Urban drainage pipe network detection difficulty prediction method based on entropy weight method and grey clustering

Abstract

The invention discloses an urban drainage pipe network detection difficulty prediction method based on an entropy weight method and grey clustering. The method comprises the following steps: Step 1, determining influence factors of pipe network detection difficulty; step 2, determining a single index category threshold value of each influence factor in the step 1; 3, constructing a whitening weight function and a weight coefficient calculation model of each index in the step 2; step 4, endowing each index weight to the step 3 and calculating a comprehensive weight coefficient matrix; and Step 5, judging the difficulty level to which each project belongs. A difficulty grade evaluation model based on grey clustering is constructed, a key influence factor is selected by adopting a statistical survey method, the key influence factor is weighted by adopting an entropy weight method, and the selected factor is subjected to difficulty grade comprehensive evaluation by grey weighted clustering.

Description

Technical Field

[0001] The present invention relates to the technical field of operation and maintenance of urban drainage pipe networks, and in particular to a method for predicting the detection difficulty of urban drainage pipe networks based on entropy weight method and grey clustering. Background Art

[0002] Drainage network inspection is to detect and judge the function and structure of the pipeline through manual, CCTV and sonar inspection methods to determine the operating status of the pipeline. Regular inspection of the pipeline, timely diagnosis of pipeline defects, and effective repair measures can reduce the safety risks during pipeline operation, reduce the probability of accidents, and extend the service life of the pipeline. At present, the pipeline inspection project volume is mainly determined by the quotas issued by various provinces and cities to determine the unit price of the project volume. The quotas limit the scope and methods of use of the drainage network inspection price. Most quotas give the comprehensive unit price (yuan / m) of the rainwater, sewage or rainwater and sewage combined pipeline dredging inspection costs. The difficulty of drainage network dredging inspection in practice varies greatly. Under a relatively fixed comprehensive unit price, the inspection progress will decrease with the difficulty of inspection, and some "difficult pipe sections" with high inspection difficulty are the core issues of urban drainage network management. It is difficult to effectively play the macro-management role of price, the "invisible hand", by adopting a relatively fixed comprehensive unit price for pipeline dredging inspection. Therefore, reasonable engineering volume pricing for pipeline inspection is an important guarantee for improving inspection efficiency and a practical problem that needs to be solved urgently. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method for predicting the difficulty of urban drainage pipe network detection based on entropy weight method and gray clustering, to construct a difficulty level assessment model based on gray clustering, to select key influencing factors by statistical survey method, to weight the key influencing factors by entropy weight method, and to conduct a comprehensive evaluation of the difficulty level of the selected factors by gray weighted clustering.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0005] The urban drainage pipe network detection difficulty prediction method based on entropy weight method and grey clustering includes the following steps: Step 1, determining the influencing factors of pipe network detection difficulty;

[0006] Step 2, determine the single indicator category boundary value of each influencing factor in Step 1;

[0007] Step 3, construct the whitening weight function and weight coefficient calculation model of each indicator in Step 2;

[0008] Step 4: Assign weights to each indicator in Step 3 and calculate the comprehensive weight coefficient matrix;

[0009] Step 5. Determine the difficulty level of each item.

[0010] In the above Step 1, the factors affecting the difficulty of pipe network detection include burial depth, sewage volume area ratio, siltation volume area ratio, pipe diameter and temperature.

[0011] The calculation method of sewage volume area ratio in Step 1 above is:

[0012] Assume that the height of the sewage layer is h2, the pipe diameter is d, and the sewage interface area is S. When 0≤h2≤d / 2:

[0013]

[0014] When d / 2<h2≤d:

[0015]

[0016] So, when 0≤h2≤d:

[0017]

[0018] Assume that the proportion of sewage volume to area is η 污水 ,but:

[0019]

[0020] The calculation method for the proportion of siltation area in the above Step 1 is:

[0021] Assume that the height of the sedimentation layer is h3, the pipe diameter is d, and the sewage volume area ratio is η 淤积 , when 0≤h3≤d, we have:

[0022]

[0023] According to formulas (4) and (5), the proportion of sewage volume area and siltation volume area can be obtained.

