Early warning system for damage risk of underground pipeline in third-party construction

Abstract

The invention relates to the technical field of risk early warning, and discloses a third-party construction underground pipeline damage risk early warning system, which is used for solving the problems that the risk of pipeline damage caused by third-party construction cannot be identified, the inspection efficiency is high and the cost is relatively high. The present invention comprises: a first database for storing historical data; the second database is used for storing the training data set and the test data set; the first processing unit performs training optimization of the evaluation model based on the training data set, and performs test of an optimization result through the test data set; the second processing unit calculates the probability of damage according to the information of the construction project to be evaluated, and the risk levels are matched according to a preset step to serve as the output of an evaluation result; and the terminal equipment is in communication connection with the second processing unit, inputs to-be-evaluated construction project information and outputs early warning information. According to the invention, the risk of pipeline damage of the construction project can be early warned accurately, the possibility of accident occurrence is reduced, and the operation cost of an operation and maintenance unit is saved.

Description

Technical Field

[0001] The present invention relates to the technical field of risk warning, and in particular to a third-party construction underground pipeline damage risk warning system. Background Art

[0002] Underground pipelines are important sources of energy, water, and heat in daily urban life. The safe operation of underground pipelines is related to the peace of thousands of households. With the rapid development of cities, various types of construction operations are increasing. In recent years, incidents of construction damaging gas pipelines have occurred from time to time, posing challenges to the safe operation of cities. Third-party construction damaging underground pipelines refers to the process of construction, excavation, and other operations near underground pipelines, which may cause gas pipelines to be damaged or ruptured due to improper operation or negligence, leading to serious consequences such as gas, hot water, electricity leakage, fire, and even explosion.

[0003] When underground pipeline operation and maintenance units discover third-party construction, they can only adopt encrypted operation, real-time monitoring, and rely on subjective experience to judge the possibility of damage to underground pipelines by the construction. They are unable to make an objective quantitative evaluation, which leads to a mismatch between the control measures taken and the construction hazard level, thus causing pipeline damage accidents. Even if no pipeline accidents are caused in time, excessive management will increase manpower and material costs and increase the burden on operation and maintenance companies. Summary of the invention

[0004] The present invention provides a third-party construction underground pipeline damage risk early warning system, which is used to solve the problems of failure to identify the risk of third-party construction damaging pipelines and high inspection efficiency and cost.

[0005] The present invention proposes a third-party construction underground pipeline damage risk early warning system, including a first database, a second database, a first processing unit, a second processing unit, and at least one terminal device;

[0006] The first database is a historical data storage database, including historical third-party construction data and underground pipeline protection area grid division data;

[0007] The second database is a training data set and a test data set storage database; based on the preset key factor information type that affects the occurrence of construction damage, the key factor information and the result data of the damage are extracted from the first database and stored in the training data set and the test data set respectively;

[0008] The first processing unit is an evaluation model optimization unit; the evaluation model is trained and optimized based on the training data set of the second database, and the optimization result is tested through the test data set;

[0009] The second processing unit is an evaluation model operation unit; based on the set trigger conditions, it synchronizes the optimized evaluation model parameters in the first processing unit; calculates the probability of damage according to the information of the construction project to be evaluated input by the terminal device, and outputs the evaluation result according to the preset ladder matching risk level;

[0010] The terminal device sends the input construction project information to be evaluated to the second processing unit, receives the evaluation result of the second processing unit and outputs it as early warning information.

[0011] Preferably, the evaluation model is constructed based on logistic regression, and the logistic regression hypothesis function is as follows:

[0012]

[0013] Among them, P(y=1|x;θ) represents the probability of construction damage to the target when the key factor information x is input, and g(θ T x) is the Sigmoid function, θ is the weight vector, x is the key factor information, and y = 1 indicates that construction damage has occurred.

[0014] Preferably, the evaluation model based on logistic regression is optimized as follows:

[0015] The cross entropy error is used to represent the difference between the model output and the true result. The particle swarm algorithm is used to search for the global optimal solution and optimize the weight parameter θ to minimize the cross entropy error.

