A road and bridge icing prediction method

Abstract

The present invention provides a method for predicting icing of a road or bridge. First, environmental parameters of the road or bridge pavement are collected, and each environmental parameter is an influencing factor. Then, an initial threshold is set for each environmental parameter. The initial threshold is used to judge whether the road surface state is icy under the single factor condition. The icy road surface state is recorded as +T, and the non-icy road surface state is recorded as ‑T, where T is an arbitrary positive number. Then, the global weight of each environmental parameter under the condition of multiple factors is calculated by the ANP network analysis method. Each factor is multiplied by the corresponding global weight and added to obtain a comprehensive index K. If the comprehensive index K is greater than 0, the road surface state is icy; if K is less than 0, the road surface state is non-icy; if K is equal to 0, the road surface state is in an icing critical state. The present invention has high judgment accuracy, greatly reduces energy consumption, and reduces the occurrence of traffic accidents. The judgment matrix of the present invention can be continuously updated according to the judgment result.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent construction, and relates to a road and bridge snow and ice melting auxiliary technology, and specifically to a road and bridge icing prediction method. Background Art

[0002] As a type of smart city construction, intelligent monitoring systems have been applied to various fields. For example, in the construction of the national-level Internet of Vehicles pilot zone project in Xiqing District, Tianjin, the intelligent network system was used to exchange information with the surrounding environment, realize interactive interconnection, and build a more dimensional intelligent transportation system. In the Xiangyang and Shenzhen Pingshan Internet of Vehicles projects, the vehicle-road collaborative edge computing platform, the autonomous driving cloud platform, and the traffic cloud control platform have significantly improved traffic efficiency, optimized travel experience, and effectively improved driving safety. However, in the existing technology, whether the road surface of the bridge is frozen does not depend on the temperature alone. If the judgment of whether it is frozen is inaccurate, it will either cause traffic accidents or cause a lot of energy waste, which is not conducive to the construction of smart cities. Therefore, there is an urgent need for a technology that can accurately judge and predict whether the road surface of the bridge is frozen. Summary of the invention

[0003] The purpose of the present invention is to improve the accuracy of prediction and judgment of whether the road surface of a road or bridge is frozen through a road or bridge icing prediction method, so as to make relevant plans in advance, respond in time, and improve the intelligent monitoring and management of roads and bridges.

[0004] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0005] A road and bridge icing prediction method comprises the following steps:

[0006] Step 1: Collect environmental parameters of the road surface, each of which is an influencing factor;

[0007] Step 2: Set an initial threshold for each environmental parameter. First, use the initial threshold to determine whether the road surface is icy under the single factor condition. The icy road surface is recorded as +T, and the non-icy road surface is recorded as -T, where T is an arbitrary positive number.

[0008] Step 3, the global weight of each environmental parameter under the condition of multiple factors is analyzed by ANP network analysis method;

[0009] Step 4: Multiply each factor by the corresponding global weight and add them together to obtain a comprehensive index K; if the comprehensive index K is greater than 0, the road surface condition is icy; if K is less than 0, the road surface condition is non-icy; if K is equal to 0, the road surface condition is a critical icing state.

[0010] Furthermore, the environmental parameters include temperature, humidity, snowfall, wind speed, atmospheric pressure, water vapor pressure, PM2.5 and PM10.

[0011] Furthermore, the specific method of step 3 is:

[0012] Step 3.1, first construct an ANP network structure model, which is divided into a control layer and a network layer. The control layer is further divided into a target layer and a criterion layer. The target layer is to turn on / off the snow-melting and ice-melting equipment. The network layer is composed of environmental parameters, and each parameter of the environmental parameters is defined as an element of the network layer.

[0013] Step 3.2: For the criterion layer, the weight is calculated by AHP; for the network layer, the local weight of each influencing factor under multiple criteria needs to be calculated by weighted supermatrix, and the final global weight is obtained by multiplying the criterion layer weight and the local weight of the network layer.

[0014] Furthermore, the criterion layer is road surface condition, and the weight is 1.

