represents a control vector. The dth component value of the control vector represents the probability value of the dth strategy. The control strategy with the largest probability value is selected as the output. The control strategy group contains all executable control strategies. A control strategy includes all control manhole cover angles. represents the set of all cells of the data matrix, is the ninth weight parameter, is the seventh bias parameter.

7. A flood prevention intelligent control method according to claim 6, characterized in that: The steps to control model training include: Step 401, initializing the parameters of the control model; Step 402, observe Drainage network data at the moment , Control strategies executed at all times , Drainage network data at the moment , execute control strategy Rewards ; ； in, express The number of manhole covers under control at all times, Indicates Control the depth of water accumulation in the manhole cover. Indicates Control the rainfall around the manhole cover. Indicates The proportion of water level in the sewer pipe connected to the control manhole cover, Indicates Control the opening angle of the manhole cover. , , , , are the ninth, tenth, eleventh, twelfth and thirteenth weight coefficients; Step 403, then calculate the control strategy error: ； in express The control strategy error at time represents the discount factor, , Represents the control model input The maximum probability value in the control vector output when Represents the control model input The control vector outputted at this time corresponds to the control strategy The probability value of Step 404, update the control model, the updated formula is as follows: ； , represents the step size of deep learning, Indicates delivery update; Step 405, iterate steps 402-404 until the control model converges or the number of iterations reaches a set value; the default value of this value is 30.

8. A flood control intelligent control system, characterized in that: It is used to store computer-readable instructions, and when the computer-readable instructions are read, a flood prevention intelligent control method as described in any one of claims 1-7 can be executed.

9. A monitoring manhole cover, characterized in that: It is used for communication connection with a controller, and the controller is used for executing a flood prevention intelligent control method as described in any one of claims 1-7.

Images