Earth and rockfill dam compaction quality optimal sampling scheme determination method based on counting sampling principle

Abstract

The invention discloses an earth and rockfill dam compaction quality optimal sampling scheme determination method based on a counting sampling principle. The method comprises the following steps: determining a first statistical model under the principle of preferentially protecting a transfer party and a second statistical model giving consideration to acceptance risk control of the construction party and the transfer party according to the risk tolerance of the construction party and the transfer party; carrying out simulation solution by adopting a Monte Carlo method to respectively obtain optimal sampling schemes based on different statistical models; and on the basis of the determined optimal sampling scheme, generating a discrete sampling point spatial distribution pattern diagram meeting the spatial aggregation state control requirement by utilizing a spatial arrangement optimization model of the discrete sampling points. On the basis that the detection cost is not remarkably increased, the probability that the statistical model has authenticity abandoning and authenticity accepting errors under the condition of limited small samples is effectively controlled, the spatial arrangement pattern is reasonably optimized from the spatial aggregation state angle of discrete sampling points, and the reasonable rights and interests of a construction party and a transfer party are fully guaranteed.

Description

Technical Field

[0001] The invention relates to the technical field of water conservancy and hydropower engineering, and in particular to a method for determining an optimal sampling scheme for compaction quality of an earth-rock dam based on a counting sampling principle. Background Art

[0002] The rolling construction process is an important link in the quality control of earth-rock dam construction. The quality of its construction will directly affect the final realization of the structural performance indicators of earth-rock dam such as density, strength, and stiffness, and is related to the long-term operation safety of earth-rock dam. According to the current construction specification of rolled earth-rock dam in my country (DL / T 5129-2013), the compaction quality control of earth-rock dam mainly relies on density detection methods based on discrete sampling points (such as sand filling method, additional mass method, nuclear density meter method, etc.). However, due to the detection cost and its interference with the continuous construction process, the actual discrete sampling point inspection sample data obtained is very limited, generally less than 1% of the entire construction warehouse surface. Therefore, the compaction quality assurance of earth-rock dam is limited by the limited small sample data of discrete sampling points, and its detection accuracy is also significantly affected.

[0003] Generally speaking, the random sampling process of the compaction quality of earth-rock dams can be generalized to conform to the normal distribution form, and the compaction quality detection index is established accordingly, and the compaction quality of the entire construction bin surface is macro-controlled with a sampling failure rate of 10%.

[0004] However, the current index system established based on the normal distribution population may not be able to adapt well to the compaction quality acceptance under the condition of limited small samples. In practical application, there are the following problems:

[0005] First, under the condition of finite small samples, the random sampling process of the compaction quality of earth-rock dams should actually be attributed to the hypergeometric distribution problem based on the counting sampling principle. The hypergeometric distribution problem has an essential difference in the way it considers the true rejection error (referring to the situation that the sample is rejected according to the sampling inspection standard but is actually qualified) and the false acceptance error (referring to the situation that the sample is accepted according to the sampling inspection standard but is actually unqualified) from the random sampling process after generalization with reference to the normal distribution problem.

[0006] Second, for discrete testing units, after limiting the sampling failure rate, the sampling failure control number in the sampling plan will be rounded down, causing the testing standard to fluctuate periodically as the sample size increases;

[0007] Third, after limiting the sampling failure rate, the range of sampling plans available is limited, which is not conducive to controlling the probability of statistical errors such as true rejection and false acceptance.

[0008] Fourth, the influence of spatial autocorrelation of discrete sampling points under spatial aggregation state on the accuracy of characterization of compaction quality within the construction warehouse surface is not considered. Summary of the invention

[0009] In view of the shortcomings of the prior art, the present invention proposes a method for determining the optimal sampling scheme for the compaction quality of earth-rock dams based on the counting sampling principle, which is used to optimize the sampling scheme based on discrete sampling points used in the compaction quality acceptance process during the construction period. Without significantly increasing the detection cost, the acceptance risk of the construction party and the transferee is effectively reduced, and the reasonable rights and interests of both parties are fully protected.