[0024] The calculation method of the whitening weight function in the above Step 3 is:

[0025] The whitening weight function is divided into three types: high class, medium class and low class. The mathematical model is as follows: Assume that the lower limit of the high class is H, the middle limit of the medium class is Z, the upper limit of the low class is L, and record d ij is the observed value of the i-th evaluation object with respect to the j-th indicator; is the whitening weight function value of the j index k subclass; the gray class whitening weight function and category weight coefficient calculation formula are as follows:

[0026] High Type:

[0027]

[0028] Medium type:

[0029]

[0030] Low Type:

[0031]

[0032] The process of calculating the indicator weights and comprehensive weight coefficient matrix in the above Step 4 is as follows:

[0033] Grey clustering is used for weighted clustering. The comprehensive clustering of each single indicator adopts the entropy weight method. The scores of m experts for p indicators are selected to form an m×p matrix, which is defined as the original matrix X ij , where X ij It refers to the score of the i-th expert on the j-th indicator. The matrix is ​​as follows:

[0034]

[0035] Standardize the data:

[0036] Let the score of the j index of the i-th expert be x ij (i=1,2,...,n;j=1,2,...,p), perform standardization to obtain X i ' j Matrix, where:

[0037]

[0038] Calculate the information entropy of each indicator:

[0039] The information entropy H of the jth index j It can be calculated by the following formula:

[0040]

[0041] Calculate the weight of each factor:

[0042]

[0043] That is, the weight of each indicator in the comprehensive evaluation is obtained; the comprehensive evaluation coefficient of each evaluation object is calculated:

[0044]

[0045] Then the i-th evaluation object belongs to the k-th * A class.

[0046] The present invention provides a method for predicting the difficulty of urban drainage pipe network detection based on entropy weight method and grey clustering. By constructing a difficulty level assessment model based on grey clustering, a statistical survey method is used to select key influencing factors, the entropy weight method is used to weight the key influencing factors, and grey weighted clustering is used to conduct a comprehensive evaluation of the difficulty level of the selected factors. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0048] Figure 1 This is a schematic diagram of the algorithm flow of the model of the present invention;

[0049] Figure 2 Schematic diagram of factors affecting the difficulty of urban drainage pipe network detection in the present invention;

[0050] Figure 3 This is a schematic diagram of the proportion of sewage volume and siltation area in the urban drainage network;

[0051] Figure 4 Schematic diagram of three types of whitening weight function images. DETAILED DESCRIPTION

[0052] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings and embodiments.

[0053] Description of indicators that affect the difficulty of drainage network detection: Assume that the set of indicators that affect the difficulty of drainage network detection is X = {X1, X2, X3, ..., X i ,...,X p}, where X i is the ith indicator.

[0054] Description of difficulty level comment set: Let difficulty level comment set C = {C1, C2, C3, ..., C j ,...,C s}, where C j is the i-th comment. The comment set is language data, described by three language evaluation values: “difficult”, “medium”, and “easy”.

[0055] Design of algorithm for urban drainage network detection difficulty level assessment model based on grey clustering:

[0056] The main factors affecting the difficulty of pipe network inspection were determined through investigation, and the difficulty level was determined by using grey weighted clustering. The algorithm flow is as follows: Figure 1 As shown, the main steps are as follows:

[0057] Step 1 Determine the factors affecting the difficulty of pipe network inspection;

[0058] Step 2 Determine the single indicator category boundary value of each influencing factor;

[0059] Step 3 constructs the whitening weight function and weight coefficient calculation formula of each index;

[0060] Step 4 assign weights to each indicator and calculate the comprehensive weight coefficient matrix;

[0061] Step 5: Determine the difficulty level of each item.

[0062] Determine influencing factors and category boundaries:

[0063] There are many factors that affect the difficulty of urban drainage network inspection. This study is based on the management practice of urban pipe network repair projects in Wuhu City, Anhui Province. Through questionnaire surveys and expert interviews, it is found that the main factors affecting the difficulty of urban drainage network inspection are burial depth ( Figure 2 (a)) Sewage volume area ratio ( Figure 2 (b)) and the proportion of sedimentation area ( Figure 2 (c)), pipe diameter, and temperature.

[0064] ① Calculation of sewage volume and area ratio

[0065] like Figure 2 As shown in (b), let the height of the sewage layer be h2, the pipe diameter be d, and the shaded area be S. When 0≤h2≤d / 2:

[0066]

[0067] When d / 2<h2≤d:

[0068]

[0069] So, when 0≤h2≤d:

[0070]

[0071] Assume that the proportion of sewage volume to area is η 污水 ,but:

[0072]

[0073] ② Calculation of the proportion of sedimentation area

[0074] like Figure 2 As shown in (c), let the height of the sedimentation layer be h3, the pipe diameter be d, and the sewage volume area ratio be η 淤积 , when 0≤h3≤d, according to the derivation process of ①, we can get:

[0075]

[0076] According to formulas (4) and (5), the proportion of sewage volume area and sedimentation volume area can be obtained as follows: Figure 3 shown.

