Wire breakage prediction model construction and wire breakage prediction method for prestressed concrete cylinder pipe

Abstract

The application discloses a wire breakage prediction model construction method and a wire breakage prediction method for a prestressed concrete cylinder pipe, and the wire breakage prediction model construction method comprises the following steps: obtaining historical wire breakage data of target pipe sections in which wire breakage occurs in the prestressed concrete cylinder pipe, wherein the historical wire breakage data comprises service time length, wire breakage time and position information of the target pipe sections in which wire breakage occurs; performing segmentation processing on the target pipe sections according to the position information of the target pipe sections, and performing wire breakage data statistics according to the wire breakage time corresponding to the target pipe sections included in each pipe section to obtain wire breakage data statistical results corresponding to each pipe section, wherein the wire breakage data statistical results comprise position information corresponding to each pipe section, wire breakage time, service time length and wire breakage quantity of the target pipe sections included in each pipe section; and training a preset model according to the wire breakage data statistical results until a preset training condition is met, so as to obtain a wire breakage prediction model for the prestressed concrete cylinder pipe.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of big data analysis, and in particular to a broken wire prediction model construction method and a broken wire prediction method for a prestressed concrete cylinder pipe. BACKGROUND

[0002] With the rapid development of economy and the accelerated urbanization, the demand for urban water supply is increasing. At the same time, in order to ensure the supply of water resources in the city, the urban water supply network system is also becoming larger and larger. The safety of water supply for urban residents is closely related to the urban water supply network system. Due to the characteristics of prestressed concrete cylinder pipe (PCCP), such as composite structure, high internal and external pressure resistance, good joint sealing, strong anti-seismic ability, and convenient and fast construction, PCCP is widely used in long-interval water supply trunk lines and urban water supply projects. However, as the service life of PCCP increases, risks such as broken wires may occur, and regular detection of broken wires is required to take timely measures. The loss caused by the failure to detect and repair in time is difficult to estimate, which not only makes it difficult to ensure the safety of urban water supply, but also may have a huge impact on the lives of residents along the water supply pipeline and the local ecological environment.

[0003] At present, the research and application of PCCP pipe broken wire mainly focuses on PCCP monitoring and broken wire detection and other related aspects. When the pipe breaks, the broken wire position is determined in time through related detection equipment, and then the broken pipe is maintained. More research is on the detection principle, and there is a lack of early prediction before the broken wire occurs. SUMMARY

[0004] Therefore, the technical problem to be solved by the present application is to overcome the defect that the existing detection method lacks early prediction before the broken wire occurs, thereby providing a broken wire prediction model construction method and a broken wire prediction method for a prestressed concrete cylinder pipe, a device and an electronic equipment.

[0005] According to a first aspect, the present application discloses a broken wire prediction model construction method, comprising: obtaining historical broken wire data of a plurality of target pipe sections in a prestressed concrete cylinder pipe in which a broken wire occurs, the historical broken wire data including a use time length, a broken wire time and position information of the target pipe section in which the broken wire occurs; segmenting the plurality of target pipe sections according to the position information of the plurality of target pipe sections and obtaining a broken wire data statistical result corresponding to each pipe section according to the broken wire time corresponding to the target pipe section included in each pipe section, the broken wire data statistical result including position information corresponding to each pipe section, the broken wire time, the use time length and the number of broken wires of the target pipe section included; training a preset model according to the broken wire data statistical result until a preset training condition is met, to obtain a broken wire prediction model for the prestressed concrete cylinder pipe.

[0006] The filament breakage prediction model construction method provided by the application can realize prediction of the number of filament breaks of the prestressed concrete cylinder pipe, and the number of filament breaks of the prestressed concrete cylinder pipe predicted by the filament breakage prediction model can be used to intervene in the pipe section that has filament breaks in advance.

[0007] Optionally, the training of the preset model according to the filament breakage data statistical result until the preset training condition is met to obtain the filament breakage prediction model of the prestressed concrete cylinder pipe comprises: performing expansion processing on the filament breakage data statistical result by using a preset interpolation method to obtain expansion data; and training the prediction model by using the expansion data.

[0008] Through the above method, the expansion data obtained by performing expansion processing on the filament breakage data statistical result has more universal data characteristics and meets the requirement of the amount of data for training the preset model.

[0009] Optionally, the preset interpolation method comprises a cubic interpolation method.

[0010] The expansion data obtained by performing expansion on the filament breakage data statistical result by using the cubic interpolation method can generate a smooth curve as a whole. Because it is easy to generate a relatively sharp fluctuation, the highest point of the curve is higher than the highest node, and the lowest point of the curve is lower than the lowest node, so that the expanded data is independent of each other and close to the actual data.

[0011] Optionally, the preset model comprises a width echo state network model.

[0012] The width echo state is formed by combining a width learning system and an echo state network, and has the advantage of a width learning incremental algorithm. For the weights that change due to the addition of training data or the addition of hidden layer nodes, the width echo network can quickly update the original model using these data, so as to learn a more actual rule.