[0010] To achieve the above object, the present invention proposes a method for determining the optimal sampling scheme for compaction quality of earth-rock dam based on the counting sampling principle, which is particularly characterized in that it includes the following steps:

[0011] S1) According to the risk tolerance of the construction party and the transferee, the statistical model on which the best sampling scheme depends is determined. If the construction party has fully self-checked the compaction quality of the constructed warehouse surface and has sufficient confidence in the construction quality, the first statistical model under the principle of giving priority to protecting the transferee is adopted to avoid the long-term operation safety risk of the earth-rock dam caused by the occurrence of pseudo-errors. The first statistical model is specifically shown as follows:

[0012]

[0013] In the formula,

[0014] n is the sample size in the sampling scheme, that is, the number of discrete sampling points,

[0015] N is the number of independent inspection units divided in the construction warehouse.

[0016] T β is the sampling failure rate q in the construction warehouse that is acceptable to the handover party β The estimated number of unqualified units to be inspected,

[0017] (n, 0) means that the number of unqualified measurement points allowed to appear in the n discrete sampling points randomly selected in the sampling plan is 0.

[0018] β is the probability of a pseudo-error in sampling detection,

[0019] β0 is the control probability of pseudo-error in sampling detection,

[0020] q β is the sampling failure rate acceptable to the transferor,

[0021] is the number of combinations of n discrete sampling point units randomly selected from the N units to be inspected in the construction warehouse surface,

[0022] To NT from the construction warehouse β The number of combinations of n discrete sampling point units randomly selected from the qualified units to be inspected,

[0023] floor(·) is the floor rounding function;

[0024] If the construction party fails to conduct sufficient self-inspection on the compaction quality of the constructed warehouse surface, resulting in the construction party's lack of confidence in the construction quality, and at the same time, increasing the sampling inspection sample size n will not have a significant impact on the continuous construction process, and the construction party is willing to bear the additional inspection costs incurred by expanding the sampling inspection sample size n, the second statistical model that takes into account the acceptance risk control of both the construction party and the handover party is adopted. The second statistical model is specifically shown as follows:

[0025]

[0026] In the formula,

[0027] n is the sample size in the sampling plan, that is, the number of discrete sampling points,

[0028] m is the maximum number of unqualified measurement points allowed in the sampling plan,

[0029] N is the number of independent inspection units divided in the construction warehouse.

[0030] T β is the sampling failure rate q in the construction warehouse that is acceptable to the handover party β The estimated number of unqualified units to be inspected,

[0031] T α is the sampling failure rate q used by the construction party for self-inspection control in the construction warehouse α The estimated number of unqualified units to be inspected,

[0032] (n, m) means that the maximum number of unqualified measurement points allowed in the n discrete sampling points randomly selected in the sampling plan is m.

[0033] β is the probability of a pseudo-error in sampling detection,

[0034] β0 is the control probability of pseudo-error in sampling detection,

[0035] α is the probability of false negatives in sampling detection,

[0036] α0 is the control probability of false positive errors in sampling detection,

[0037] γ is the risk self-control reduction coefficient of the construction party,

[0038] q βis the sampling failure rate acceptable to the transferor,

[0039] q α The sampling failure rate used by the construction party for self-inspection control.

[0040] is the number of combinations of n discrete sampling point units randomly selected from the N units to be inspected in the construction warehouse surface,

[0041] T from the construction warehouse β The number of combinations of t discrete sampling point units randomly selected from the unqualified units to be inspected,

[0042] T from the construction warehouse α The number of combinations of t discrete sampling point units randomly selected from the unqualified units to be inspected,

[0043] To NT from the construction warehouse β The number of combinations of nt discrete sampling point units randomly selected from the qualified units to be inspected,

[0044] To NT from the construction warehouse α The number of combinations of nt discrete sampling point units randomly selected from the qualified units to be inspected,

[0045] floor(·) is the floor rounding function;

[0046] S2) Based on the selected first statistical model or second statistical model, a Monte Carlo method is used to perform simulation and solve, to obtain specific values ​​of the sample capacity n in the sampling plan and the maximum number m of unqualified measurement points allowed in the sampling plan, and to obtain the optimal sampling plan (n, 0) or (n, m) based on different statistical models respectively;

[0047] S3) Based on the determined optimal sampling scheme (n, 0) or (n, m), n discrete sampling points are randomly selected from the N independent units to be inspected in the construction warehouse surface, and the spatial coordinates (x i ,y i ) to generate a spatial distribution pattern map of discrete sampling points that meets the requirements of spatial aggregation state control, which is used to guide sampling inspection at earth-rock dam construction sites.