[0077] Combining the characteristics of each indicator and the quantitative scale standard of a single indicator, the category boundaries of each indicator were determined, as shown in Table 1.

[0078] Table 1 Category boundary values ​​of each indicator

[0079]

[0080] Construct the whitening weight function of each indicator:

[0081] Generally, the whitening weight function is divided into three basic types, and its mathematical model is as follows: Assume that the lower limit of the high class is H, the middle limit of the middle class is Z, and the upper limit of the low class is L, and record d ij is the observed value of the i-th evaluation object with respect to the j-th indicator; is the whitening weight function value of the j index k subclass; the gray class whitening weight function and category weight coefficient calculation formula are as follows:

[0082] ①High type

[0083]

[0084] ②Medium type

[0085]

[0086] ③Low type

[0087]

[0088] Three types of whitening weight function images are as follows Figure 4 As shown, the whitening weight function value u of different gray classes of each indicator is calculated according to the whitening weight function ijk .

[0089] Indicator weight calculation and comprehensive clustering:

[0090] The comprehensive clustering of each single indicator in the grey clustering adopts the fixed weight clustering based on the entropy weight method. The scores of m experts on p indicators are selected to form an m×p matrix, which is defined as the original matrix X ij , where X ij It refers to the score of the i-th expert on the j-th indicator.

[0091]

[0092] ① Standardize the data

[0093] Let the score of the j index of the i-th expert be x ij (i=1,2,...,n;j=1,2,...,p), perform standardization to obtain X i' j Matrix; where:

[0094]

[0095] ② Calculate the information entropy of each indicator

[0096] The information entropy H of the jth index j It can be calculated by the following formula:

[0097]

[0098] ③Calculate the weight of each factor

[0099]

[0100] The weight of each indicator in the comprehensive evaluation is obtained. Calculate the comprehensive evaluation coefficient of each evaluation object:

[0101]

[0102] Then the i-th evaluation object belongs to the k-th * A class.

[0103] Example:

[0104] In order to test the feasibility of this research method for classifying the difficulty level of urban drainage network inspection, 100 pipe sections in the urban drainage network inspection project in Wuhu City, Anhui Province were selected, and their burial depth, sewage area ratio, siltation area ratio, pipe diameter, and pipe temperature were investigated respectively, and 5 index values ​​of each inspected pipe section were obtained. Due to space limitations, only the index values ​​of the first 10 pipe sections are listed here, as shown in the table; according to the basic steps of this research method, grey variable weight clustering was performed on the inspection items of each pipe section, and the comprehensive inspection difficulty level of each pipe section was analyzed.

[0105] Table 2 Index values ​​of each pipe section

[0106]

[0107] 1) Determine the whitening weight function of the single indicator measurement

[0108] According to the above steps, the whitening weight functions of the five indicators are first obtained, among which the temperature inside the tube is a mixed polarity indicator, and its weight function calculation formula is as follows:

[0109]

[0110] ②Medium type

[0111] ③Low type

[0112] 2) Calculate the single indicator measurement matrix

[0113] Substitute the category boundary values ​​of burial depth, sewage area ratio, siltation area ratio, pipe diameter, and pipe temperature into the above-mentioned whitening weight function and weight coefficient calculation formula, and substitute the index value of each project to obtain the weight coefficient matrix. The table shows the weight coefficients corresponding to the three different levels of difficulty, medium, and easy of the five indicators. Due to space limitations, only the weight coefficients of the first 10 pipe sections are listed here.

[0114] Table 3 Statistics of weight coefficients of different gray categories of five indicators

[0115]

[0116] 3) Determine the weight of each indicator

[0117] According to the description above, 10 experts were selected to score the five indicators to form a 10×5 original scoring table, as shown in the table. After standardization according to formula (10-25), X' was obtained.

[0118] Table 4 Original expert scoring table

[0119]

[0120]

[0121] Table 5 Standardized expert scoring table

[0122]

[0123] The information entropy of each indicator is calculated, and the weights of the five indicators are obtained, as shown in Table 6.