[0013] According to a second aspect, the application also discloses a filament breakage prediction method for a prestressed concrete cylinder pipe, comprising the following steps: acquiring use data of a to-be-predicted prestressed concrete cylinder pipe section, wherein the use data comprises position information of the pipe section and starting use time of a target pipe section contained in the pipe section; inputting the use data into a filament breakage prediction model for a prestressed concrete cylinder pipe, which is constructed by using the filament breakage prediction model construction method according to the first aspect or any optional embodiment of the first aspect; and determining a time when filament breakage occurs in the to-be-predicted prestressed concrete cylinder pipe section and a number of filament breaks according to an output result of the filament breakage prediction model.

[0014] The prestressed concrete cylinder pipe wire breaking prediction method provided by the application realizes wire breaking prediction of the prestressed concrete cylinder pipe, changes the research on the detection principle of wire breaking monitoring, realizes automatic and accurate prediction of wire breaking, and is more likely to meet the urban water supply demand.

[0015] Optionally, after the time and the number of wire breakages of the to-be-predicted prestressed concrete cylinder pipe section are determined according to the output result of the wire breaking prediction model, the method further includes: evaluating the time and the number of wire breakages of the to-be-predicted prestressed concrete cylinder pipe section according to a regression model evaluation index, and determining an error value corresponding to the time and the number of wire breakages of the to-be-predicted prestressed concrete cylinder pipe section.

[0016] Through the above method, the error value corresponding to the time and the number of wire breakages of the to-be-predicted prestressed concrete cylinder pipe section is determined, so that the accuracy of the prediction result can be evaluated.

[0017] According to a third aspect, the application also discloses a wire breaking prediction model construction device, including: a first acquisition module, configured to acquire historical wire breaking data of a plurality of target pipe sections in which wire breaking occurs in a prestressed concrete cylinder pipe, the historical wire breaking data including a use duration, a wire breaking time and position information of the target pipe sections in which wire breaking occurs; a determination module, configured to segment the plurality of target pipe sections according to the position information of the plurality of target pipe sections and to obtain a wire breaking data statistical result corresponding to each pipe section by performing wire breaking data statistics according to the wire breaking time corresponding to the target pipe sections included in each pipe section, the wire breaking data statistical result including position information corresponding to each pipe section, the wire breaking time, the use duration and the number of target pipe sections included in each pipe section; and a training module, configured to train a preset model according to the wire breaking data statistical result until a preset training condition is met, to obtain a wire breaking prediction model of the prestressed concrete cylinder pipe.

[0018] According to a fourth aspect, the embodiments of the present application further disclose a wire breakage prediction device for a prestressed concrete cylinder pipe. The device comprises: a second obtaining module configured to obtain usage data of a prestressed concrete cylinder pipe segment to be predicted, the usage data comprising position information of the pipe segment and starting usage time of a target pipe joint contained in the pipe segment; a prediction module configured to input the usage data into a wire breakage prediction model for a prestressed concrete cylinder pipe, which is constructed by using the wire breakage prediction model construction method according to the first aspect or any possible implementation manner of the first aspect; and an output module configured to determine time and number of wire breakage of the prestressed concrete cylinder pipe segment to be predicted according to an output result of the wire breakage prediction model.

[0019] According to a fifth aspect, the embodiments of the present application further disclose an electronic device, comprising: at least one processor; and a memory connected with the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the wire breakage prediction model construction method according to the first aspect or any possible implementation manner of the first aspect, or perform the steps of the wire breakage prediction method for a prestressed concrete cylinder pipe according to the second aspect or any possible implementation manner of the second aspect.

[0020] According to a sixth aspect, the embodiments of the present application further disclose a computer readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the wire breakage prediction model construction method according to the first aspect or any possible implementation manner of the first aspect, or the steps of the wire breakage prediction method for a prestressed concrete cylinder pipe according to the second aspect or any possible implementation manner of the second aspect are implemented. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the drawings needed in the specific embodiments or the prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0022] Figure 1 a flow chart of a specific example of the wire breakage prediction model construction method in the embodiments of the present application;

[0023] Figure 2 a schematic diagram of a specific example of the wire breakage prediction model construction method in the embodiments of the present application;

[0024] Figure 3 a schematic diagram of a specific example of the wire breakage prediction model construction method in the embodiments of the present application;

[0025] Figure 4 A flow chart of one specific example of the broken wire prediction method for the prestressed steel cylinder concrete pipe in the embodiment of the present application;

[0026] Figure 5 A schematic diagram of one specific example of the broken wire prediction method for the prestressed steel cylinder concrete pipe in the embodiment of the present application;

[0027] Figure 6 A principle block diagram of one specific example of the broken wire prediction model construction device in the embodiment of the present application;

[0028] Figure 7 A principle block diagram of one specific example of the broken wire prediction device for the prestressed steel cylinder concrete pipe in the embodiment of the present application;

[0029] Figure 8 A specific example of the electronic device in the embodiment of the present application. DETAILED DESCRIPTION

[0030] The technical solutions of the present application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0031] In the description of the present application, it should be noted that the orientations or positional relationships indicated by the terms “center”, “upper”, “lower”, “left”, “right”, “vertical”, “horizontal”, “inner”, “outer” and the like are based on the orientations or positional relationships shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms “first”, “second”, “third” are only for the purpose of description, and cannot be understood as indicating or implying relative importance.

