CN114429034A - Pressure monitoring point moving arrangement method for water supply network hydraulic model water quantity checking - Google Patents

Abstract

The invention discloses a pressure monitoring point moving arrangement method for water supply network hydraulic model water quantity checking, which comprises the following steps: dividing a checking period according to the total number of the nodes of the pipe network and the number of the pressure monitoring sensors; obtaining a Jacobian matrix of node pressure relative to node water quantity based on an initialized pipe network hydraulic model; solving a monitoring point moving arrangement scheme according to a Jacobian matrix and an improved implicit enumeration optimization method; and checking and calculating the node water quantity parameters of the water supply network hydraulic model according to the monitoring data acquired by the monitoring point position deployment scheme in each checking period, and improving the calculation precision. On the premise of equivalent cost, the invention can multiply the acquisition amount of monitoring data, increase the pressure monitoring space density of the pipe network, acquire more pipe network running state information and is beneficial to improving the model checking precision; on the premise that the model checking precision is equivalent, the method can greatly reduce the hardware cost, the construction cost and the maintenance cost of model parameter checking.

Description

A method for moving and arranging pressure monitoring points for water quantity check of hydraulic model of water supply network

technical field

The invention relates to the technical field of modeling of urban water supply systems, in particular to a method for moving and arranging pressure monitoring points for checking the water quantity of a hydraulic model of a water supply pipe network.

Background technique

The urban water supply pipe network system is the lifeline of a city and an essential and important municipal infrastructure in a modern city. In recent years, with the rapid development of urbanization, the scale of the urban water supply pipe network system has continued to expand, and the complexity has gradually increased. In addition, the aging of the pipe network and the high leakage rate of the urban water supply pipe network are difficult to reduce. Therefore, the effective management of the pipe network The requirements and technical difficulty of operation and maintenance are getting higher and higher. The use of modern digital information means to efficiently manage urban water affairs has become the consensus of the water industry at home and abroad. The establishment of hydraulic model of water supply pipe network is the core component of information management, and it is one of the core technical means to realize rational planning and transformation of pipe network, system condition diagnosis and real-time operation optimization.

Under the current strategic background of the rapid development of "smart cities" and "smart water affairs" labelled by Internet+, Internet of Things, big data, artificial intelligence, etc., the modeling of water supply pipeline networks has been paid more and more attention. Some domestic cities such as Beijing and Shanghai , Guangzhou, Foshan and other water supply companies have successively established hydraulic models of water supply network. The pipeline network model is a simulation model that can accurately represent the actual system under a specific modeling purpose. At present, the hydraulic model of the pipeline network has played an important role in the planning and design of the pipeline network, and the reconstruction and expansion of the pipeline network. However, due to the limitation of the accuracy of the pipe network model, the in-depth application of the pipe network model has encountered related technical bottlenecks in recent years. For example, in-depth applications such as pipeline leakage identification and control, pipeline network abnormal state diagnosis, pipeline network water quality simulation, and energy-saving optimization based on hydraulic model checking are all due to the accuracy of the pipeline network model, resulting in unsatisfactory application results.

In the process of hydraulic modeling of the water supply network, the monitoring data such as the pressure and flow of the pipe network are often used to check the pipe network model to improve the accuracy of the pipe network model. Generally speaking, for a water supply network system, the more monitoring points are arranged, the higher the calibration accuracy of the model will be. However, limited by economic investment, the number of monitoring equipment is relatively limited. Some scholars or engineers have proposed various methods for optimizing the layout of monitoring points (eg, CN105894130B, CN109930658B) to optimize the location of monitoring equipment monitoring points; however, limited by the relatively limited number of monitoring equipment, even if monitoring points are selected In the case of site optimization and fixed monitoring, the spatial monitoring density of the entire water supply pipe network is still low, and the lack of real-time operating status monitoring information of the pipe network is still the key bottleneck that makes it difficult to improve the accuracy of the hydraulic model of the pipe network. Genetic and BP neural network, particle swarm algorithm (PSO), sparrow search algorithm and other random search algorithms are often used for checking (CN108898512A, CN112149358A, CN112163301A, CN112733443A, etc.), but this kind of random algorithm has low computational efficiency and is difficult to apply to Large-scale pipeline network and application scenarios of real-time demand. To sum up, in the face of the bottleneck problem that it is difficult to improve the checking accuracy of the hydraulic model of the pipeline network due to relatively few monitoring equipment and insufficient monitoring information, the present invention proposes an innovative idea of mobile monitoring, aiming to perform rotation through mobile equipment. Monitoring, increase the amount of monitoring data several times, and build corresponding model checking methods, effectively break through the technical bottleneck, and promote the development of pipeline network model checking theory and technology.