[0048] Furthermore, in step S2), the specific steps of using the Monte Carlo method to solve the first statistical model are as follows:

[0049] S2A1) Determine the sampling failure rate q of the compaction quality on the surface of the construction bin to be inspected that is acceptable to the handover party β, the entire construction warehouse is discretized into N independent inspection units, random numbers are generated by software, and T is randomly assigned in the N independent inspection units. β Units are defective and unqualified units;

[0050] S2A2) specifying a set of retrieval number pairs (n, 0) as an initial sampling scheme;

[0051] S2A3) randomly select n discrete sampling points from N independent units to be inspected in the construction warehouse surface. When the number of units to be inspected with defects in the n discrete sampling points is equal to 0, it is considered that the group of sampling results can pass the inspection;

[0052] S2A4) repeating step S2A3) for no less than 10,000 times, counting the frequency of occurrence of the sampling results that pass the test, and using the frequency as the probability β of a pseudo-error in the sampling test;

[0053] S2A5) setting the control probability β0 of the occurrence of nano-pseudo errors in sampling detection, when the probability of the occurrence of nano-pseudo errors in sampling detection obtained by simulation β>β0, the sample capacity n is increased by 1 sampling unit;

[0054] S2A6) Repeat steps S2A3) to S2A5) until the obtained sampling scheme satisfies the condition β≤β0. The sampling scheme at this time is the optimal sampling scheme (n, 0) under the principle of giving priority to protecting the transferred party.

[0055] Furthermore, in S2), on the basis of satisfying the first statistical model to effectively control the probability β of the occurrence of false positive errors in sampling detection, the second statistical model is solved by the Monte Carlo method, and the probability α of the occurrence of the occurrence of the false positive errors in sampling detection is additionally considered to be effectively controlled. The specific steps are as follows:

[0056] S2B1) Determine the sampling failure rate q used by the construction party for self-inspection control α , the entire construction warehouse is discretized into N independent inspection units, random numbers are generated by software, and T is randomly assigned in the N independent inspection units. α Units are defective and unqualified units;

[0057] S2B2) using the sampling scheme obtained by solving the first statistical model using the Monte Carlo method as a preliminary sampling scheme (n, m) for solving the second statistical model;

[0058] S2B3) randomly select n discrete sampling points from N independent units to be inspected in the construction warehouse surface. When the number of units to be inspected with defects among the n discrete sampling points selected is less than or equal to m, it is considered that the group of sampling results can pass the inspection;

[0059] S2B4) repeating step S2B3) for no less than 10,000 times, counting the frequency of occurrence of sampling results that fail the test, and using the frequency as the probability α of a true rejection error in the sampling test;

[0060] S2B5) setting the control probability of sampling detection to reject the true error α0, when the probability of sampling detection to reject the true error obtained by simulation α>α0, the maximum number m of unqualified measurement points allowed in the sampling plan is increased by 1 sampling unit;

[0061] S2B6) using the sampling plan adjusted in step S2B5) as the initial sampling plan (n, m);

[0062] S2B7) repeating step S2B3) for no less than 10,000 times, counting the frequency of occurrence of the sampling results that pass the test, and using the frequency as the probability β of a pseudo-error in the sampling test;

[0063] S2B8) is compared with the set control probability β0 of the occurrence of nano-pseudo-error in the sampling detection. When the simulated probability of the occurrence of nano-pseudo-error in the sampling detection β>β0, the sample capacity n is increased by 1 sampling unit;

[0064] S2B9) repeating steps S2B7) to S2B8) until the obtained sampling plan satisfies the condition β≤β0, and using the obtained sampling plan as the adjusted primary sampling plan (n, m) for solving the second statistical model;

[0065] S2B 10 ) Repeat steps S2B3)~S2B9) until the obtained sampling plan satisfies the conditions of β≤β0 and α≤α0 at the same time. The sampling plan at this time is the optimal sampling plan (n, m) that takes into account the acceptance risk control of both the construction party and the handover party.