[0032] In the description of the present application, it should be noted that unless otherwise explicitly specified and limited, the terms “mounting”, “connection”, “connecting” should be understood broadly, for example, it can be fixed connection, or detachable connection, or integrally connected; it can be mechanical connection, or electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, or the internal communication of two elements, it can be wireless connection, or wired connection. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0033] In addition, the technical features involved in the different embodiments of the application described below can be combined with each other as long as they do not conflict with each other.

[0034] The embodiment of the application discloses a broken wire prediction method for a prestressed concrete cylinder pipe, as shown in the figure, the method comprises the following steps: Figure 1

[0035] Step 101, historical broken wire data of a target pipe section in which a broken wire occurs in a prestressed concrete cylinder pipe is acquired, the historical broken wire data comprises a service time length of the target pipe section, a broken wire time and position information of the target pipe section.

[0036] Exemplarily, the prestressed concrete cylinder pipe (PCCP) of the target pipe section can be a PCCP pipe of any pipe section, the service time length of the target pipe section can be a time period from starting to use the target pipe section to the time when the broken wire occurs in the target pipe section, the position of the broken wire in the PCCP pipe and the time of the broken wire can be determined by a broken wire detection device, and the position information of the target pipe section can comprise but is not limited to geographical position information and spatial distribution information of the target pipe section. In the embodiment of the application, the occurrence trend of the broken wire in the PCCP pipe section is related to factors such as the soil, water quality, water pressure of the environment in which the pipe section is located and the service time length of the pipe section. When the service time length is the same and the length of each pipe section is relatively short, the similar pipe sections have similar broken wire occurrence trends due to the similar environment in which the pipe sections are located. The pipe section number of the PCCP represents the spatial arrangement sequence of the pipe section, and the geographical position information and the spatial distribution information of the pipe section can be represented by the pipe section number of the PCCP.

[0037] Step 102, the plurality of target pipe sections are subjected to segmentation processing according to the position information of the plurality of target pipe sections, and broken wire data statistics are performed according to the broken wire time corresponding to the target pipe sections included in each pipe section to obtain a broken wire data statistical result corresponding to each pipe section, the broken wire data statistical result comprising position information corresponding to each pipe section, broken wire time, service time length and broken wire number of the target pipe sections included in each pipe section.

[0038] Exemplarily, the segmentation processing process can comprise but is not limited to dividing pipe sections of different numbers of target pipe sections into one pipe section according to the position information of the target pipe sections. In the embodiment of the application, the similar pipe sections have similar broken wire occurrence trends due to the similar environment in which the pipe sections are located. In order to analyze the broken wire law from the spatial distribution and also in order to increase the data amount, the pipe sections are divided into one section according to the pipe sections of different numbers, and the broken wire number is summarized according to a preset time period as a unit. The preset time period can comprise but is not limited to one month, so as to statistically analyze the broken wire number law of the pipe sections. For example, 58 pipe sections are divided into 12 sections, 19 sections and 7 sections according to 5 pipe sections, 3 pipe sections and 8 pipe sections respectively, and the broken wire data statistical result of a total of 38 pipe sections in each month is obtained.

[0039] ​In step 103, the preset model is trained according to the filament breakage data statistical result until a preset training condition is met, so as to obtain a filament breakage prediction model of the prestressed steel cylinder concrete pipe.

[0040] Exemplarily, the type of the preset model is not limited in the embodiments of the present application, as long as the filament breakage prediction model of the prestressed steel cylinder concrete pipe obtained by training can effectively and reliably realize filament breakage prediction.

[0041] The filament breakage prediction method of the prestressed steel cylinder concrete pipe provided by the present application can obtain the filament breakage data statistical result corresponding to each pipe section by statistically processing the historical filament breakage data of the target pipe section, train the preset model by using the filament breakage data statistical result corresponding to each pipe section, and obtain the filament breakage prediction model.

[0042] As an optional embodiment of the present application, the training of the preset model according to the filament breakage data statistical result until the preset training condition is met, so as to obtain the filament breakage prediction model of the prestressed steel cylinder concrete pipe, comprises: expanding the filament breakage data statistical result by using a preset interpolation method to obtain expanded data; and training the prediction model by using the expanded data.