[0015] Furthermore, in step 3.2, the weighted supermatrix is ​​obtained as follows:

[0016] S1, group the environmental parameters according to the element categories to obtain several element groups,

[0017] S2, then compare the importance of each influencing factor within and between element groups, and get the judgment matrix D ij ;

[0018] S3, calculate the relative importance M of each row of the judgment matrix by the square root method;

[0019] S4, normalize the relative importance M obtained to obtain the characteristic value w of each row, and finally obtain D j The eigenvector D of each judgment matrix when it is the secondary criterion w ij , where i represents the i-th element group, and j represents the j-th element as the subcriterion;

[0020] S5. According to the relative importance of the elements, the eigenvectors of each judgment matrix are combined into an unweighted supermatrix W

[0021]

[0022] n represents the number of element groups, W ik (i,k=1,2..n) represents the comparison of the relative importance of the elements in the i-th element group with the elements in the k-th element group as the secondary criterion, through W ik Construct the unweighted supermatrix W;

[0023] S6, taking the road surface condition as the main criterion and each element group as the secondary criterion to construct a judgment matrix, compare the relative importance of each element group, calculate the eigenvalue aij combination, and obtain the weighted matrix A;

[0024]

[0025] aij represents the eigenvalue obtained by comparing the relative importance of the i-th element group with other element groups when the j-th element group is used as the secondary criterion, and n is the matrix order, that is, the number of element groups;

[0026] S7, weighted supermatrix W'=AW, when W' exists to an infinite power, the columns of the weighted supermatrix W' will tend to the same vector, and a limit supermatrix will be obtained. At this time, the column vector of the limit supermatrix is ​​the local weight of each element under the road condition criterion; the formula of the limit supermatrix is: ;

[0027] Furthermore, in step S1, the environmental parameters of the network layer are divided into three groups, namely element group one of weather influencing factors, element group two of atmospheric influencing factors, and element group three of air quality influencing factors. The elements in element group one are temperature, humidity, snowfall, and wind speed; the elements in element group two are atmospheric pressure and water vapor pressure; and the elements in element group three are PM2.5 and PM10.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] (1) The present invention first judges the road surface state and equipment operation status under single-factor conditions through the designed initial threshold, and then makes decisions on the start and stop of equipment under multi-factor conditions through the ANP network analysis method and issues corresponding instructions to the equipment monitoring module. This method takes into account the mutual influence between various factors or adjacent levels and can be used to solve various quantitative and non-quantitative, rational and non-rational decision-making problems.

[0030] (2) Compared with other icing identification methods, the ANP network analysis method used in the present invention calculates the weights of various influencing factors in an empirical manner, does not require the collection of a large number of samples to establish a simulation model, is simple and efficient, and the judgment matrix can be continuously modified according to the judgment results. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 The figure is a flow chart of a method for predicting road and bridge icing according to an embodiment of the present invention.

[0032] Figure 2 The figure is a flow chart of the ANP network analysis method according to an embodiment of the present invention.

[0033] Figure 3This is a structural model diagram of the ANP network analysis method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0034] The present invention will be described in detail below in conjunction with the accompanying drawings and embodiments. The embodiments shown in the accompanying drawings are only used to explain the present invention but cannot limit the present invention.

[0035] like Figure 1 As shown, the present invention provides a road and bridge icing prediction method, comprising the following steps:

[0036] Step 1: Collect environmental parameters of the road surface, including temperature, humidity, snowfall, wind speed, atmospheric pressure, water vapor pressure, PM2.5 and PM10, each of which is an influencing factor;

[0037] Step 2: Set an initial threshold for each environmental parameter. First, use the initial threshold to determine whether the road surface is icy under the single factor condition. The icy road surface is recorded as +T, and the non-icy road surface is recorded as -T. T is any positive number, such as 1.

[0038] Step 3, the global weight of each environmental parameter under the condition of multiple factors is analyzed by ANP network analysis method;

[0039] Step 4: Multiply each factor by the corresponding global weight and add them together to obtain a comprehensive index K; if the comprehensive index K is greater than 0, the road surface condition is icy; if K is less than 0, the road surface condition is non-icy; if K is equal to 0, the road surface condition is a critical icing state.