[0066] Furthermore, in S2A5), the control probability β0 of a false positive error in sampling detection can generally be set to 10%; in S2B5), the control probability α0 of a false positive error in sampling detection can generally be set to 5%.

[0067] Furthermore, step S2) can be implemented by programming with Matlab program.

[0068] Furthermore, in step S3), the specific steps of obtaining the spatial distribution pattern diagram of discrete sampling points that meets the spatial aggregation state control requirements are as follows:

[0069] S31) According to the shape of the actual construction warehouse surface, the N independent units to be inspected are numbered one by one according to a certain sorting principle;

[0070] S32) Based on the optimal sampling scheme (n, 0) or (n, m) determined in step S2), n discrete sampling point units are randomly selected from the N independent units to be inspected in the construction warehouse surface, and the spatial coordinates (x ) of the center positions of the selected n discrete sampling points are determined according to the numbers of the discrete sampling point units. i ,y i );

[0071] S33) Using the determined spatial coordinates (x) of the center position of the discrete sampling point i ,y i ), calculate the nearest point index R of the optimal sampling scheme (n, 0) or (n, m) in the spatial distribution within the construction warehouse;

[0072] S34) judging whether the optimal sampling scheme (n, 0) or (n, m) meets the control requirements of the spatial aggregation state according to the size of the nearest point index R; the calculation model of the nearest point index R is as follows:

[0073]

[0074]

[0075] In the formula,

[0076] r e is the average distance between n discrete sampling points,

[0077] r T is the theoretical standard distance considering the regional boundary effect,

[0078] r i is the distance between the ith discrete sampling point and the nearest discrete sampling point,

[0079] n is the sample size in the optimal sampling scheme,

[0080] N is the number of independent inspection units discretely separated from the construction warehouse surface.

[0081] D is the theoretical density of n discrete sampling points evenly distributed within the construction warehouse.

[0082] B is the perimeter of the construction warehouse surface,

[0083] A is the area of ​​the construction warehouse;

[0084] S35) If the nearest point index R≤0.5, the n discrete sampling points are concentratedly distributed and do not meet the control requirements of the spatial aggregation state. Steps S32) to S34) need to be repeated until R>0.5 is satisfied and the n discrete sampling points are randomly or uniformly distributed. The sampling scheme at this time is the optimal sampling scheme after spatial layout optimization considering the distribution pattern of discrete sampling points.

[0085] The advantages of the present invention are:

[0086] 1. The present invention fully considers the different risk tolerance of the construction party and the transfer party, and adopts the first statistical model under the principle of giving priority to protecting the transfer party and the second statistical model taking into account the acceptance risk control of the construction party and the transfer party as the basis for solving the optimal sampling plan for the compaction quality of the earth-rock dam, thereby ensuring the reasonable rights and interests of the construction party and the transfer party;

[0087] 2. The present invention considers the discreteness of the detection units that obey the hypergeometric distribution, breaks through the limitation of fixed sampling failure rate in the existing specifications, and searches for the optimal sampling scheme that meets the statistical model conditions and has the minimum sample size n from the integer number pair (n, m), and its search range is wider;

[0088] 3. The present invention reasonably optimizes the spatial layout pattern from the perspective of the spatial aggregation state of discrete sampling points to prevent the discrete sampling points from being affected by their spatial autocorrelation due to spatial aggregation, resulting in reduced representation accuracy;

[0089] 4. The present invention provides an effective solution method for two types of statistical models through the Monte Carlo method, comprehensively considers the different needs of the construction party and the transfer party, and adjusts the risk control points (q 1-α ,1-α0),(q β , β0) automatically matches the optimal sampling scheme;