[0043] Exemplarily, the specific type of the preset interpolation method is not limited in the embodiments of the present application, as long as the expansion processing of the filament breakage data statistical result can be effectively and reliably realized. The data characteristics of the expanded data obtained by expanding the filament breakage data statistical result are more universal, and can better meet the requirement of the data quantity when training the preset model.

[0044] As an optional embodiment of the present application, the preset interpolation method comprises a cubic interpolation method.

[0045] Exemplarily, the broken wire data statistics result is limited, considering the requirement of the preset model for the data amount and the universality of the data characteristics, it is necessary to expand the broken wire data statistics result to a larger capacity. In the obtained historical broken wire data, the data of each pipe section is not affected by other pipe sections, and the broken wire data of each pipe section is independent, so the broken wire data between each pipe section in the broken wire data statistics result is also independent, so the method of uniformly inserting a certain number of data between each two data cannot be used to expand the broken wire data statistics result, because the newly inserted data may not be located between two data, in order to make the expanded data closer to the real data, a cubic interpolation method is used to expand the broken wire data statistics result. The expanded data obtained by using the cubic interpolation method can generate a smooth curve as a whole. Due to the possibility of generating relatively severe fluctuations, the highest point of the curve is higher than the highest node, and the lowest point of the curve is lower than the lowest node, so that the expanded data is independent of each other and close to the real data.

[0046] In the embodiment of the application, the cubic interpolation method is to solve a cubic equation, divide the known broken wire data statistics result corresponding data into a small interval one by one, and then fit the data to a curve. Assuming that there are n+1 data, the first data is x0=A, and the n+1 data is xn+1=B, the data interval [A, B] is divided into n intervals [x0, x1], [x1, x2], …, [xn, xn+1], and the cubic spline function equation is defined as shown in formula (1), wherein a, b, c and d represent undetermined coefficients: n n-1 n i i i i

[0047] i i i i 2 i 3 (1)

[0048] The cubic spline function S(x) needs to meet the following three conditions:

[0049] 1. In each sub-interval [x i , x i+1 ], S(x) = S i (x) is a cubic equation;

[0050] 2. Meet the interpolation condition: S i (x) = y i ​​​​​​​​​​​​​​(i = 0, 1, 2...n)

[0051] 3. The first and second derivatives of S(x) exist and are continuous, making the curve smooth.

[0052] Since the data interval endpoints are known and determined, S i (x) at the two endpoints of the interval is the boundary value A and B, as shown in equation (2) and equation (3), and S i (x), the first derivative S' i (x) and the second derivative S" i (x) can also be listed, as shown in equation (4), equation (5) and equation (6).

[0053] S' i (x) = A (2)

[0054] S' i (x n ) = B (3)

[0055] S i (x) = a i +b i (x-x i )+c i (x-x i ) 2 +d i (x-x i ) 3 (4)

[0056] S' i (x) = b i +2c i (x-x i )+3d i (x-x i ) 2 (5)

[0057] S" i (x) = 2c i +6d i (x-x i ) (6)

[0058] Through the operation of the above formula, the four unknowns a i , b i , c i and d i in the cubic function equation can be calculated, so that the data can be inserted to fit the data to the curve. The broken wire data statistical results are expanded using the cubic interpolation method, and the expanded data can reflect the broken wire trend of the PCCP pipe, and can also increase the data quantity to meet the training needs of the subsequent model.

[0059] As an optional embodiment of the present invention, the preset model includes a width echo state network model.

[0060] For example, the Wide Echo State Network (BESN) model is a combination of a wide learning system and an echo state network (ESN) model. It has the advantages of the wide learning incremental algorithm. For weights that change due to the addition of training data or hidden layer nodes, the wide echo network can quickly use this data to update the original model, thereby learning a more realistic pattern. In this embodiment, a BESN network model is established based on the mapping relationship between the time and number of broken wires in the expanded data, and the model is trained. The BESN network model consists of an input layer, a mapping layer, a reinforcement layer, and an output layer. The model structure is defined as follows: the number of neurons in the mapping layer is n, the number of ESN units in the reinforcement layer is m, and the input variables are as shown in equation (7), where u(t) represents the input variables, u1(t), u2(t), ..., u m (t) represents a total of m sets of input variables:

[0061] u(t)=[u1(t),u2(t),...,u m (t)] T (7)

[0062] The expression for the mapping layer is as described in equation (8), where Z1, Z2, ..., Z n Representing n groups of mapping nodes in the mapping layer, the expression is similar in form to the input variables:

[0063] Z = [Z1, Z2, ..., Zn] n (8)

[0064] The expression for the reinforcement layer is shown in equation (9), where φ i W is a linear or nonlinear activation function. ei With β ei These are the random weights and the bias, Z. i For the i-th group of mapping nodes in the mapping layer:

[0065] Z i =φ i (XW ei +β ei (9)

[0066] Concatenate n groups of mapping layer nodes into Z. n = [Z1, Z2, ..., Z n ], using H j Let j represent the node of the j-th enhancement layer containing r neurons. Then:

[0067] H j= ξ j (Z n W hi + β hj ) (10)

[0068] In formula (10), ξ j is a nonlinear activation function, W hi and β hj are random weights and biases respectively, m groups of enhanced nodes are spliced into H m = [H1, H2,..., H m ], denoted as A = [Z n |H m ], A represents the combination of mapping layer nodes and enhanced layer nodes, and the output Y of the BESN network model is as shown in formula (11), wherein, W out represents an output weight matrix:

[0069] Y = [Z1, Z2,..., Z n |H1, H2,..., H m ] W out = [Z n |H m ] W out = AW out (11)

[0070] Since W ei , β ei , W hi , β hj are randomly generated and remain unchanged during the training process, the only weight that the network needs to learn is the output weight W m . After each sample is input into the BESN network, it is first trained and learned by the mapping layer through full connection learning, and then the output is input into the enhanced layer. Each neural unit of the enhanced layer is an echo state network unit, the reservoir pool update process of each echo state network unit is as shown in formula (12), the output of each echo state network unit is as shown in formula (13), and the output weight matrix calculation of the echo state network unit is as shown in formula (14).

[0071] x(t+1) = (1-α)x(t) + αf(W in u(t) + Wx(t)) (12)

[0072] y(t+1) = g(W out x(t+1)) (13)

[0073] W out = (X T X) -1 X T Y (14)

[0074] In formula (12), (13), (14), x is a reservoir state, y is an output of an echo state network, W in is a weight matrix of an input connected to a reservoir, W is a weight matrix of a connection within the reservoir, W out is a weight matrix of a connection of the reservoir to an output, a is a proportional coefficient, f and g are respectively an activation function of a reservoir cell and an activation function of an output cell, and t represents a filament breaking time. All states of the i-th ESN cell form H i , x represents a reservoir state matrix, and y represents a real output of the model.

[0075] In the embodiments of the present application, in order to improve the accuracy of data prediction, the echo state network (ESN) is changed to a width echo state network. The input data is generated into a first feature mapping node through a mapping function, and then transmitted to a second feature node. The network diagram of the echo state network (ESN) is as shown in Figure 2 , and the network diagram of the width echo state network model (BESN) is as shown in Figure 3 . Similarly, the feature vector is extracted and feature sparsification and nonlinear mapping are performed, and then the output is input to a reinforcement layer. Each neural unit of the reinforcement layer is an echo state network unit (ESN). The reservoir in the echo state network is used to process input information. Multiple reinforcement layer neural units can perform parallel processing on the information input to the reinforcement layer, thereby obtaining more rich state information. The output layer obtains the output weight by calculating the linear relationship between the state of the reservoir and the output data of the mapping layer. The network training process updates and records the state information in parallel through the mapping layer and the reinforcement layer. The changes of the time series data are recorded in the state information of the mapping layer and the reinforcement layer. Finally, the relationship between the state information and the real output is calculated through the linear fitting relationship of the output layer to obtain the output weight, and the network training is completed.

[0076] The embodiments of the present application also disclose a filament breaking prediction method for a prestressed concrete cylinder pipe, as shown in Figure 4 , the method comprises the following steps:

[0077] In step 201, usage data of a prestressed concrete cylinder pipe segment to be predicted is acquired. The usage data comprises position information of the pipe segment and starting usage time of a target pipe joint contained.

[0078] Exemplarily, the prestressed concrete cylinder pipe segment to be predicted can be any prestressed concrete cylinder pipe segment, and the usage data comprises position information and starting usage time of the pipe segment. In the embodiments of the present application, the PCCP pipe joint number can be used to represent the geographical position information and spatial distribution information of the pipe joint. The acquisition of the pipe joint number corresponding to the prestressed concrete cylinder pipe segment to be predicted can determine the position information of the pipe segment.

[0079] Step 202, input the use data into the broken wire prediction model of the prestressed concrete cylinder pipe constructed by the broken wire prediction model construction method in the above embodiment.

[0080] Step 203, determine the time and the number of broken wires when the prestressed concrete cylinder pipe segment to be predicted occurs broken wires according to the output result of the broken wire prediction model.

[0081] Exemplarily, the use data is input into the broken wire prediction model of the prestressed concrete cylinder pipe constructed by the broken wire prediction model construction method in the above embodiment, and the time and the number of broken wires when the prestressed concrete cylinder pipe segment to be predicted occurs broken wires can be determined according to the output result. In the embodiment of the application, it is assumed that the latest detection broken wire data and the data of 12 months before the current month of each pipe joint (pipe segment) are input into the trained broken wire prediction model, and the model can automatically calculate the number of broken wires of each pipe joint in the future 1-6 months, and the prediction result can be viewed through a prediction curve diagram, wherein the broken wire number prediction result of 12 PCCP pipes is as shown in the following figure: Figure 5

[0082] The broken wire prediction method of the prestressed concrete cylinder pipe provided by the application realizes the broken wire prediction of the prestressed concrete cylinder pipe, changes the research on the detection principle of the broken wire monitoring, realizes the automatic and accurate prediction of the broken wire, and is more easy to meet the urban water supply demand.

