A method for optimizing the layout of railway bridge holes based on an expert knowledge base

Abstract

The present invention relates to the field of layout of spans of railway bridges, and specifically discloses an optimization method for layout of spans of railway bridges based on an expert knowledge base, including: constructing an expert knowledge base; determining the layout range through the longitudinal section of the line, the horizontal single, and the bridge starting height; determining the clearance requirements and weights of each control point; starting from the control point with the highest weight, selecting the span with the smallest beam length that meets the clearance requirements in the expert knowledge base and arranging it at the control point; using standard simply supported beams for layout adjustment between adjacent control points; calculating the layout of curves to determine the accurate positions of piers and abutments; when spans are arranged throughout the bridge range, obtaining a span scheme, calculating the weight score and relevant costs of this scheme, repeating the above process to obtain different combined schemes; selecting the scheme with the highest weight score and the lowest cost from various schemes, and outputting the layout of spans of the bridge. The present invention is based on an expert knowledge base, combines different schemes, and automatically optimally selects the scheme with reasonable spans and the lowest cost.

Description

Technical Field

[0001] The present invention relates to the technical field of railway bridge span layout, and particularly relates to an optimization method for railway bridge span layout based on an expert knowledge base. Background Art

[0002] With the development of modern social transportation, the addition of road interchanges has led to a large number of interchange bridge constructions, which involves the layout of bridge spans, that is, the division of bridge spans or the selection of bridge pier positions. It is an important content in the overall design of bridges or the architectural design of bridges.

[0003] The layout of railway bridge spans is the primary task in bridge design. Manual span layout has problems such as large workload, low efficiency, and difficulty in finding the most cost-saving solution. Therefore, there is an urgent need for a method that can automatically select the most reasonable and cost-saving span layout plan. Summary of the Invention

[0004] To solve the problems existing in the prior art, the present invention provides an optimization method for railway bridge span layout based on an expert knowledge base. By combining field survey data with line horizontal and vertical data, the layout range and clearance requirements of control points are determined; different weights are given to each control point according to its importance; according to the required clearance, the corresponding span is selected from the expert knowledge base and arranged at the control point; for the section between two control points, standard simply supported beams are used for layout adjustment to obtain a span layout plan, and the weight score and related costs of this plan are calculated; repeat the above process of selecting different spans from the expert knowledge base to obtain different combined plans; select the plan with the highest weight score and the lowest cost from various plans. The present invention is based on an expert knowledge base, combines different plans, and automatically selects the most reasonable and cost-saving optimized span layout plan, solving the problems mentioned in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solution: An optimization method for railway bridge span layout based on an expert knowledge base, comprising the following steps:

[0006] S1. Construct an expert knowledge base;

[0007] S2. Determine the layout range through the line vertical section, horizontal single, and bridge starting height;

[0008] S3. Determine the clearance requirements of each control point;

[0009] S4. Determine the weight of each control point;

[0010] S5. Load the expert knowledge base, starting from the control point with the highest weight coefficient, select the span with the minimum beam length that meets the clearance requirements in the expert knowledge base, and arrange it at the control point;

[0011] S6. Select adjacent control point hole spans for layout, and use standard simply supported beams to conduct layout adjustment between adjacent control points to obtain the optimal simply supported beam combination for adjacent control points;

[0012] S7. According to the curve of the longitudinal section of the line, determine the accurate positions of piers and abutments through curve layout calculation by the method of bisecting the middle sag;

[0013] S8. When the hole spans are laid out in the entire bridge range, a hole span scheme is obtained. Calculate the weight score and relevant costs of this scheme, and repeat the above process to obtain different combination schemes;

[0014] S9. Select the scheme with the highest weight score and the lowest cost from various schemes, and output the bridge hole span layout.

[0015] Preferably, in step S1, construct an expert knowledge base, including the following steps:

[0016] S11. Sort out the bridge span layout schemes of existing projects, especially the special structures of crossing structures;

[0017] S12. Classify according to different regional geology, including the loess plain area and the rocky hilly area;

[0018] S13. Cover the main information of hole span patterns, main clear spans, single and double tracks, pier heights, longitudinal and transverse widths of piers, longitudinal and transverse widths of pile caps, and construction costs;

[0019] S14. Construct an expert knowledge base with complete bridge span patterns, including simply supported beams, continuous beams, rigid frames, tied arches, and portal piers.