SUMMARY OF THE INVENTION

In view of the above problems existing in the prior art, the present application provides a method for moving and arranging pressure monitoring points for checking the water quantity of a hydraulic model of a water supply pipe network. A sufficient number of pressure monitoring data of the pipe network can be obtained through a limited number of pressure monitoring sensors. , so as to improve the model calibration accuracy.

The technical scheme of the present invention is as follows:

A method for moving and arranging pressure monitoring points for checking the water quantity of a hydraulic model of a water supply pipe network, comprising the following steps:

S1, formulate a check plan, which is divided into round_num check cycles, the number of monitoring points in each check cycle is sens_num, and the length of the check cycle is t; the monitoring points are the installation positions of the pressure monitoring sensors;

The total number of monitoring points in the verification plan is:

length(sens)=sens_num×round_num

The total length of the check plan is:

T=t×round_num

S2. Based on the initialized hydraulic model of the pipe network, the Jacobian matrix HQ of the node pressure and the node water volume is obtained;

S3. According to the Jacobian matrix HQ and the improved implicit enumeration optimization method, solve the monitoring point moving scheme sens_place; the monitoring point moving scheme sens_place refers to: when changing from the current check cycle to the next check cycle, the monitoring point needs to be changed from Which position to change the current position to;

The mathematical description method of the monitoring point movement scheme sens_place is to define the positions of different monitoring points in each check cycle as a vector, that is, the monitoring point position deployment scheme sens_place[k], where k represents the kth check cycle; The sum of the monitoring point position deployment schemes sens_place[k] of all verification periods is the monitoring point movement scheme sens_place;

S4. According to the time sequence of the verification plan, select the monitoring point location deployment scheme sens_place[1] in the first verification cycle, and calculate the node water volume of the hydraulic model of the water supply network through an iterative method;

S5. Calculate the optimal parameter estimation value of the hydraulic model of the water supply pipe network in the checking period, and set the calculation result as the parameter of the hydraulic model of the water supply pipe network;

S6. Select the monitoring point location deployment scheme sens_place[2] in the second check cycle, repeat S4 and S5, and gradually improve the calculation accuracy of the hydraulic model of the water supply network by adjusting the parameters until all check cycles are processed. .

Further, step S2 includes the following steps:

S21. Set the node association matrix A of the water supply network:

S22. Use the Haicheng-Wilhelm equation to calculate the partial differential of the head loss to the pipe section:

Among them: h is the head loss, _Ku is the unit conversion factor, d, L, q and c are the pipe diameter (mm), pipe length (m), water volume (L/s) and Haicheng-Williams coefficient;

S23. Write the partial differential of the head loss to the pipe section in the form of a diagonal matrix:

S24. Calculate the Jacobian matrix HQ of the nodal pressure with respect to the nodal water volume according to the following formula:

HQ=-(ABA ^T ) ^-1 .