[0083] As an optional embodiment of the application, after determining the time and the number of broken wires when the prestressed concrete cylinder pipe segment to be predicted occurs broken wires according to the output result of the broken wire prediction model, the method further comprises: evaluating the time and the number of broken wires when the prestressed concrete cylinder pipe segment to be predicted occurs broken wires according to a regression model evaluation index, and determining the error value corresponding to the time and the number of broken wires when the prestressed concrete cylinder pipe segment to be predicted occurs broken wires.

[0084] Exemplarily, the regression evaluation index can include but is not limited to root mean square error (RMSE), mean square error (MSE), square absolute percentage error (SMAPE), and mean absolute error (MAE), and the specific calculation formula is as follows:

[0085]

[0086] ​

[0087]

[0088]

[0089] In equations (14), (15), (16), and (17), n represents the number of samples. Let y(k) represent the k-th predicted value and y(k) represent the k-th true value. The smaller the RMSE, MSE, SMAPE, and MAE, the smaller the prediction error, indicating a better model fit. In this embodiment, it is assumed that the latest detected wire breakage data of each pipe section (segment) and the data up to the current month for the previous 12 months are input into the trained wire breakage prediction model. The model can automatically calculate the number of wire breaks for each pipe section in the next 1-6 months. The error of the calculated number of wire breaks for each pipe section in the next 1-6 months is evaluated based on the regression evaluation index. The error evaluation results of the predicted number of wire breaks for 12 PCCP pipe segments are shown in Table 1 below.

[0090] Table 1 Evaluation Indicators for Predicted Number of Broken Filaments in 12 PCCP Segments

[0091]

[0092] This invention also discloses a device for constructing a broken wire prediction model, such as... Figure 6 As shown, the device includes: a first acquisition module 301, used to acquire historical wire breakage data of target pipe sections in prestressed steel cylinder concrete pipes where wire breakage has occurred, the historical wire breakage data including the usage duration, wire breakage time, and location information of the target pipe sections where wire breakage has occurred; a determination module 302, used to segment the multiple target pipe sections according to the location information of the multiple target pipe sections and to perform wire breakage data statistics based on the wire breakage time of the target pipe sections contained in each pipe section to obtain the wire breakage data statistics result corresponding to each pipe section, the wire breakage time of the target pipe sections contained in each pipe section, the usage duration, and the number of broken wires; and a training module 303, used to train a preset model according to the wire breakage data statistics result until the preset training conditions are met to obtain a wire breakage prediction model for prestressed steel cylinder concrete pipes.

[0093] The filament breakage prediction model construction device provided by the application can realize prediction of the number of filament breaks of the prestressed concrete cylinder pipe, and the number of filament breaks of the prestressed concrete cylinder pipe predicted by the filament breakage prediction model can be used to intervene in the pipe section with filament breaks in advance.

[0094] As an optional embodiment of the application, the training module comprises: expanding the filament breakage data statistical result by using a preset interpolation method to obtain expansion data; and training the prediction model by using the expansion data.

[0095] As an optional embodiment of the application, the preset interpolation method comprises a cubic interpolation method.

[0096] As an optional embodiment of the application, the preset model comprises a wide echo state network model.

[0097] The application also discloses a filament breakage prediction device for a prestressed concrete cylinder pipe, as shown in the accompanying drawings. Figure 7 The device comprises: a second acquisition module 501 configured to acquire use data of a pipe section to be predicted, the use data comprising position information of the pipe section and starting use time of a target pipe section contained in the pipe section; a prediction module 502 configured to input the use data into a filament breakage prediction model for a prestressed concrete cylinder pipe, which is constructed by using the filament breakage prediction model construction method of the first aspect or any optional embodiment of the first aspect; and an output module 503 configured to determine time and number of filament breaks of the pipe section to be predicted when filament breaks occur in the pipe section to be predicted according to an output result of the filament breakage prediction model.

[0098] As an optional embodiment of the application, the filament breakage prediction device for a prestressed concrete cylinder pipe further comprises an evaluation module configured to evaluate time and number of filament breaks of the pipe section to be predicted when filament breaks occur in the pipe section to be predicted according to a regression model evaluation index, and determine an error value corresponding to the time and number of filament breaks of the pipe section to be predicted when filament breaks occur in the pipe section to be predicted.

[0099] The application also provides an electronic device, as shown in the accompanying drawings. Figure 8 The electronic device can comprise a processor 401 and a memory 402, wherein the processor 401 and the memory 402 can be connected through a bus or other means, Figure 8 for example, the bus connection is taken as an example.

[0100] The processor 401 can be a central processing unit (CPU). The processor 401 can also be other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, or a combination of the above.