[0016] Preferably, the cross entropy error BCE is

[0017] BCE=-w(y′-log(y))+(1-y′)log(1-y)

[0018] w is the weight parameter, which is used to adjust the weight of the first item;

[0019] y′ is the predicted label, representing the model’s predicted probability for the positive class, y′=P(y=1|x;θ);

[0020] y is the true label, which indicates the true probability of the positive class and takes the value of 0 or 1.

[0021] Preferably, the optimization method of the evaluation model based on logistic regression is:

[0022] Step S101, initializing the particle swarm: randomly generating a group of particles to form an initial particle swarm. Each particle represents a candidate solution for a group of weight parameters θ;

[0023] Step S102, calculating the objective function value: for each particle, calculating its corresponding cross entropy error;

[0024] Step S103, update pBest and gBest: record the optimal solution pBest found by each particle so far and the optimal solution gBest found by the entire particle swarm so far;

[0025] Step S104, update particle speed and position: update the speed and position of each particle according to pBest and gBest. The speed update formula is:

[0026] v′=v+c1·rand·(pBest-p)+c2·rand·(gBest-p)

[0027] The position update formula is:

[0028] p′=p+v′

[0029] Among them, v′ is the updated speed, v is the original speed, c1 and c2 are the preset empirical weights, p′ is the updated position, p is the original position;

[0030] Step S105, iterative optimization: repeat steps S102 to S104 until a predetermined number of iterations is reached or the algorithm converges;

[0031] Step S106, output the optimal solution: gBest is the global optimal solution found by the algorithm, that is, the optimal weight parameter θ.

[0032] Preferably, the preset experience weights c1 and c2 are both 2.

[0033] Preferably, the key factor information includes construction data, geographic data, pipeline data, weather data, and major event data;

[0034] in,

[0035] Construction data includes construction unit, construction month and date, construction content, and construction tools;

[0036] Geographic data includes grid attribute characteristics of the underground pipeline protection area to which the address location belongs;

[0037] Pipeline data include pipeline laying time, pipeline material, historical maintenance times, pipeline pressure, pipeline medium and pipeline ownership unit;

[0038] Weather data includes temperature, weather conditions, rain, snow and precipitation for the previous N days; N is the number of days for the preliminary review;

[0039] The major event data is the number of days with major events in the previous N days.

[0040] Preferably, the terminal device is provided with a feedback unit, and after the evaluated construction project is completed, the project information and the information of the actual damage are stored as new data in the first database.

[0041] Preferably, the second processing unit synchronizes the optimized evaluation model parameters in the first processing unit, and the triggering condition is:

[0042] Set a time period, or the accuracy improvement of the test data set reaches a set value.

[0043] Preferably, it further comprises a central control alarm unit, wherein the central control alarm unit is communicatively connected with the second processing unit;

[0044] When a construction project to be evaluated with a probability of damage greater than a set value appears in the second processing unit, the information of the construction project to be evaluated and the probability of damage are sent to the central control alarm unit and an alarm is issued.

[0045] At least one of the above technical solutions adopted in the embodiments of the present invention can achieve the following beneficial effects:

[0046] Through the present invention, it is possible to accurately evaluate the possibility of construction in the underground pipeline area causing damage to the underground pipeline. According to the evaluation results, the operation and maintenance unit takes corresponding control measures, which not only reduces the possibility of pipeline accidents caused by third-party construction in the underground pipeline, but also saves the operating costs of the operation and maintenance unit. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0048] Figure 1 A schematic diagram of a framework of a third-party construction underground pipeline damage risk early warning system provided in an embodiment of this specification;

[0049] Figure 2 The figure is a flow chart of an optimization method of an evaluation model based on logistic regression in one embodiment of the present specification. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0051] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0052] Figure 1A schematic diagram of a framework of a third-party construction underground pipeline damage risk warning system provided in an embodiment of this specification, the system includes a first database, a second database, a first processing unit, a second processing unit, and at least one terminal device.

[0053] The first database is a historical data storage database, including historical third-party construction data and underground pipeline protection area grid division data. The underground pipeline protection area grid division data includes map information, area division data of square grids according to the set minimum size, the number of each grid, and the regional characteristics of the grid (such as shopping malls, parks, green spaces, schools, hospitals, etc.).