[0020] Preferably, in step S3, the determination of the clearance requirements for each control point specifically includes: combining the three-dimensional ground model and the line horizontal and vertical data, automatically obtaining the existing control point information. When the distance between adjacent control points is less than 32 m, it is considered as one control point, and it is supplemented, corrected, and improved through field survey data including planning and relocation survey information to obtain the single and double tracks, straight and curved lines, pier heights, clearance L j需 and the data of the skew angle θ required for each control point.

[0021] Preferably, in step S4, the determination of the weights of each control point specifically includes: giving different weights to each control point according to the importance of the crossed terrain type, and its characteristic is that the higher the level, the higher the weight that is more difficult to change.

[0022] Preferably, in step S5, the hole spans that meet the clearance requirements specifically include:

[0023] For the case where the pile cap cannot intrude into the space below the clearance width when crossing the railway, calculate the clear span value L according to the following formula j :

[0024] L j = L - C - D×tan(θ);

[0025] L j represents the calculated net span value, L represents the main span length, C represents the longitudinal width of the bearing platform, D represents the transverse width of the bearing platform, and θ represents the skew angle between the line and the structure;

[0026] For the case where the bearing platform intrudes into the space under the net width when crossing a general road or pipeline, the net span value L is calculated according to the following formula j :

[0027] L j = L - A - B×tan(θ);

[0028] L j represents the calculated net span value, L represents the main span length, A represents the longitudinal width of the pier, B represents the transverse width of the pier, and θ represents the skew angle between the line and the structure;

[0029] Calculate the net span value L provided by each hole - span pattern j , according to the pier height, single - or double - track, and the required clearance L j需 factors, select the hole - span that meets L j需 ≤L j in the expert knowledge base, that is, the hole - span that meets the clearance requirements.

[0030] Preferably, in step S6, it specifically includes the following: Adopt a combination of multiple standard simply - supported beams of 16m, 20m, 24m, and 32m for layout adjustment, and the layout principles are as follows:

[0031] d) ΔL = the net distance between two control points - the sum of the beam lengths of the simply - supported beam layout spans, and the smaller ΔL is, the better;

[0032] e) Try to use a larger - span layout;

[0033] f) The smaller the difference in beam spans between adjacent holes - spans is, the better;

[0034] According to the above principles, use a policy - based reinforcement learning algorithm to obtain the optimal simply - supported beam combination between adjacent control points, and the formula is expressed as follows:

[0035] Q t (s,a) = Q t-1 (s,a)+ɑ(R(s,a)+γmaxQ(s’,a’)-Q t-1 (s,a))

[0036] where, Q t (s,a) is the Q - value to be updated corresponding to (s,a) in the t - th process, Q t-1(s, a) is the Q value corresponding to (s, a) in the t-1 process, where s represents the state, a represents the action, ɑ represents the learning rate, R(s, a) is the reward value corresponding to (s, a), γ represents the discount factor, s’ represents the previous state, a’ represents the previous action, and maxQ(s’, a’) represents the maximum reward that can be obtained after the previous action.

[0037] Preferably, in step S8, the calculation of the scheme weight score and related costs is expressed by the following formula:

[0038]

[0039] Among them, V(s1) is the total score of the scheme weight coefficient, n is the total number of control points, β is the coefficient when passing or not passing, k is the control point number, and R(k, β) is the weight coefficient obtained by the control point; C(s1) is the total cost of the scheme, m is the total number of spans of the whole bridge, j is the span number, and E(j) is the cost corresponding to the j-th span.