[0101] The memory 402, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs and modules, such as program instructions / modules corresponding to the prestressed steel cylinder concrete pipe broken wire prediction method in the embodiments of the present application. The processor 401 performs various functional applications and data processing of the processor by running the non-transitory software programs, instructions and modules stored in the memory 402, that is, implements the broken wire prediction model construction method in the above method embodiments, or implements the prestressed steel cylinder concrete pipe broken wire prediction method in the above method embodiments.

[0102] The memory 402 can include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application required by a function; the data storage area can store data created by the processor 401, etc. In addition, the memory 402 can include a high-speed random access memory, and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory 402 can optionally include a memory disposed remotely with respect to the processor 401, and these remote memories can be connected to the processor 401 through a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0103] The one or more modules are stored in the memory 402, and when executed by the processor 401, perform the broken wire prediction model construction method as shown in Figure 1 the embodiments, or perform the prestressed steel cylinder concrete pipe broken wire prediction method as shown in Figure 4 the embodiments.

[0104] The above electronic device specific details can be understood by referring to the corresponding related descriptions and effects of the embodiments shown in Figure 1 or Figure 4 the embodiments, which will not be repeated here.

[0105] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The program can be stored in a computer readable storage medium. When the program is executed, the program can include the processes of the above-mentioned embodiment methods. The storage medium can be a magnetic disc, an optical disc, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD), etc. The storage medium can also include a combination of the above-mentioned types of memories.

[0106] Although the embodiments of the present application are described in conjunction with the drawings, various modifications and changes can be made by those skilled in the art without departing from the spirit and scope of the present application, and such modifications and changes fall within the scope defined by the appended claims.

Claims

1. A method for constructing a broken wire prediction model, characterized in that, The method comprises the following steps: obtaining historical broken wire data of a plurality of target pipe sections in which broken wires occur in prestressed concrete cylinder pipes, the historical broken wire data comprising service time length, broken wire time and position information of the target pipe sections in which broken wires occur; segmenting the plurality of target pipe sections according to the position information of the plurality of target pipe sections and obtaining broken wire data statistical results corresponding to each pipe section according to the broken wire time of the target pipe sections included in each pipe section, the broken wire data statistical results comprising position information, broken wire time, service time length and broken wire number corresponding to each pipe section, the segmenting process comprising dividing pipe sections of different numbers of target pipe sections into one pipe section according to the position information of the target pipe sections; training a preset model according to the broken wire data statistical results until a preset training condition is met to obtain a broken wire prediction model of the prestressed concrete cylinder pipe, the preset model comprising a width echo state network model, the width echo state network model comprising an input layer, a mapping layer, a reinforcement layer and an output layer, the number of neurons in the mapping layer being n, the number of echo state network units in the reinforcement layer being m, and the input variable being as shown in the following formula: wherein, represents an input variable, represents an input of m groups of variables; the expression of the mapping layer is as shown in the following formula: wherein, represents n groups of mapping nodes in the mapping layer; the expression of the reinforcement layer is as shown in the following formula: wherein, is a linear or non-linear activation function, and are random weights and biases, respectively, is the i-th set of mapping nodes in the mapping layer; The n groups of mapping layer nodes are spliced into , where represents the jth group of enhanced layer nodes containing r neurons, and has: wherein, is a non-linear activation function, and are random weights and bias, respectively; m groups of enhancement nodes are concatenated as , denoted as , A represents the combination of the mapping layer node and the enhancement layer nodes, the output of the wide echo state network model is shown as follows: wherein, represents an output weight matrix; after each sample is input into the width echo state network model, the original data is trained by full connection learning through the mapping layer, and then the output is input into the reinforcement layer, each neuron in the reinforcement layer is an echo state network unit, and the reserve pool updating process of each echo state network unit is as shown in the following formula: the output of each echo state network unit is as shown in the following formula: the output weight matrix of the echo state network unit is calculated as shown in the following formula: wherein, is the reservoir state, is the output of the echo state network, is the weight matrix connecting the input to the reservoir, is the weight matrix connecting the reservoir internally, is the weight matrix connecting the reservoir to the output, is a scaling factor, and are the activation function of the reservoir units and the activation function of the output units, respectively, t denotes the time of filament breakage, all states of the i-th echo state network unit form , X denotes the reservoir state matrix, Y denotes the true output of the model.

2. The method of claim 1, wherein, the training of the preset model according to the broken wire data statistical results until the preset training condition is met to obtain the broken wire prediction model of the prestressed concrete cylinder pipe comprises: performing expansion processing on the broken wire data statistical results by using a preset interpolation method to obtain expansion data; training the prediction model by using the expansion data.