[0040] like Figure 2 and Figure 3 As shown, the specific steps of ANP network analysis method are as follows:

[0041] Step 3.1, first construct an ANP network structure model, which is divided into a control layer and a network layer. The control layer is further divided into a target layer and a criterion layer. The target layer is to turn on / off the snow-melting and ice-melting equipment. The network layer is composed of environmental parameters, and each parameter of the environmental parameters is defined as an element of the network layer.

[0042] Step 3.2: For the criterion layer, the weight is calculated by AHP; for the network layer, the local weight of each influencing factor under multiple criteria needs to be calculated by weighted supermatrix, and the final global weight is obtained by multiplying the criterion layer weight and the local weight of the network layer.

[0043] Specific calculation case: A set of initial thresholds are designed, namely temperature (0℃), humidity (80%), snowfall (0.25mm), wind speed (2m / s), atmospheric pressure (101.3Kpa), water vapor pressure (0.1Kpa), PM2.5 (25.0), PM10 (37.0), current (100A), voltage (230V), and electric energy (23kw.h), as the basis for preliminarily judging whether the road surface is icy under single factor conditions.

[0044] For the road surface condition, the initial threshold is first used to determine whether the road surface condition is icy under single-factor conditions, and then the global weight of each environmental parameter under multi-factor conditions is analyzed through the ANP network analysis method. The judgment result under single-factor conditions and the global weight under multi-factor conditions are combined to comprehensively determine whether the road surface condition is icy.

[0045] The ANP network analysis method is to make decisions on the target under the influence of multiple factors. The ANP network structure model constructed first is divided into two parts: the control layer and the network layer, wherein the control layer is further divided into the target layer and the criterion layer. The weight of the criterion layer can be calculated by the AHP hierarchical analysis method. For the network layer, the local weights of each influencing factor under multiple criteria need to be calculated by the weighted supermatrix method, and the final global weight can be obtained by multiplying the criterion layer weight and the local weight of the network layer.

[0046] In this embodiment, the target layer is to turn on / off the snow-melting and ice-melting equipment (A), and the criterion layer only considers the road surface state (B1). Because the criterion layer has only one factor, the weight can be calculated without the AHP analytic hierarchy process.

[0047] The network layer is composed of multiple element groups, each of which contains multiple elements, and the elements are interconnected and interdependent. There is also a relationship of dependence and feedback between the element groups. Element group one of the network layer of this embodiment is weather influencing factors (C1), element group two is atmospheric influencing factors (C2), and element group three is air quality influencing factors (C3). The elements in element group one are temperature (D1), humidity (D2), snowfall (D3), and wind speed (D4). The elements in element group two are atmospheric pressure (D5) and water vapor pressure (D6). The elements in element group three are PM2.5 (D7) and PM10 (D8). Then, the relationship between the influencing factors is studied through expert surveys and group discussions, or the degree of mutual influence between the elements is given based on empirical values, and an influencing factor table is prepared. The final ANP model is as follows: Figure 3 shown.

[0048] The local weight calculation method of each influencing factor under multiple criteria is as follows:

[0049] In this embodiment, the unweighted supermatrix W and the weighted matrix A are first established based on the road surface state (B1), and then the unweighted supermatrix W and the weighted matrix A are multiplied to obtain the weighted supermatrix W'. Finally, the limit supermatrix W' is calculated ∞ The local weight of each element based on the road surface condition (B1) can be obtained.

[0050] The unweighted supermatrix W is an unweighted supermatrix that takes the road surface condition (B1) as the main criterion, one element as the secondary criterion, and compares the relative importance of other elements. First, the importance of each influencing factor within and between groups must be compared pairwise to obtain a judgment matrix.

[0051] When D1, D2, D3, D4 in C1 are j When (j=1,2,......,8) is the secondary criterion, 8 4×4 initial judgment matrices D can be obtained.