[0040] The beneficial effects of the present invention are as follows: By constructing an expert knowledge base, the method of the present invention controls each control point, adjusts with a standard simply supported beam between two control points, determines the accurate positions of piers and abutments through curve layout calculation on the curve, and through the weight score and related costs of the obtained multiple combination schemes, it can not only greatly improve the efficiency of bridge layout, but also obtain the most reasonable and lowest-cost scheme for span layout. Description of the Drawings

[0041] Figure 1 It is a schematic flow chart of the steps of the railway bridge span layout optimization method based on an expert knowledge base in the embodiment of the present invention. Detailed Embodiments

[0042] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0043] Please refer to Figure 1 , the present invention provides a technical solution: A railway bridge span layout optimization method based on an expert knowledge base, including the following steps:

[0044] S1. Construct an expert knowledge base, including the following steps:

[0045] S11. Sort out the bridge span layout schemes of existing projects, especially the special structures of crossing structures;

[0046] S12. Classify according to different regional geology, including, for example, the loess plain area and the rocky hilly area;

[0047] S13. Cover the main information including the span pattern, main clear span, single / double track, pier height, longitudinal and transverse widths of the pier, longitudinal and transverse widths of the bearing platform, construction cost, etc.;

[0048] S14. Build an expert knowledge base with perfect bridge span patterns, including simply supported beams, continuous beams, rigid frames, tied arches, portal piers, etc.

[0049] To obtain more accurate optimization results, make corrections according to the characteristics of different projects, such as information on pier bodies, bearing platform sizes, construction costs, etc. For reference to the bridge span expert knowledge base in the loess plain area, see Table 1 below.

[0050] Table 1 Reference for the bridge span expert knowledge base in the loess plain area

[0051]

[0052]

[0053] S2. Determine the layout range through the line vertical profile, horizontal single, and bridge starting height; determine the mileage of the two abutments.

[0054] S3. Determine the clearance requirements for each control point, specifically including: combining the 3D ground model and the line horizontal and vertical data to automatically obtain the existing control point information. When the distance between adjacent control points is less than 32m, consider them as one control point, and supplement, correct, and improve through field survey data including planning and relocation survey information to obtain the single / double track, straight / curved line, pier height, clearance L j需 and the data of the skew angle θ for each control point.

[0055] S4. Determine the weights of each control point, specifically including: giving different weights to each control point according to the importance of the crossed terrain type. Its characteristic is that the higher the level, the higher the weight of the more difficult-to-change ones. The importance weight table of bridge crossing control points is shown in Table 2.

[0056] Table 2 Importance weight coefficient table of bridge crossing control points

[0057] Serial number Control point type Weight coefficient Remarks 1 High-speed railway 10 2 Regular-speed railway 9 3 Special railway line 8 4 Great rivers and big rivers 7 5 Expressway 6 6 Important oil and gas pipelines 5 7 National highways and provincial highways 4 8 Important main canals 3 9 General rivers and ditches 2 10 General pipelines and roads 1

[0058] S5. Import the expert knowledge base. Starting from the control point with the highest weight coefficient, select the span with the smallest beam length that meets the clearance requirements in the expert knowledge base and layout it to the control point; the expert knowledge base contains various span forms and their costs for crossing control points with existing experience, and continuously supplements and improves as the design experience increases. The more span patterns it contains, the more likely the obtained layout plan will be more economical and reasonable, thus obtaining a more optimized span combination.

[0059] The hole spans that meet the clearance requirements specifically include:

[0060] For the case where the bearing platform cannot intrude into the space below the clear width range when crossing a railway, the net span value L is calculated according to the following formula j :

[0061] L j = L - C - D×tan(θ);

[0062] L j represents the calculated net span value, L represents the main span length, C represents the longitudinal width of the bearing platform, D represents the transverse width of the bearing platform, and θ represents the skew angle between the line and the structure;

[0063] For the case where the bearing platform intrudes into the space below the clear width range when crossing a general road or pipeline, the net span value L is calculated according to the following formula j :

[0064] L j = L - A - B×tan(θ);

[0065] L j represents the calculated net span value, L represents the main span length, A represents the longitudinal width of the pier, B represents the transverse width of the pier, and θ represents the skew angle between the line and the structure;

[0066] Calculate the net span value L provided by each hole span pattern, and select the hole span that meets the clearance requirements from the expert knowledge base according to the pier height, single / double track, and the required clearance L j , considering the factors of pier height, single / double track, and the required clearance L j需 such that the selected hole span satisfies L j需 ≤ L j That is, the hole spans that meet the clearance requirements.