3. The method of claim 2, wherein, The preset interpolation method comprises a cubic interpolation method.

4. A broken wire prediction method of a prestressed concrete cylinder pipe, characterized by, The method comprises the following steps: obtaining service data of a to-be-predicted prestressed concrete cylinder pipe section, the service data comprising position information of the pipe section and starting service time of target pipe sections included in the pipe section; inputting the service data into a broken wire prediction model of a prestressed concrete cylinder pipe constructed by using the broken wire prediction model construction method in any one of claims 1-3; determining time and broken wire number when the to-be-predicted prestressed concrete cylinder pipe section occurs broken wires according to the output result of the broken wire prediction model.

5. The method of claim 4, wherein, After the time and the broken wire number when the to-be-predicted prestressed concrete cylinder pipe section occurs broken wires are determined according to the output result of the broken wire prediction model, the method further comprises: evaluating the time and the broken wire number when the to-be-predicted prestressed concrete cylinder pipe section occurs broken wires according to a regression model evaluation index to determine an error value corresponding to the time and the broken wire number when the to-be-predicted prestressed concrete cylinder pipe section occurs broken wires.

6. A broken filament prediction model construction apparatus characterized by comprising: The method comprises the following steps: The first obtaining module is configured to obtain historical broken wire data of a plurality of target pipe sections in which broken wires occur in the prestressed concrete cylinder pipe, wherein the historical broken wire data comprises a service time length, a broken wire time and position information of the target pipe sections in which broken wires occur; The determining module is configured to perform segmented processing on the plurality of target pipe sections according to the position information of the plurality of target pipe sections, and perform broken wire data statistics according to the broken wire time corresponding to the target pipe sections included in each pipe section to obtain a broken wire data statistical result corresponding to each pipe section, wherein the broken wire data statistical result comprises position information corresponding to each pipe section, the broken wire time, the service time length and the number of broken wires of the target pipe sections included in each pipe section, and the segmented processing comprises dividing pipe sections of different numbers of target pipe sections into one pipe section according to the position information of the target pipe sections. The training module is configured to train a preset model according to the broken wire data statistical result until a preset training condition is met to obtain a broken wire prediction model of the prestressed concrete cylinder pipe, wherein the preset model comprises a width echo state network model, and the width echo state network model comprises an input layer, a mapping layer, a reinforcement layer and an output layer, the number of neurons in the mapping layer is n, the number of echo state network units in the reinforcement layer is m, and an input variable is as shown in the following formula: wherein, represents an input variable, represents an input of m groups of variables; An expression of the mapping layer is as shown in the following formula: wherein, represents n groups of mapping nodes in the mapping layer; An expression of the reinforcement layer is as shown in the following formula: wherein, is a linear or non-linear activation function, and are random weights and biases, respectively, is the i-th set of mapping nodes in the mapping layer; The n groups of mapping layer nodes are spliced into , where represents the jth group of enhanced layer nodes containing r neurons, and there is wherein, is a non-linear activation function, and are random weights and biases, respectively; m groups of enhancement nodes are concatenated as , denoted as , A represents the combination of the mapping layer node and the enhancement layer nodes, the output of the wide echo state network model is given by the following equation: wherein, represents an output weight matrix; After each sample is input into the width echo state network model, the original data is trained and learned through full-connection learning of the mapping layer, and then the output is input into the reinforcement layer, each neuron in the reinforcement layer is an echo state network unit, and a reservoir pool updating process of each echo state network unit is as shown in the following formula: An output of each echo state network unit is as shown in the following formula: An output weight matrix of the echo state network unit is calculated as shown in the following formula: wherein, is the reservoir state, is the output of the echo state network, is the weight matrix connecting the input to the reservoir, is the weight matrix connecting the reservoir to the output, is the weight matrix connecting the reservoir to the output, is the scaling factor, and are the activation function of the reservoir units and the activation function of the output units, respectively, t denotes the break time, and all states of the i-th echo state network unit form , X denotes the reservoir state matrix, and Y denotes the true output of the model.

7. A broken wire prediction device for a prestressed concrete cylinder pipe, characterized by The method comprises the following steps: The second obtaining module is configured to obtain service data of a to-be-predicted prestressed concrete cylinder pipe section, wherein the service data comprises position information of the pipe section and a starting service time of the target pipe sections included in the pipe section. The prediction module is configured to input the service data into the broken wire prediction model of the prestressed concrete cylinder pipe constructed by using the broken wire prediction model construction method according to any one of claims 1-3. The time and the number of broken wires when the to-be-predicted prestressed concrete cylinder pipe section occurs broken wires are determined according to an output result of the broken wire prediction model.

8. An electronic device, comprising: The method comprises the following steps: At least one processor; and a memory connected with the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the broken wire prediction model construction method according to any one of claims 1-3, or perform the steps of the broken wire prediction method of the prestressed concrete cylinder pipe according to claim 4 or 5.

9. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the broken wire prediction method of the prestressed concrete cylinder pipe according to any one of claims 1-3, or implement the steps of the broken wire prediction method of the prestressed concrete cylinder pipe according to claim 4 or 5.

Images