[0090] The method for determining the best sampling scheme for compaction quality of earth-rock dam based on the counting sampling principle of the present invention is stricter than the acceptance standard recommended by the current specification, and effectively controls the probability of two types of statistical errors, namely, true rejection and false acceptance, in the statistical model under the condition of limited small samples, and fully protects the reasonable rights and interests of the construction party and the transfer party; at the same time, the spatial layout pattern of the obtained best sampling scheme is reasonably optimized from the perspective of the spatial aggregation state of discrete sampling points, so as to prevent the discrete sampling points from being affected by their spatial autocorrelation due to spatial aggregation, and effectively ensure the rationality of the judgment of the acceptance statistical model; in addition, the present invention provides two types of statistical models and effective implementation methods for spatial layout optimization, which has the characteristics of flexible and convenient application and strong applicability, and is convenient for on-site implementation. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] Figure 1 It is an implementation flow chart of the method for determining the best sampling scheme for compaction quality of earth-rock dam based on the counting sampling principle of the present invention;

[0092] Figure 2 It is a spatial position distribution diagram of discrete sampling points of an optimized best sampling scheme in a specific implementation method. DETAILED DESCRIPTION

[0093] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0094] like Figure 1 As shown, the method for determining the optimal sampling scheme for compaction quality of earth-rock dam based on the counting sampling principle of the present invention comprises the following steps:

[0095] S1) The construction party and the transfer party negotiate and analyze their respective risk tolerance. The statistical model on which the best sampling plan depends is determined based on the consensus reached by both parties on the confidence of the compaction quality of the construction bin surface passing the inspection, the possible subsequent impact of the two types of statistical errors, and the construction party's willingness to bear the additional costs of additional testing. The specific statistical model selection is based on the following:

[0096] If the construction party has fully self-checked the compaction quality of the constructed warehouse surface and has sufficient confidence in the construction quality, the first statistical model under the principle of giving priority to protecting the transferee is adopted to avoid the long-term operation safety risk of the earth-rock dam caused by the occurrence of false positive errors. The first statistical model is specifically shown as follows:

[0097]

[0098] In the formula,

[0099] n is the sample size in the sampling scheme, that is, the number of discrete sampling points,

[0100] N is the number of independent inspection units divided in the construction warehouse.

[0101] T β is the sampling failure rate q in the construction warehouse that is acceptable to the handover party β The estimated number of unqualified units to be inspected,

[0102] (n, 0) means that the number of unqualified measurement points allowed to appear in the n discrete sampling points randomly selected in the sampling plan is 0.

[0103] β is the probability of a pseudo-error in sampling detection,

[0104] β0 is the control probability of pseudo-error in sampling detection,

[0105] q β is the sampling failure rate acceptable to the transferor,

[0106] is the number of combinations of n discrete sampling point units randomly selected from the N units to be inspected in the construction warehouse surface,

[0107] To NT from the construction warehouse βThe number of combinations of n discrete sampling point units randomly selected from the qualified units to be inspected,

[0108] floor(·) is the floor rounding function;

[0109] If the construction party fails to conduct sufficient self-inspection on the compaction quality of the constructed warehouse surface, resulting in the construction party's lack of confidence in the construction quality, and at the same time, increasing the sampling inspection sample size n will not have a significant impact on the continuous construction process, and the construction party is willing to bear the additional inspection costs incurred by expanding the sampling inspection sample size n, the second statistical model that takes into account the acceptance risk control of both the construction party and the handover party is adopted. The second statistical model is specifically shown as follows:

[0110]

[0111] In the formula,

[0112] n is the sample size in the sampling plan, that is, the number of discrete sampling points,

[0113] m is the maximum number of unqualified test points allowed in the sampling plan.

[0114] N is the number of independent inspection units divided in the construction warehouse.

[0115] T β is the sampling failure rate q in the construction warehouse that is acceptable to the handover party β The estimated number of unqualified units to be inspected,

[0116] T α is the sampling failure rate q used by the construction party for self-inspection control in the construction warehouse α The estimated number of unqualified units to be inspected,

[0117] (n, m) means that the maximum number of unqualified points allowed in the n discrete sampling points randomly selected in the sampling plan is m.