Further, step S3 includes the following steps:

S31. According to the Jacobian matrix HQ, construct the objective function:

Among them: n is the total number of nodes in the water supply network, and sens is the index of the monitoring point location;

S32, use the improved implicit enumeration optimization method to solve the objective function f(sens), and obtain a vector, that is, the monitoring point position general arrangement sensall_vector; the monitoring point position general arrangement sensall_vector describes the node index set of all monitoring points;

S33. Use the kmeans++ method to cluster all the monitoring points in the sensall_vector according to the spatial position, and the number of classification groups cluster_num is equal to the number of monitoring points sens_num;

S34. Sort the monitoring points in each classification group according to the 1-norm of the sensitivity vector from large to small to obtain a monitoring point position sequence; the sensitivity vector refers to a row in the Jacobian matrix HQ, in which the monitoring point The sensitivity vector of point i is the ith row of the Jacobian matrix HQ; the monitoring point position sequence represents the different positions of the same pressure monitoring sensor in each calibration cycle;

If the length of the monitoring point position sequence is less than the check period round_num, add monitoring points to the end of the sequence in a circular manner until the length of the sequence is equal to the check period round_num;

S35. For the first check cycle, select the first data from the monitoring point position sequence of each classification group in turn, and obtain a vector whose dimension is equal to the number of monitoring points sens_num, which is the first check cycle The monitoring point position deployment scheme sens_place[1]; for the second check cycle, the second data is selected from the monitoring point position sequence of each classification group in turn, and a vector whose dimension is equal to the number of monitoring points sens_num is also obtained, This vector is the monitoring point position deployment scheme sens_place[2] of the second check cycle; and so on, the monitoring point position deployment scheme of all check periods is obtained, which is the monitoring point movement scheme sens_place.

Further, step S4 includes the following steps:

S41. Divide the current calibration cycle into M time periods, marked as: period t ₁ , period t ₂ , . . . , period t _M ;

S42, set a vector Q[t] for describing the water quantity arrangement of all nodes; the dimension of the vector Q[t] is equal to the total number of nodes of the model, and the value of each dimension represents the water quantity of the corresponding node respectively;

S43. For the time period t ₁ , set the initial value Q[t ₁ ] of Q[t] as the average value of the total water volume of the model:

Q[t ₁ ]=ones(1,n)*Q _avg

Q _avg =Q _total /n

Where: Q _total is the total water volume of the model (L/s), Q _avg is the average water volume of the model nodes (L/s), and ones(1,n) is a row vector whose length is the total number of nodes n and the elements are all 1;

S44. Assuming that it is currently the k-th calibration cycle, read the monitoring point position deployment scheme sens_place[k], and set the pressure error amount to be eliminated by the target iteration:

dH=H _o [sens]-H _s [sens]

Where: H _o is the pressure (m) measured by the pressure monitoring sensor, and H _s is the pressure (m) after each iteration correction;

S45, solve the Jacobian matrix HQ of the pipe network model when the water quantity is arranged as Q[t ₁ ];

S46, solve the equation HQ×dQ=dH, and obtain the correction amount dQ of Q[t ₁ ];

S47. Calculate the water quantity arrangement Q[t ₁ ] _n+1 of the next iteration by using the formula Q[t ₁ ] _n+1 =Q[t ₁ ] _n +dQ;

S48, the iteration stops when dH is less than the set allowable error threshold, and the result obtained at this time is the final result of the water quantity arrangement Q[t ₁ ] in the time period t ₁ ;

S49, take Q[t ₁ ] as the initial value, calculate the water quantity arrangement Q[t ₂ ] of the time period t ₂ ; and so on, repeat S45 to S48 until the water quantity arrangement Q[t] of all time periods is calculated;

S410 , arranging all the obtained water consumption in a single period in time sequence to obtain a node water consumption pattern; the node water consumption pattern refers to a coefficient of the water consumption of a water node in the pipe network model changing with time.