[0054] The second database is a training data set and a test data set storage database; based on the preset key factor information types that affect the occurrence of construction damage, the key factor information and the result data of the damage are extracted from the first database and stored in the training data set and the test data set respectively. The training set is used to train the evaluation model, and the test set is used to test and verify the effect of the evaluation model.

[0055] In this example, the key factor information includes construction data, geographic data, pipeline data, weather data, and major event data; among them:

[0056] Construction data includes construction unit, construction month and date, construction content, construction tools, and whether construction is carried out at night;

[0057] Geographic data includes the grid attribute characteristics of the underground pipeline protection area to which the address location belongs (such as shopping malls, parks, green spaces, schools, hospitals, etc.);

[0058] Pipeline data include pipeline laying time, pipeline material, historical maintenance times, pipeline pressure, pipeline medium and pipeline ownership unit;

[0059] Weather data includes temperature, weather conditions, rain, snow and precipitation for the previous N days; N is the number of days for the preliminary review;

[0060] The major event data refers to the number of days with major events in the previous N days. Major events are collective activities of a city-wide nature, major holidays such as May Day, National Day, Spring Festival and other set holidays, or temporary events.

[0061] The first processing unit is an evaluation model optimization unit; the unit performs training optimization of the evaluation model based on the training data set of the second database, and tests the optimization result through the test data set.

[0062] In order to illustrate the relationship between each factor and whether construction damage occurs, the logistic regression model is used here to deduce the importance of each factor. Logistic regression is a common model for analyzing the influence relationship of factors in statistical learning, such as risk factor analysis. Logistic regression is suitable for analyzing the relationship between a group of factors and a qualitative result, such as whether a person's gender, age, and family conditions will affect his purchase of a certain item. The requirement of this project is to analyze the relationship between factors such as grid characteristics, weather, construction history, and qualitative results such as whether construction accidents occur or not, so it is suitable to use logistic regression for processing.

[0063] The evaluation model is built based on logistic regression. Assuming there is a set of factors x = {xi}, i = 1, 2, ..., n, and the target result is y = {0, 1}, we define the weight vector as θ = {θi}, i = 1, 2, ..., n. The logistic regression hypothesis function is shown in formula (1):

[0064]

[0065] Among them, P(y=1|x;θ) represents the probability of construction damage to the target when the key factor information x is input, and g(θ T x) is the Sigmoid function, θ is the weight vector, x is the key factor information, and y = 1 indicates that construction damage has occurred.

[0066] The optimization method of the evaluation model based on logistic regression is as follows: cross entropy error is used to represent the difference between the model output and the true result, the global optimal solution is searched through the particle swarm algorithm, and the weight parameter θ is optimized to minimize the cross entropy error.

[0067] The cross entropy error BCE is shown in formula (2):

[0068] BCE=-w(y′-log(y))+(1-y′)log(1-y) (2)

[0069] w is the weight parameter, which is used to adjust the weight of the first item;

[0070] y′ is the predicted label, representing the model’s predicted probability for the positive class, y′=P(y=|1x;θ);

[0071] y is the true label, which indicates the true probability of the positive class and takes the value of 0 or 1.

[0072] Estimating the weight parameter θ is equivalent to finding the parameter θ that minimizes the above cross entropy. In other words, the estimation of parameter θ is essentially an optimization problem. This problem can be solved by gradient descent. However, for the data of this project, there are many factors involved, and the Sigmoid function itself is not convex. Using the gradient descent method often does not find the global optimal solution. To overcome this problem, the particle swarm algorithm is used here.

[0073] Particle Swarm Optimization (PSO) is a bio-inspired optimization algorithm in the field of computational intelligence and a type of swarm intelligence optimization algorithm. The advantages of this method include: the ability to find the global optimal solution of the objective function, without requiring the objective function to be convex or differentiable.