[0067] S6. Select the adjacent control point hole spans and use standard simply supported beams to adjust the layout between adjacent control points to obtain the optimal simply supported beam combination for adjacent control points;

[0068] Specifically, it includes the following: Use a combination of multiple standard simply supported beams with lengths of 16m, 20m, 24m, and 32m for layout adjustment, and the layout principles are as follows:

[0069] g) ΔL = the net distance between two control points - the sum of the layout beam lengths of the simply supported beams. The smaller ΔL is, the better;

[0070] h) Try to use a larger span for layout;

[0071] i) The smaller the difference between the beam spans of adjacent holes, the better;

[0072] According to the above principles, use a policy-based reinforcement learning algorithm to obtain the optimal simply supported beam combination for adjacent control points, and the formula is expressed as follows:

[0073] Q t(s,a) = Q t-1 (s,a) + ɑ(R(s,a) + γmaxQ(s’,a’) - Q t-1 (s,a))

[0074] where Q t (s,a) is the Q-value to be updated for (s,a) corresponding to process t, and Q t-1 (s,a) is the Q-value for (s,a) corresponding to process t-1, s represents the state, a represents the action, ɑ represents the learning rate, R(s,a) is the reward value corresponding to (s,a), γ represents the discount factor, s’ represents the previous state, a’ represents the previous action, and maxQ(s’,a’) represents the maximum reward that can be obtained after the previous action.

[0075] If the requirement of the control point L j需 ≤ L j cannot be satisfied by any simply supported beam combination, the weight score of this control point is 0, and the layout of the hole spans of adjacent control points is continued.

[0076] S7. According to the curve of the longitudinal section of the line, the accurate positions of the piers and abutments are determined by calculating the curve layout through the method of bisecting the middle ordinate;

[0077] S8. When the hole spans are arranged throughout the bridge, a hole span scheme is obtained. Calculate the weight score and related costs of this scheme. Based on the previous scheme, for the control point with the highest weight coefficient among those with a weight score of 0, select a larger L j hole span pattern for layout to make it satisfy L j需 ≤ L j condition to obtain a new hole span layout scheme; repeat the above process to obtain different combination schemes;

[0078] The calculation of the weight score and related costs of the scheme is expressed by the following formula:

[0079]

[0080] where V(s1) is the total score of the scheme weight coefficient, n is the total number of control points, β is the coefficient when passing or not passing, k is the control point number, and R(k,β) is the weight coefficient obtained by the control point; C(s1) is the total cost of the scheme, m is the total number of spans of the whole bridge, j is the hole span number, and E(j) is the cost corresponding to the j-th span.

[0081] S9. Select the scheme with the highest weight score and the lowest cost from various schemes, and output the bridge hole span layout. The formula is as follows: F = Max(V(s1,s1,s1...,s n )).and.Min(C(s1,s1,s1...,s m ))。

[0082] Based on the expert knowledge base, the present invention conducts control on each control point, uses a standard simply supported beam to adjust between two control points, determines the accurate positions of piers and abutments through curve layout calculation when on a curve, and through weight scores and relevant costs for various obtained combination schemes, can not only greatly improve the efficiency of bridge layout, but also obtain the most reasonable and lowest-cost scheme for span layout.