Further, the specific steps of improving the implicit enumeration optimization method described in step S32 are as follows:

S51. Set a vector sensall_init, which is the node index set of all monitoring points;

S52. Randomly arrange all nodes to obtain an in-point group N _in (i), and select an in-point in sequence;

S53. Randomly arrange all the initial solution vector elements to obtain an out point group N _out (j), and replace the out points with the in points in order to obtain a new solution sensall_new;

S54. Substitute the original solution sensall_init and the new solution sensall_new into the objective function of step S31, solve the original solution f(sensall_init) and the new solution f(sensall_new), and select a better solution to enter the next iteration;

S55. Repeat S52, S53, and S54 until a better solution cannot be generated, and terminate; the final solution obtained is the sensall_vector of the general arrangement of monitoring point positions.

Further, the specific method of step S44 is as follows:

Denote HQ as A, dH as b, and dQ as the unknown x to be solved, then the process of solving Ax=b by iterative method is:

b ₂ =bA·dx ₁

b ₃ =b ₂ -A·dx ₂

...

b _n+1 =b _n -A·dx _n

...

After the iteration is over, the number of iterations is N, then the value of x is:

Among them: sum(A, axis=0) represents the sum of the columns of the matrix A, sum(A, axis=1) represents the sum of the rows of the matrix A, and abs(A) represents the absolute value of the matrix A, where Multiplication and division are both multiplication and division of counterpoints.

The beneficial technical effects of the present invention are:

(1) Under the premise of the same cost, the present invention can expand the collection amount of monitoring data several times, increase the monitoring density of the pressure space of the pipe network, and obtain more information on the operation status of the pipe network, so as to improve the hydraulic model checking of the water supply pipe network. Accuracy provides a solid data foundation;

(2) On the premise that the model checking accuracy is comparable, the present invention can greatly reduce the hardware cost, construction cost and maintenance cost of model parameter checking; Reduce computing time, improve computing efficiency, and have better adaptability to the application scenarios of real-time verification;

(3) The present invention has the ability to accumulate and absorb the continuous accumulation of monitoring data, that is, with the accumulation of monitoring data in the time dimension and the space dimension, the calibration accuracy of the pipe network is gradually improved in the overall trend, approaching when all nodes are laid out Corresponding calibration accuracy when monitoring equipment is available.

The invention can be applied to model parameter checking of water supply pipe network, running state diagnosis, etc., and has wide application prospect.

Description of drawings

Accompanying drawing 1 is the pipe network topology diagram of the embodiment;

Accompanying drawing 2 is the schematic diagram of monitoring point movement scheme;

3 is a schematic diagram of the process of moving and arranging monitoring points of the pipeline network of the embodiment;

Accompanying drawing 4 is the variation curve of absolute pressure checking error of all nodes in the mobile monitoring process of the pipeline network of the embodiment;

Accompanying drawing 5 is the variation curve of relative calibration error of all-node flow in the mobile monitoring process of the pipeline network of the embodiment;

FIG. 6 is the variation curve of the relative calibration error of the flow rate of the whole pipe section during the mobile monitoring process of the pipe network of the embodiment.

Detailed ways

The present invention will be described in detail below with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are only some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

The embodiment is a water supply pipe network model of an engineering case, as shown in FIG. 1 . The pipe network system has 4 water supply sources, each with its own time pattern, the total number of water nodes is 484, the number of node water consumption patterns is 17, the nodes are divided into 17 groups and assigned random water consumption patterns, the baseline water consumption is randomly set, and the pipeline There are 567 pipes in total, and the roughness coefficient of the pipes is randomly set, ranging from 90 to 130.

The steps of parameter checking are as follows:

S1. Formulate a check plan, which is divided into round_num check cycles. The number of monitoring points in each check cycle is sens_num (the number of pressure monitoring sensors that can be moved), and the length of the check cycle is t; the monitoring point is the pressure Monitor where the sensor is installed.

For the example, round_num=20, sens_num=10, t=24 hours.