[0074] In this embodiment, the optimization method of the evaluation model based on logistic regression is as follows: Figure 2 As shown, specifically including:

[0075] Step S101, initializing the particle swarm: randomly generating a group of particles to form an initial particle swarm. Each particle represents a candidate solution for a group of weight parameters θ;

[0076] Step S102, calculating the objective function value: for each particle, calculating its corresponding cross entropy error;

[0077] Step S103, update pBest and gBest: record the optimal solution pBest found by each particle so far and the optimal solution gBest found by the entire particle swarm so far;

[0078] Step S104, update particle speed and position: update the speed and position of each particle according to pBest and gBest. The speed update formula is shown in formula (3):

[0079] v′=v+c1·rand·(pBest-p)+c2·rand·(gBest-p) (3)

[0080] The position update formula is shown in formula (4):

[0081] p′=p+v′ (4)

[0082] Where v′ is the updated speed, v is the original speed, c1 and c2 are both preset empirical weights (the classic value is 2 for both), p′ is the updated position, p is the original position, and rand is a random number between 0 and 1, which is used to increase the uncertainty of the search and prevent the algorithm from falling into the local optimal value.

[0083] Step S105, iterative optimization: repeat steps S102 to S104 until a predetermined number of iterations is reached or the algorithm converges;

[0084] Step S106, output the optimal solution: gBest is the global optimal solution found by the algorithm, that is, the optimal weight parameter θ.

[0085] The second processing unit is an evaluation model running unit; based on the set trigger conditions, it synchronizes the optimized evaluation model parameters in the first processing unit; calculates the probability of damage according to the information of the construction project to be evaluated input by the terminal device, and outputs the evaluation result as the risk level according to the preset ladder matching.

[0086] In this example, the risk level is divided into three levels: high risk, medium risk, and low risk according to the probability of damage occurring.

[0087] The second processing unit synchronizes the optimized evaluation model parameters in the first processing unit, and the triggering condition is: a set time period, or the accuracy improvement range tested by the test data set reaches a set value.

[0088] The terminal device sends the input construction project information to be evaluated to the second processing unit, receives the evaluation result of the second processing unit and outputs it as early warning information.

[0089] The terminal equipment is generally deployed in each grassroots project management department, and sends the information of the construction project to be evaluated to the second processing unit through the communication link for analysis and evaluation, and outputs the probability and risk level of damage.

[0090] The terminal device is also provided with a feedback unit, which stores the project information and the information of the actual damage as new data into the first database after the evaluated construction project is completed.

[0091] During specific use, when the number of updated data in the first database exceeds the preset threshold, the newly added data will be extracted according to the requirements of the training data set and the test data set, the training data set and the test data set will be updated, the evaluation model will be optimized and tested, the test results will be commissioned, and the optimized model parameters will be retained.

[0092] The present example further includes a central control alarm unit, which is in communication with the second processing unit; when a construction project to be evaluated with a probability of damage greater than a set value appears in the second processing unit, the construction project information to be evaluated and the probability of damage are sent to the central control alarm unit and an alarm is issued. The central control alarm unit is generally set in a centralized management department of a district or city.

[0093] The present invention can evaluate the risk of third-party construction damaging underground pipelines in underground pipeline protection areas based on historical construction data of urban areas in combination with other factors, guide pipeline right-holding units to optimize inspection routes, reduce the possibility of third-party construction damaging pipelines, and save operating costs for operation and maintenance units.

[0094] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A third-party underground pipeline damage risk early warning system, characterized in that: comprising a first database, a second database, a first processing unit, a second processing unit, and at least one terminal device; The first database is a historical data storage database, including historical third-party construction data and underground pipeline protection area grid division data; The second database is a training data set and a test data set storage database; based on the preset key factor information type that affects the occurrence of construction damage, the key factor information and the result data of the damage are extracted from the first database and stored in the training data set and the test data set respectively; The first processing unit is an evaluation model optimization unit; Performing training optimization of the evaluation model based on the training data set of the second database, and testing the optimization result through the test data set; The second processing unit is an evaluation model operation unit; based on the set trigger conditions, it synchronizes the optimized evaluation model parameters in the first processing unit; calculates the probability of damage according to the information of the construction project to be evaluated input by the terminal device, and outputs the evaluation result according to the preset ladder matching risk level; The terminal device sends the input construction project information to be evaluated to the second processing unit, receives the evaluation result of the second processing unit and outputs it as early warning information.