[0083] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for optimizing the layout of railway bridge holes based on an expert knowledge base, characterized in that, It includes the following steps: S1. Construct an expert knowledge base; S2. Determine the layout range based on the line vertical profile, horizontal single, and bridge starting height; S3. Determine the clearance requirements for each control point; S4. Determine the weights of each control point; S5. Load the expert knowledge base. Starting from the control point with the highest weight coefficient, select the smallest-span hole that meets the clearance requirements and has the minimum beam length in the expert knowledge base, and lay it out to the control point. The hole spans that meet the clearance requirements specifically include: For the case where the bearing platform cannot intrude into the space under the clear width range across the railway, calculate the clear span value L according to the following formula j :[[]]END]] L j = L - C - D × tan(θ); L j It represents the calculated clear span value, L represents the main span length, C represents the longitudinal width of the pile cap, D represents the transverse width of the pile cap, and θ represents the skew angle between the line and the structure; For the case where the bearing platform invades the space below the clear width range across a general road or pipeline, calculate the clear span value L according to the following formula j :[[]]END]] L j = L - A - B × tan(θ); L j represents the calculated clear span value, L represents the main span length, A represents the longitudinal width of the pier, B represents the transverse width of the pier, and θ represents the skew angle between the line and the structure; Calculate the net span value L provided by each hole span pattern j , based on the pier height, single / double track, and the required clearance L j需 factors, select a hole span that meets L j需 ≤L j from the expert knowledge base, that is, the hole span that meets the clearance requirement; S6. Select the hole spans of adjacent control points for layout. Use standard simply supported beams to adjust the layout between adjacent control points to obtain the optimal simply supported beam combination for adjacent control points. Specifically, it includes the following: Use a variety of standard simply supported beam combinations of 16m, 20m, 24m, and 32m for layout adjustment. The layout principles are as follows: a) ΔL = the net distance between two control points - the sum of the beam lengths of the simply supported beam layout spans. The smaller ΔL is, the better; b) Try to use larger-span layouts; c) The smaller the difference in beam spans between adjacent holes is, the better; According to the above principles, use a policy-based reinforcement learning algorithm to obtain the optimal simply supported beam combination for adjacent control points. The formula is expressed as follows: Q t (s,a) = Q t-1 (s,a) + ɑ(R(s,a) + γmaxQ(s’,a’) - Q t-1 (s,a)) Among them, Q t (s,a) is the Q value to be updated for (s,a) corresponding to the t process, and Q t-1 (s,a) is the Q value of (s,a) corresponding to the t-1 process, s represents the state, a represents the action, ɑ represents the learning rate, R(s,a) is the reward value corresponding to (s,a), γ represents the discount factor, s’ represents the previous state, a’ represents the previous action, and maxQ(s’,a’) represents the maximum reward that can be obtained next after the previous action; S7. According to the curve of the line vertical profile, calculate and determine the accurate positions of the piers and abutments through the mid-ordinate bisection method for curve layout; S8. When the hole spans are laid out throughout the bridge range, a hole span plan is obtained. Calculate the weight score and relevant costs of this plan, and repeat the above process to obtain different combination plans; S9. Select the plan with the highest weight score and the lowest cost from various plans and output the bridge hole span layout.

2. The method for optimizing the layout of railway bridge holes based on an expert knowledge base according to claim 1, wherein: In step S1, when constructing the expert knowledge base, it includes the following steps: S11. Sort out the bridge span layout plans of existing projects, including the special structures of crossing structures; S12. Classify according to different regional geology, including the loess plain area and the rocky hilly area; S13. Cover the main information of hole span patterns, main clear spans, single or double tracks, pier heights, longitudinal and transverse widths of piers, longitudinal and transverse widths of pile caps, and construction costs; S14. Construct an expert knowledge base with complete bridge span patterns, including simply supported beams, continuous beams, rigid frames, tied arches, and portal piers.

3. The method for optimizing the pier hole layout of a railway bridge based on an expert knowledge base according to claim 1, wherein: In step S3, the determination of the clearance requirements for each control point specifically includes: combining the three-dimensional ground model and the horizontal and vertical line data to automatically obtain the existing control point information. When the distance between adjacent control points is less than 32 m, they are considered as one control point, and the supplementary correction and improvement are carried out through the field survey data including the survey information of planning and relocation, so as to obtain the single / double track, straight / curved line, pier height, and clearance L required for each control point. j需 And the data of the skew angle θ.

4. The method for optimizing the layout of railway bridge holes based on an expert knowledge base according to claim 1, characterized in that: In step S4, the determination of the weights of each control point specifically includes: Give different weights to each control point according to the importance of the crossing ground object type. Its characteristic is that the higher the level, the higher the weight that is more difficult to change.

5. The optimized method for arranging bridge openings of railway bridges based on an expert knowledge base according to claim 1, wherein: In step S8, the calculation of the plan weight score and relevant costs is expressed by the following formula: Among them, V(s1) is the total score of the plan weight coefficient, n is the total number of control points, β is the coefficient when passing or not passing, k is the control point serial number, R(k,β) is the weight coefficient obtained by the control point; C(s1) is the total construction cost of the plan, m is the total number of spans of the whole bridge, j is the hole span serial number, and E(j) is the construction cost corresponding to the j-th span.

Images