The total number of monitoring points in the verification plan is:

length(sens)=sens_num×round_num=10×20=200

The total length of the check plan is:

T=t×round_num=24×20=480 (hours)

S2. Based on the initialized hydraulic model of the pipe network, the Jacobian matrix HQ of the node pressure and the node water quantity is obtained. Specific steps are as follows:

(2-1) Set the node association matrix A of the water supply network:

(2-2) Use the Haicheng-Wilhelm equation to calculate the partial differential of the head loss to the pipe section:

Among them: h is the head loss, _Ku is the unit conversion factor, d, L, q and c are the pipe diameter (mm), pipe length (m), water volume (L/s) and Haicheng-Williams coefficient;

(2-3) Write the partial differential of the head loss to the pipe section in the form of a diagonal matrix:

(2-4) Calculate the Jacobian matrix HQ of the nodal pressure with respect to the nodal water volume according to the following formula:

HQ=-(ABA ^T ) ^-1

S3. According to the Jacobian matrix HQ and the improved implicit enumeration optimization method, solve the monitoring point movement scheme sens_place. The monitoring point moving scheme sens_place refers to: when changing from the current check cycle to the next check cycle, the monitoring point needs to be changed from the current position to which position, as shown in FIG. 2 .

The mathematical description method of the monitoring point movement scheme sens_place is to define the positions of different monitoring points in each check cycle as a vector, that is, the monitoring point position deployment scheme sens_place[k], where k represents the kth check cycle; The sum of the monitoring point position deployment schemes sens_place[k] of all verification periods is the monitoring point moving scheme sens_place.

Specific steps are as follows:

(3-1) According to the Jacobian matrix HQ, construct the objective function:

Among them: n is the total number of nodes in the water supply network, and sens is the index of the monitoring point location.

(3-2) Use the improved implicit enumeration optimization method to solve the objective function f(sens), and obtain a vector, that is, sensall_vector, which is the general arrangement of monitoring points. The monitoring point location general arrangement sensall_vector describes the node index set of all monitoring points. Specific steps are as follows:

(3-2-1) Set a vector sensall_init, which is the node index set of all monitoring points;

(3-2-2) Randomly arrange all nodes to obtain an in-point group N _in (i), and select an in-point in sequence;

(3-2-3) Randomly arrange all the initial solution solution vector elements to obtain an out point group N _out (j), and let the in points replace the out points in order to obtain a new solution sensall_new;

(3-2-4) Substitute the original solution sensall_init and the new solution sensall_new into the objective function of step S31, solve the original solution f(sensall_init) and the new solution f(sensall_new), and select a better solution to enter the next iteration;

(3-2-5) Repeat S52, S53, and S54 until a better solution cannot be generated, and terminate; the final solution obtained is the sensall_vector of the general arrangement of monitoring points.

(3-3) Use the kmeans++ method to cluster all the monitoring points in the sensall_vector according to the spatial position, and the number of classification groups cluster_num is equal to the number of monitoring points sens_num.

(3-4) Sort the monitoring points in each classification group according to the 1-norm of its sensitivity vector from large to small to obtain the monitoring point position sequence; the sensitivity vector here refers to the values in the Jacobian matrix HQ A certain row, in which the sensitivity vector of monitoring point i is the ith row of Jacobian matrix HQ; if the length of the monitoring point position sequence is less than the check period round_num, the monitoring points are added to the end of the sequence in a circular manner until the sequence is The length of is equal to the check cycle round_num; the monitoring point position sequence represents the different positions of the same pressure monitoring sensor in each check cycle.

(3-5) For the first check cycle, select the first data from the monitoring point position sequence of each classification group in turn, and obtain a vector whose dimension is equal to the number of monitoring points sens_num, which is the first data The monitoring point location deployment scheme sens_place[1] of the check cycle; for the second check cycle, select the second data from the monitoring point position sequence of each classification group in turn, and also obtain a dimension equal to the number of monitoring points sens_num , which is the monitoring point position deployment scheme sens_place[2] of the second check cycle; and so on, the monitoring point position deployment scheme of all check cycles is obtained, which is the monitoring point movement scheme sens_place. The monitoring point movement process of the embodiment is shown in FIG. 3 .