[0052] When D5, D6 in C2 are j When (j=1,2,......,8) is the secondary criterion, 8 2×2 initial judgment matrices D can be obtained.

[0053] When D7, D8 in C3 are j When (j=1,2,......,8) is the secondary criterion, 8 2×2 initial judgment matrices D can be obtained.

[0054] The initial judgment matrix D can calculate the relative importance M of each row by the square root method. i , after normalization, we can get the eigenvector w of each matrix i , taking the judgment matrix of temperature as the secondary criterion in element group 1 as an example, the judgment matrix and eigenvector w are shown in Table 1:

[0055] Table 1 Initial judgment matrix D based on temperature in element group 1 11

[0056]

[0057] The table shows the relative importance between two elements. The relative importance M of each row of the judgment matrix is ​​calculated by the square root method. i ; The relative importance M i Perform normalization to obtain the eigenvalue w of each row i ,

[0058] The square root method is to calculate the nth root of the product of the elements of each row of the judgment matrix to obtain the relative importance M of each row. i (i=1, 2....n), the relative importance of each row M i(i=1, 2....n) to form a relative importance vector M, where n is the matrix hierarchy. Then the obtained relative importance M is normalized to obtain the eigenvalue w of each row i (i=1, 2....n), that is, the weight, the characteristic value w of each row i The eigenvalue vector w is formed. Finally, the maximum eigenvalue λ is calculated. max . Where M, w and λ max The formulas are shown in formula (1), (2), and (3). In formula (3), w i is the relative importance M of the i-th row i Normalized eigenvalues.

[0059] Formula (1)

[0060] is the initial judgment matrix D 11 The element in the i-th row and j-th column, n is the initial judgment matrix D 11 The order of D 11 , n=4.

[0061] For example, for the first row,

[0062] M i is the relative importance of the i-th row in the judgment matrix D.

[0063] Formula (2)

[0064] For example, for the first row,

[0065] Then the maximum eigenvalue is calculated using formula (3) and (4) and a consistency check is performed. The final result CR < 0.1, which meets the requirements.

[0066] Formula (3)

[0067] In the above formula, Si is the i-th row value of the test vector S, and the test vector is equal to the initial judgment matrix D multiplied by the eigenvalue vector w, that is, S=Dw.

[0068] The consistency test first calculates the consistency ratio CR = CI / RI. When CR < 0.1, the requirement is met. When CR > 0.1, the judgment matrix needs to be readjusted and corrected. RI can be obtained by looking up the table.

[0069] Formula (4)

[0070] Initial judgment matrix D 11 The eigenvalue vector w is used as the judgment matrix , we get the characteristic vector D obtained by taking the judgment matrix with temperature as the secondary criterion in element group 1 as an example w 11 =w, that is:

[0071]

[0072] Similarly, this method can be used to calculate the judgment matrix between elements in other groups. Finally, we can get D j When it is the secondary criterion, the eigenvector of each judgment matrix is ​​the weight value D w ij , where i represents the i-th (i=1,2,3) element group, and j represents the j-th (j=1,2,......,8) element as the secondary criterion.

[0073] The judgment matrices with each element in element group 1 as sub-criteria are: , , , , , , and ;

[0074] The judgment matrices with the elements in element group 2 as sub-criteria are: , , , , , , and ;

[0075] The judgment matrices with the elements in element group three as sub-criteria are: , , , , , , and ;

[0076] The above judgment matrix is ​​composed into an unweighted supermatrix W, which is composed as follows:

[0077]

[0078] W ij (i,j=1,2,3) represents the comparison of the relative importance of the elements in the i-th element group with the elements in the j-th element group as the secondary criterion. ij Construct the unweighted supermatrix W.