[0118] β is the probability of a pseudo-error in sampling detection,

[0119] β0 is the control probability of pseudo-error in sampling detection,

[0120] α is the probability of false negatives in sampling detection,

[0121] α0 is the control probability of false positive errors in sampling detection,

[0122] γ is the risk self-control reduction coefficient of the construction party,

[0123] q β is the sampling failure rate acceptable to the transferor,

[0124] q αThe sampling failure rate used by the construction party for self-inspection control.

[0125] is the number of combinations of n discrete sampling point units randomly selected from the N units to be inspected in the construction warehouse surface,

[0126] T from the construction warehouse β The number of combinations of t discrete sampling point units randomly selected from the unqualified units to be inspected,

[0127] T from the construction warehouse α The number of combinations of t discrete sampling point units randomly selected from the unqualified units to be inspected,

[0128] To NT from the construction warehouse β The number of combinations of nt discrete sampling point units randomly selected from the qualified units to be inspected,

[0129] To NT from the construction warehouse α The number of combinations of nt discrete sampling point units randomly selected from the qualified units to be inspected,

[0130] floor(·) is the floor function.

[0131] S2) Based on the selected first statistical model or second statistical model, the Monte Carlo method is used to perform simulation and solution to obtain the specific values ​​of the sample capacity n in the sampling plan and the maximum number m of unqualified measurement points allowed in the sampling plan, and the optimal sampling plans (n, 0) or (n, m) based on different statistical models are obtained respectively.

[0132] Specifically, the specific steps of using the Monte Carlo method to solve the first statistical model are as follows:

[0133] S2A1) Determine the sampling failure rate q of the compaction quality on the surface of the construction bin to be inspected that is acceptable to the handover party β , the entire construction warehouse is discretized into N independent inspection units, random numbers are generated by software, and T is randomly assigned in the N independent inspection units. β Units are defective and unqualified units;

[0134] S2A2) specifying a set of retrieval number pairs (n, 0) as an initial sampling plan, wherein the sample size n in the sampling plan needs to be sufficiently small, such as n=4;

[0135] S2A3) randomly select n discrete sampling points from N independent units to be inspected in the construction warehouse surface. When the number of units to be inspected with defects in the n discrete sampling points is equal to 0, it is considered that the group of sampling results can pass the inspection;

[0136] S2A4) repeating step S2A3) for no less than 10,000 times, counting the frequency of occurrence of the sampling results that pass the test, and using the frequency as the probability β of a pseudo-error in the sampling test;

[0137] S2A5) setting the control probability β0 of the occurrence of nano-pseudo errors in sampling detection, when the probability of the occurrence of nano-pseudo errors in sampling detection obtained by simulation β>β0, the sample capacity n is increased by 1 sampling unit;

[0138] S2A6) Repeat steps S2A3) to S2A5) until the obtained sampling scheme satisfies the condition β≤β0. The sampling scheme at this time is the optimal sampling scheme (n, 0) under the principle of giving priority to protecting the transferred party.

[0139] Specifically, the specific steps of using the Monte Carlo method to solve the second statistical model are as follows:

[0140] S2B1) Determine the sampling failure rate q used by the construction party for self-inspection control α , the entire construction warehouse is discretized into N independent inspection units, random numbers are generated by software, and T is randomly assigned in the N independent inspection units. α Units are defective and unqualified units;

[0141] S2B2) using the sampling scheme obtained by solving the first statistical model using the Monte Carlo method as a preliminary sampling scheme (n, m) for solving the second statistical model;

[0142] S2B3) randomly select n discrete sampling points from N independent units to be inspected in the construction warehouse surface. When the number of units to be inspected with defects among the n discrete sampling points selected is less than or equal to m, it is considered that the group of sampling results can pass the inspection;

[0143] S2B4) repeating step S2B3) for no less than 10,000 times, counting the frequency of occurrence of sampling results that fail the test, and using the frequency as the probability α of a true rejection error in the sampling test;