S4. According to the time sequence of the verification plan, select the monitoring point location deployment scheme sens_place[1] in the first verification cycle, and calculate the node water volume of the hydraulic model of the water supply network through an iterative method. The specific steps are as follows:

(4-1) Divide the current calibration cycle into M time periods, marked as: period t ₁ , period t ₂ , ..., period t _M ;

(4-2) A vector Q[t] is set to describe the water volume arrangement of all nodes; the dimension of the vector Q[t] is equal to the total number of nodes of the model, and the value of each dimension represents the water volume of the corresponding node respectively;

(4-3) For the period t ₁ , set the initial value Q[t ₁ ] of Q[t] as the average value of the total water volume of the model:

Q[t ₁ ]=ones(1,n)*Q _avg

Q _avg =Q _total /n

Where: Q _total is the total water volume of the model (L/s), Q _avg is the average water volume of the model nodes (L/s), and ones(1,n) is a row vector whose length is the total number of nodes n and the elements are all 1;

(4-4) Assuming that the current is the kth calibration cycle, read the monitoring point position deployment scheme sens_place[k], and set the pressure error amount to be eliminated by the target iteration:

dH=H _o [sens]-H _s [sens]

Where: H _o is the pressure (m) measured by the pressure monitoring sensor, and H _s is the pressure (m) after each iteration correction;

(4-5) Solve the Jacobian matrix HQ of the pipe network model when the water quantity is arranged as Q[t ₁ ];

(4-6) Solve the equation HQ×dQ=dH, and obtain the correction amount dQ of Q[t ₁ ];

(4-7) Calculate the water quantity arrangement Q[t ₁ ] _n+1 of the next iteration through the formula Q[t ₁ ] _n+1 =Q[t ₁ ] _n +dQ;

(4-8) The iteration stops when dH is less than the set allowable error threshold, and the result obtained at this time is the final result of the water quantity arrangement Q[t ₁ ] in the time period t ₁ ;

(4-9) Taking Q[t ₁ ] as the initial value, calculate the water quantity arrangement Q[t ₂ ] of the time period t ₂ ; and so on, repeat S45 to S48 until the water quantity arrangement Q[t] of all time periods is calculated ;

(4-10) The node water consumption pattern can be obtained by connecting the water consumption of all periods.

S5. Calculate the optimal parameter estimation value of the hydraulic model of the water supply pipe network in the checking period, and set the calculation result as the parameter of the hydraulic model of the water supply pipe network;

S6. Select the monitoring point location deployment scheme sens_place[2] in the second check cycle, repeat S4 and S5, and gradually improve the calculation accuracy of the hydraulic model of the water supply network by adjusting the parameters until all check cycles are processed. .

The parameter calibration results are evaluated and calculated using the following indicators:

Nodal pressure absolute error:

Node flow relative error:

Relative error of pipe flow:

Among them, H _true , Q _true , and q _true are respectively the real node pressure, node flow, and pipeline flow value set by the case pipe network, and H _s , Q _s , and q _s are the simulated node pressure after the case pipe network is checked, Node flow, pipeline flow value.

The results of checking and evaluating the parameters of the mobile monitoring process are shown in Figures 4-6.

Although the embodiment of the present invention has been disclosed as above, it is not limited to the application listed in the description and the embodiment, and it can be applied to various fields suitable for the present invention. For those of ordinary skill in the art, various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principle and spirit of the present invention, and therefore without departing from the general concepts defined by the claims and equivalent scopes Below, the invention is not limited to the specific details.