[0079] The weighted matrix A is constructed by taking the road surface state (B1) as the main criterion and each element group as the secondary criterion to construct the initial judgment matrix, compare the relative importance of each element group, and calculate its eigenvalue a ij Combination. ij It represents the eigenvalue a obtained by comparing the relative importance of the i-th element group with other element groups when the j-th element group is used as the secondary criterion. The element groups in this example are weather influencing factors (C1), atmospheric influencing factors (C2), and air quality influencing factors (C3). Taking the secondary criterion as weather influencing factors (C1), a judgment matrix is ​​constructed to compare the relative importance of each element group, and the eigenvalue a is obtained by the square root method. 11 , a 21 , a 31 As shown in Table 2. Similarly, the eigenvalues ​​of the judgment matrix between each element group are calculated using C2 and C3 as the secondary criteria, and finally the weighted matrix A is obtained.

[0080] Table 2 Initial judgment matrix between element groups when weather influencing factors are secondary criteria

[0081]

[0082]

[0083] Similarly, by constructing the initial judgment matrix between element groups with atmospheric influence factors as the secondary criterion, and calculating the eigenvalue a 12 , a 22 , a 32 .

[0084] By constructing the initial judgment matrix between element groups with air quality influencing factors as the secondary criteria, and calculating the eigenvalue a 13 , a 23 , a 33 .

[0085] The combined weighted matrix A is as follows:

[0086]

[0087] The weighted supermatrix W'=aW, that is:

[0088]

[0089] When W' is infinite, the columns of the weighted supermatrix W' will tend to the same vector, and a limit supermatrix will be obtained. At this time, the column vector of the limit supermatrix is ​​the local weight of each element under the road condition criterion. The formula of the limit supermatrix is:

[0090] (5)

[0091] In this embodiment, the local weights of the influencing elements are temperature (0.23), humidity (0.204), snowfall (0.144), wind speed (0.109), atmospheric pressure (0.143), water vapor pressure (0.065), PM2.5 (0.052), and PM10 (0.053).

[0092] The global weight is obtained by multiplying the local weight of each influencing element of the network layer by the weight of the corresponding criterion layer. In this embodiment, the criterion layer has only one road surface state, so the criterion layer weight is (1.0), and the global weight of each influencing element is the same as the local weight.

[0093] The decision is made through programming in Python, and the judgment value of each influencing factor under the single factor condition is considered to be "1" when it is judged as icing; the judgment value of non-icing is considered to be "-1", and then the judgment value under the single factor condition is multiplied by the global weight of the corresponding influencing factor obtained under the multi-factor condition after calculation by the ANP network analysis method, so as to obtain the judgment value of each influencing factor under the multi-factor condition. These judgment values ​​are added together to make a judgment on whether the road surface state is icy under multiple influencing factors. If the result is greater than 0, it means that the road surface state is "iced"; if the result is less than 0, it means that the road surface state is "non-iced". In this embodiment, the initial thresholds of each influencing factor are set to temperature (0°C), humidity (80%), snowfall (0.25mm), wind speed (2m / s), atmospheric pressure (101.3Kpa), water vapor pressure (0.1Kpa), PM2.5 (25.0), and PM10 (37.0). Two groups of examples are taken to verify the ANP network analysis method, as shown in the following table

[0094] Table 3 Decision table

[0095]

[0096] The results obtained by ANP network analysis are the same as those of the two groups of examples. Therefore, network analysis can better reflect the complex internal relationship between various influencing factors in actual situations.

[0097] It should be noted that all the above-mentioned initial judgment matrices of the present invention are determined by empirical values, or obtained through historical data analysis, and are adopted only after passing consistency verification. During use, the initial matrix can be corrected in turn according to the actual road and bridge pavement status (such as manual verification), continuously improved, and the prediction accuracy can be improved.

[0098] In the description of the present invention, it should be understood that all terms described in the present invention have the same meaning as those generally understood by technicians in the technical field of the present invention, and are only used to facilitate the description of the embodiments described in the patent of the present invention, and cannot limit the present invention.

[0099] It should be understood that, in addition to the sensors and monitors used in this embodiment, other environmental parameter sensors that affect icing and any monitors that can monitor whether the equipment is faulty can also be used. In addition to the above 8 factors, the influencing elements taken in the ANP network analysis method in this embodiment can also be any other factors that affect road icing.

[0100] It should be understood that the limit supermatrix of the ANP network analysis method in this embodiment needs to be calculated by PYTHON.