[0144] S2B5) setting the control probability of sampling detection to reject the true error α0, when the probability of sampling detection to reject the true error obtained by simulation α>α0, the maximum number m of unqualified measurement points allowed in the sampling plan is increased by 1 sampling unit;

[0145] S2B6) using the sampling plan adjusted in step S2B5) as the initial sampling plan (n, m);

[0146] S2B7) repeating step S2B3) for no less than 10,000 times, counting the frequency of occurrence of the sampling results that pass the test, and using the frequency as the probability β of a pseudo-error in the sampling test;

[0147] S2B8) is compared with the set control probability β0 of the occurrence of nano-pseudo-error in the sampling detection. When the simulated probability of the occurrence of nano-pseudo-error in the sampling detection β>β0, the sample capacity n is increased by 1 sampling unit;

[0148] S2B9) repeating steps S2B7) to S2B8) until the obtained sampling plan satisfies the condition β≤β0, and using the obtained sampling plan as the adjusted primary sampling plan (n, m) for solving the second statistical model;

[0149] S2B 10 ) Repeat steps S2B3)~S2B9) until the obtained sampling plan satisfies the conditions of β≤β0 and α≤α0 at the same time. The sampling plan at this time is the optimal sampling plan (n, m) that takes into account the acceptance risk control of both the construction party and the handover party.

[0150] Preferably, in S2A5), the control probability β0 of a false positive error in sampling detection can generally be set to 10%; in S2B5), the control probability α0 of a false positive error in sampling detection can generally be set to 5%.

[0151] Preferably, step S2) can be implemented by programming with Matlab program.

[0152] In this embodiment, the construction pit surface that needs to be tested for compaction quality is located in the weathered material area on the downstream side of a certain earth-rock dam project in southwest my country, and its size is 80m×20m. According to the current specifications, the water filling method is used for compaction quality testing, and the test pit size is a cylindrical pit with a diameter and depth of 1m. Referring to the test pit size, the entire construction pit surface can be divided into 1m 2 The unit area is divided into 1600 independent square grid units. The risk points for self-inspection control by the construction party are (2.5%, 95%), and the risk points for acceptance control by the handover party are (10%, 10%).

[0153] In this embodiment, the construction party has fully self-inspected the constructed warehouse surface and has sufficient confidence in the construction quality. Considering the sampling inspection cost and its interference with the continuous construction process, after friendly consultation between the construction party and the transferee, the construction party agreed to use the first statistical model under the principle of giving priority to protecting the transferee for compaction quality acceptance.

[0154] In this embodiment, the first statistical model is used to perform compaction quality acceptance, and referring to step S2) for simulation and solution, the sample size n in the sampling scheme is obtained to be 22, that is, the best sampling scheme based on the first statistical model is (22, 0).

[0155] If the acceptance risk of the construction party needs to be taken into account at the same time, and the second statistical model in step S2) is used for simulation and solution, the best sampling scheme that meets the requirements is (77, 4), and its sampling frequency is 77 / 1600, which is much higher than the current standard recommended value of 1 / 5000 for sampling of dam shell materials. Therefore, the statistical model selected by the construction party and the transferee under the principle of giving priority to protecting the transferee is relatively reasonable.

[0156] S3) Based on the determined optimal sampling scheme (n, 0) or (n, m), n discrete sampling points are randomly selected from the N independent units to be inspected in the construction warehouse surface, and the spatial coordinates (x i ,y i ) to generate a spatial distribution pattern map of discrete sampling points that meets the requirements of spatial aggregation state control, which is used to guide sampling inspection at earth-rock dam construction sites.