Claims (6)

1. a pressure monitoring point moving arrangement method for water supply pipe network hydraulic model water quantity check, is characterized in that, comprises the following steps: S1, formulate a check plan, which is divided into round_num check cycles, the number of monitoring points in each check cycle is sens_num, and the length of the check cycle is t; the monitoring points are the installation positions of the pressure monitoring sensors; The total number of monitoring points in the verification plan is: length(sens)=sens_num×round_num The total length of the check plan is: T=t×round_num S2. Based on the initialized hydraulic model of the pipe network, the Jacobian matrix HQ of the node pressure and the node water volume is obtained; S3. According to the Jacobian matrix HQ and the improved implicit enumeration optimization method, solve the monitoring point moving scheme sens_place; the monitoring point moving scheme sens_place refers to: when changing from the current check cycle to the next check cycle, the monitoring point needs to be changed from Which position to change the current position to; The mathematical description method of the monitoring point movement scheme sens_place is to define the positions of different monitoring points in each check cycle as a vector, that is, the monitoring point position deployment scheme sens_place[k], where k represents the kth check cycle; The sum of the monitoring point position deployment schemes sens_place[k] of all verification periods is the monitoring point moving scheme sens_place; S4. According to the time sequence of the verification plan, select the monitoring point location deployment scheme sens_place[1] in the first verification cycle, and calculate the node water volume of the hydraulic model of the water supply network through an iterative method; S5. Calculate the optimal parameter estimation value of the hydraulic model of the water supply pipe network in the checking period, and set the calculation result as the parameter of the hydraulic model of the water supply pipe network; S6. Select the monitoring point location deployment scheme sens_place[2] in the second check cycle, repeat S4 and S5, and gradually improve the calculation accuracy of the hydraulic model of the water supply network by adjusting the parameters until all check cycles are processed. . 2. A method for moving and arranging pressure monitoring points for water supply network hydraulic model water quantity check according to claim 1, wherein step S2 comprises the following steps: S21. Set the node association matrix A of the water supply network:

S22. Use the Haicheng-Wilhelm equation to calculate the partial differential of the head loss to the pipe section:

Where: h is the head loss, _Ku is the unit conversion factor, d, L, q and c are the pipe diameter (mm), pipe length (m), water volume (L/s) and Haicheng-Williams coefficient; S23. Write the partial differential of the head loss to the pipe section in the form of a diagonal matrix:

S24. Calculate the Jacobian matrix HQ of the nodal pressure with respect to the nodal water volume according to the following formula: HQ=-(ABA ^T ) ^-1 . 3. A method for moving and arranging pressure monitoring points for checking the water quantity of a water supply pipe network hydraulic model according to claim 1, wherein step S3 comprises the following steps: S31. According to the Jacobian matrix HQ, construct the objective function:

Among them: n is the total number of nodes in the water supply network, and sens is the index of the monitoring point location; S32, use the improved implicit enumeration optimization method to solve the objective function f(sens), and obtain a vector, that is, the monitoring point position general arrangement sensall_vector; the monitoring point position general arrangement sensall_vector describes the node index set of all monitoring points; S33. Use the kmeans++ method to cluster all the monitoring points in the sensall_vector according to the spatial position, and the number of classification groups cluster_num is equal to the number of monitoring points sens_num; S34. Sort the monitoring points in each classification group according to the 1-norm of the sensitivity vector from large to small to obtain a monitoring point position sequence; the sensitivity vector refers to a row in the Jacobian matrix HQ, in which the monitoring point The sensitivity vector of point i is the ith row of the Jacobian matrix HQ; the monitoring point position sequence represents the different positions of the same pressure monitoring sensor in each calibration cycle; If the length of the monitoring point position sequence is less than the check period round_num, add monitoring points to the end of the sequence in a circular manner until the length of the sequence is equal to the check period round_num; S35. For the first check cycle, select the first data from the monitoring point position sequence of each classification group in turn, and obtain a vector whose dimension is equal to the number of monitoring points sens_num, which is the first check cycle The monitoring point location deployment scheme sens_place[1]; for the second check cycle, select the second data from the monitoring point location sequence of each classification group in turn, and also obtain a vector whose dimension is equal to the number of monitoring points sens_num, This vector is the monitoring point position deployment scheme sens_place[2] of the second check cycle; and so on, the monitoring point position deployment scheme of all check cycles is obtained, which is the monitoring point movement scheme sens_place. 4. A method for moving and arranging pressure monitoring points for water supply network hydraulic model water quantity check according to claim 1, wherein step S4 comprises the following steps: S41. Divide the current calibration cycle into M time periods, marked as: period t ₁ , period t ₂ , . . . , period t _M ; S42, set a vector Q[t] for describing the water quantity arrangement of all nodes; the dimension of the vector Q[t] is equal to the total number of nodes of the model, and the value of each dimension represents the water quantity of the corresponding node respectively; S43. For the time period t ₁ , set the initial value Q[t ₁ ] of Q[t] as the average value of the total water volume of the model: Q[t ₁ ]=ones(1,n)*Q _avg Q _avg =Q _total /n Where: Q _total is the total water volume of the model (L/s), Q _avg is the average water volume of the model nodes (L/s), and ones(1,n) is a row vector whose length is the total number of nodes n and the elements are all 1; S44. Assuming that it is currently the k-th calibration cycle, read the monitoring point position deployment scheme sens_place[k], and set the pressure error amount to be eliminated by the target iteration: dH=H _o [sens]-H _s [sens] Where: H _o is the pressure (m) measured by the pressure monitoring sensor, and H _s is the pressure (m) after each iteration correction; S45, solve the Jacobian matrix HQ of the pipe network model when the water quantity is arranged as Q[t ₁ ]; S46, solve the equation HQ×dQ=dH, and obtain the correction amount dQ of Q[t ₁ ]; S47. Calculate the water quantity arrangement Q[t ₁ ] _n+1 of the next iteration by using the formula Q[t ₁ ] _n+1 =Q[t ₁ ] _n +dQ; S48, the iteration stops when dH is less than the set allowable error threshold, and the result obtained at this time is the final result of the water quantity arrangement Q[t ₁ ] in the time period t ₁ ; S49, take Q[t ₁ ] as the initial value, calculate the water quantity arrangement Q[t ₂ ] of the time period t ₂ ; and so on, repeat S45 to S48 until the water quantity arrangement Q[t] of all time periods is calculated; S410 , arranging all the obtained water consumption in a single period in time sequence to obtain a node water consumption pattern; the node water consumption pattern refers to the coefficient of the water consumption of the water consumption nodes in the pipe network model changing with time. 5. a kind of pressure monitoring point moving arrangement method for water supply pipe network hydraulic model water quantity check according to claim 3, is characterized in that: The specific steps of improving the implicit enumeration optimization method described in step S32 are as follows: S51. Set a vector sensall_init, which is the node index set of all monitoring points; S52. Randomly arrange all nodes to obtain an in-point group N _in (i), and select an in-point in sequence; S53. Randomly arrange all the initial solution vector elements to obtain an out point group N _out (j), and replace the out points with the in points in order to obtain a new solution sensall_new; S54. Substitute the original solution sensall_init and the new solution sensall_new into the objective function of step S31, solve the original solution f(sensall_init) and the new solution f(sensall_new), and select a better solution to enter the next iteration; S55. Repeat S52, S53, and S54 until a better solution cannot be generated, and terminate; the final solution obtained is the sensall_vector of the general arrangement of monitoring point positions. 6. a kind of pressure monitoring point moving arrangement method for water supply pipe network hydraulic model water quantity check according to claim 4, is characterized in that: The specific method of step S44 is as follows: Denote HQ as A, dH as b, and dQ as the unknown x to be solved, then the process of solving Ax=b by iterative method is:

After the iteration is over, the number of iterations is N, then the value of x is:

Among them: sum(A, axis=0) represents the sum of the columns of the matrix A, sum(A, axis=1) represents the sum of the rows of the matrix A, and abs(A) represents the absolute value of the matrix A, where Multiplication and division are both multiplication and division of counterpoints.

Images