[0101] The above-mentioned implementation schemes and examples are only for a more detailed and specific understanding of the present invention, and cannot be understood as a limitation of the patent of the present invention. It should be pointed out that for ordinary technicians in this field, any modification, variation, combination or equivalent replacement of the description and drawings of the present invention is included in the patent protection scope of the present invention.

Claims

1. A road and bridge icing prediction method, characterized in that: The following steps are involved: Step 1: Collect environmental parameters of the road surface, each of which is an influencing factor; Step 2: Set an initial threshold for each environmental parameter. First, use the initial threshold to determine whether the road surface is icy under the single factor condition. The icy road surface is recorded as +T, and the non-icy road surface is recorded as -T, where T is an arbitrary positive number. Step 3, the global weight of each environmental parameter under the condition of multiple factors is analyzed by ANP network analysis method; Step 4: Multiply each factor by the corresponding global weight and add them together to obtain a comprehensive index K; if the comprehensive index K is greater than 0, the road surface state is icy; if K is less than 0, the road surface state is not icy; if K is equal to 0, the road surface state is in a critical icing state; The specific method of step 3 is: Step 3.1, first construct an ANP network structure model, which is divided into a control layer and a network layer. The control layer is further divided into a target layer and a criterion layer. The target layer is to turn on / off the snow-melting and ice-melting equipment. The network layer is composed of environmental parameters, and each parameter of the environmental parameters is defined as an element of the network layer. Step 3.2: For the criterion layer, the weight is calculated by AHP; for the network layer, the local weight of each influencing factor under multiple criteria needs to be calculated by weighted supermatrix, and the final global weight is obtained by multiplying the criterion layer weight and the local weight of the network layer; In step 3.2, the weighted supermatrix is ​​obtained as follows: S1, group the environmental parameters according to the element categories to obtain several element groups, S2, then compare the importance of each influencing factor within and between element groups, and get the judgment matrix D ij ; S3, calculate the relative importance M of each row of the judgment matrix by the square root method; S4, normalize the relative importance M obtained to obtain the characteristic value w of each row, and finally obtain D j The eigenvector D of each judgment matrix when it is the secondary criterion w ij , where i represents the i-th element group, and j represents the j-th element as the secondary criterion; S5. According to the relative importance of the elements, the eigenvectors of each judgment matrix are combined into an unweighted supermatrix W n represents the number of element groups, W ik , i,k=1,2..n, which means that the elements in the i-th element group are compared with the elements in the k-th element group as the secondary criterion for the relative importance of each other, through W ik Construct the unweighted supermatrix W; S6, taking the road surface condition as the main criterion and each element group as the secondary criterion to construct a judgment matrix, compare the relative importance of each element group, and calculate its eigenvalue a ij Combine to obtain the weighted matrix A; a ij It represents the eigenvalue obtained by comparing the relative importance of the i-th element group with other element groups when the j-th element group is used as the secondary criterion, and n is the matrix order, that is, the number of element groups; S7, weighted supermatrix W'=AW, when W' is infinitely powered, the columns of the weighted supermatrix W' will tend to the same vector, and a limit supermatrix will be obtained. At this time, the column vector of the limit supermatrix is ​​the local weight of each element under the road surface condition criterion; the formula of the limit supermatrix is: .

2. The road and bridge icing prediction method according to claim 1, characterized in that: The environmental parameters include temperature, humidity, snowfall, wind speed, atmospheric pressure, water vapor pressure, PM2.5 and PM10.

3. The road and bridge icing prediction method according to claim 1, characterized in that: The criterion layer is the road surface state, and the weight is 1.

4. The road and bridge icing prediction method according to claim 1, characterized in that: In step S1, the environmental parameters of the network layer are divided into three groups, namely, element group one of weather influencing factors, element group two of atmospheric influencing factors, and element group three of air quality influencing factors. The elements in element group one are temperature, humidity, snowfall, and wind speed; the elements in element group two are atmospheric pressure and water vapor pressure; and the elements in element group three are PM2.5 and PM10.

Images