[0124] Table 6 Information entropy of five indicators

[0125]

[0126] 4) Determine the difficulty level

[0127] The weight coefficients corresponding to the three gray classes of the five indicators of a pipeline are used as row vectors to form the weight coefficient matrix of the project. Combined with the weights of the five indicators determined in the previous article, the U under the kth gray class is calculated respectively. ik , get the comprehensive weight coefficient, and judge the gray class to which the project belongs according to the maximum value of the row vector of each project. The evaluation results are shown in the table. Due to space limitations, only the difficulty level assessment results of the first 10 pipe sections are listed here.

[0128] It can be seen from Table 7 that among the 100 pipe sections, 69 have a difficulty level of "difficult", 4 have a difficulty level of "medium", and 27 have a difficulty level of "easy". Therefore, in the process of pipe network inspection and billing, the pipe sections with a difficulty level of difficult can be multiplied by an adjustment coefficient α greater than 1 on the basis of the comprehensive unit price according to actual requirements; the pipe sections with an easy difficulty level can be multiplied by an adjustment coefficient β less than 1 on the basis of the comprehensive unit price.

[0129] Table 7 Difficulty level assessment results

[0130]

Claims

1. A method for predicting the difficulty of urban drainage network detection based on entropy weight method and grey clustering, characterized in that: The steps include: Step 1. Determine the factors affecting the difficulty of pipe network inspection; Step 2, determine the single indicator category boundary value of each influencing factor in Step 1; Step 3, construct the whitening weight function and weight coefficient calculation model of each indicator in Step 2; Step 4: Assign weights to each indicator in Step 3 and calculate the comprehensive weight coefficient matrix; Step 5. Determine the difficulty level of each item.

2. The urban drainage network detection difficulty prediction method based on entropy weight method and grey clustering according to claim 1 is characterized in that: In the above-mentioned Step 1, the factors affecting the difficulty of pipe network detection include burial depth, sewage volume area ratio, siltation volume area ratio, pipe diameter and temperature.

3. The method for predicting the difficulty of urban drainage network detection based on entropy weight method and grey clustering according to claim 2 is characterized in that: The calculation method of sewage volume area ratio in Step 1 is: Assume that the height of the sewage layer is h2, the pipe diameter is d, and the sewage interface area is S. When 0≤h2≤d / 2: When d / 2<h2≤d: So, when 0≤h2≤d: Assume that the proportion of sewage volume to area is η 污水 ,but:

4. The method for predicting the difficulty of urban drainage network detection based on entropy weight method and grey clustering according to claim 3 is characterized in that: The calculation method of the siltation area ratio in Step 1 is: Assume that the height of the sedimentation layer is h3, the pipe diameter is d, and the sewage volume area ratio is η 淤积 , when 0≤h3≤d, we have: According to formulas (4) and (5), the proportion of sewage volume area and siltation volume area can be obtained.

5. The method for predicting the difficulty of urban drainage network detection based on entropy weight method and grey clustering according to claim 4 is characterized in that: The calculation method of the whitening weight function in Step 3 is: The whitening weight function is divided into three types: high class, medium class and low class. The mathematical model is as follows: Assume that the lower limit of the high class is H, the middle limit of the medium class is Z, the upper limit of the low class is L, and record d ij is the observed value of the i-th evaluation object with respect to the j-th indicator; is the whitening weight function value of the j index k subclass; the gray class whitening weight function and category weight coefficient calculation formula are as follows: High Type: Medium type: Low Type:

6. The method for predicting the difficulty of urban drainage network detection based on entropy weight method and grey clustering according to claim 5 is characterized in that: The process of calculating the indicator weights and the comprehensive weight coefficient matrix in Step 4 is as follows: Grey clustering is used for weighted clustering. The comprehensive clustering of each single indicator adopts the entropy weight method. The scores of m experts for p indicators are selected to form an m×p matrix, which is defined as the original matrix X ij , where X ij It refers to the score of the i-th expert on the j-th indicator. The matrix is ​​as follows: Standardize the data: Let the score of the j index of the i-th expert be x ij (i=1,2,...,n;j=1,2,...,p), perform standardization to obtain X i ' j Matrix, where: Calculate the information entropy of each indicator: The information entropy H of the jth index j It can be calculated by the following formula: Calculate the weight of each factor: That is, the weight of each indicator in the comprehensive evaluation is obtained; the comprehensive evaluation coefficient of each evaluation object is calculated: Then the i-th evaluation object belongs to the k-th * A class.

Images