2. The third-party construction underground pipeline damage risk early warning system according to claim 1 is characterized in that: The evaluation model is constructed based on logistic regression, and the logistic regression hypothesis function is as follows: Among them, P(y=|1x;θ) represents the probability of construction damage to the target when the key factor information x is input, and g(θ T x) is the Sigmoid function, θ is the weight vector, x is the key factor information, and y = 1 indicates that construction damage has occurred.

3. The third-party construction underground pipeline damage risk early warning system according to claim 2 is characterized in that: The evaluation model based on logistic regression is optimized as follows: The cross entropy error is used to represent the difference between the model output and the true result. The particle swarm algorithm is used to search for the global optimal solution and optimize the weight parameter θ to minimize the cross entropy error.

4. The third-party construction underground pipeline damage risk early warning system according to claim 3 is characterized in that: The cross entropy error BCE is BCE=-w(y′-log(y))+(1-y′)log(1-y) w is the weight parameter, which is used to adjust the weight of the first item; y′ is the predicted label, representing the model’s predicted probability for the positive class, y′=P(y=|1x;θ); y is the true label, which indicates the true probability of the positive class and takes the value of 0 or 1.

5. The third-party construction underground pipeline damage risk early warning system according to claim 4 is characterized in that: The optimization method of the evaluation model based on logistic regression is: Step S101, initializing the particle swarm: randomly generating a group of particles to form an initial particle swarm. Each particle represents a candidate solution for a group of weight parameters θ; Step S102, calculating the objective function value: for each particle, calculating its corresponding cross entropy error; Step S103, update pBest and gBest: record the optimal solution pBest found by each particle so far and the optimal solution gBest found by the entire particle swarm so far; Step S104, update particle speed and position: update the speed and position of each particle according to pBest and gBest. The speed update formula is: v′=v+c1·rand·(pBest-p)+c2·rand·(gBest-p) The position update formula is: p′=p+v′ Among them, v′ is the updated speed, v is the original speed, c1 and c2 are the preset empirical weights, p′ is the updated position, p is the original position; Step S105, iterative optimization: repeat steps S102 to S104 until a predetermined number of iterations is reached or the algorithm converges; Step S106, output the optimal solution: gBest is the global optimal solution found by the algorithm, that is, the optimal weight parameter θ.

6. The third-party construction underground pipeline damage risk early warning system according to claim 5 is characterized in that: The preset experience weights c1 and c2 are both 2.

7. The third-party construction underground pipeline damage risk early warning system according to any one of claims 1 to 6, characterized in that: The key factor information includes construction data, geographic data, pipeline data, weather data, and major event data; in, Construction data includes construction unit, construction month and date, construction content, and construction tools; Geographic data includes grid attribute characteristics of the underground pipeline protection area to which the address location belongs; Pipeline data includes pipeline laying time, pipeline material, historical maintenance times, pipeline pressure, pipeline medium and pipeline ownership unit; Weather data includes temperature, weather conditions, rain, snow and precipitation for the previous N days; N is the number of days for the preliminary review; The major event data is the number of days with major events in the previous N days.

8. The third-party construction underground pipeline damage risk early warning system according to any one of claims 1 to 6, characterized in that: The terminal device is provided with a feedback unit, and after the evaluated construction project is completed, the project information and the information of the actual damage are stored as new data in the first database.

9. The third-party construction underground pipeline damage risk early warning system according to any one of claims 1 to 6, characterized in that: The second processing unit synchronizes the optimized evaluation model parameters in the first processing unit, and its triggering condition is: Set a time period, or the accuracy improvement of the test data set reaches a set value.

10. The third-party construction underground pipeline damage risk early warning system according to any one of claims 1 to 6, characterized in that: It also includes a central control alarm unit, which is communicatively connected with the second processing unit; When a construction project to be evaluated with a probability of damage greater than a set value appears in the second processing unit, the information of the construction project to be evaluated and the probability of damage are sent to the central control alarm unit and an alarm is issued.

Images