[0157] Specifically, the specific steps for obtaining the spatial distribution pattern map of discrete sampling points that meets the spatial aggregation state control requirements are as follows:

[0158] S31) According to the shape of the actual construction warehouse surface, the N independent units to be inspected are numbered one by one according to a certain sorting principle;

[0159] S32) Based on the optimal sampling scheme (n, 0) or (n, m) determined in step S2), n discrete sampling point units are randomly selected from the N independent units to be inspected in the construction warehouse surface, and the spatial coordinates (x ) of the center positions of the selected n discrete sampling points are determined according to the numbers of the discrete sampling point units. i ,y i );

[0160] S33) Using the determined spatial coordinates (x) of the center position of the discrete sampling point i ,y i ), calculate the nearest point index R of the optimal sampling scheme (n, 0) or (n, m) in the spatial distribution within the construction warehouse;

[0161] S34) judging whether the optimal sampling scheme (n, 0) or (n, m) meets the control requirements of the spatial aggregation state according to the size of the nearest point index R; the calculation model of the nearest point index R is as follows:

[0162]

[0163] In the formula,

[0164] r e is the average distance between n discrete sampling points,

[0165] r T is the theoretical standard distance considering the regional boundary effect,

[0166] r i is the distance between the ith discrete sampling point and the nearest discrete sampling point,

[0167] n is the sample size in the optimal sampling scheme,

[0168] N is the number of independent inspection units discretely separated from the construction warehouse surface.

[0169] D is the theoretical density of n discrete sampling points evenly distributed within the construction warehouse.

[0170] B is the perimeter of the construction warehouse surface,

[0171] A is the area of ​​the construction warehouse;

[0172] S35) If the nearest point index R≤0.5, the n discrete sampling points are concentratedly distributed and do not meet the control requirements of the spatial aggregation state. Steps S32) to S34) need to be repeated until R>0.5 is satisfied and the n discrete sampling points are randomly or uniformly distributed. The sampling scheme at this time is the optimal sampling scheme after spatial layout optimization considering the distribution pattern of discrete sampling points.

[0173] In this embodiment, according to the determined optimal sampling scheme (22, 0), 22 discrete sampling points are randomly generated in the construction warehouse surface, such as Figure 2 As shown. By calculating the nearest point index R of the spatial distribution pattern of the 22 discrete sampling points generated by the characterization, it can be seen that the distribution of the 22 discrete sampling points meets the spatial aggregation state control requirements (R = 1.12> 0.5), which effectively limits the influence of the spatial autocorrelation of the discrete sampling points on the characterization accuracy of the compaction quality. Therefore, the compaction quality of the earth-rock dam construction bin surface can be sampled and tested on site according to the spatial distribution pattern diagram of the 22 discrete sampling points: if all the points in the 22 discrete sampling points meet the qualified standards, the compaction quality of the construction bin surface passes the inspection and can enter the next construction process; otherwise, it cannot pass the inspection and needs to be re-rolled within the range of the unqualified measuring points.

[0174] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A method for determining the optimal sampling scheme for compaction quality of earth-rock dam based on counting sampling principle, characterized in that: The steps include: S1) According to the risk tolerance of the construction party and the transferee, the statistical model on which the best sampling scheme depends is determined. If the construction party has fully self-checked the compaction quality of the constructed warehouse surface and has sufficient confidence in the construction quality, the first statistical model under the principle of giving priority to protecting the transferee is adopted to avoid the long-term operation safety risk of the earth-rock dam caused by the occurrence of pseudo-errors. The first statistical model is specifically shown as follows: In the formula, n is the sample size in the sampling scheme, that is, the number of discrete sampling points, N is the number of independent inspection units divided in the construction warehouse. T β is the sampling failure rate q in the construction warehouse that is acceptable to the handover party β The estimated number of unqualified units to be inspected, (n, 0) means that the number of unqualified measurement points allowed to appear in the n discrete sampling points randomly selected in the sampling plan is 0. β is the probability of a pseudo-error in sampling detection, β0 is the control probability of pseudo-error in sampling detection, q β is the sampling failure rate acceptable to the transferor, is the number of combinations of n discrete sampling point units randomly selected from the N units to be inspected in the construction warehouse surface, To NT from the construction warehouse β The number of combinations of n discrete sampling point units randomly selected from the qualified units to be inspected, floor(·) is the floor rounding function; If the construction party fails to conduct sufficient self-inspection on the compaction quality of the constructed warehouse surface, resulting in the construction party's lack of confidence in the construction quality, and at the same time, increasing the sampling inspection sample size n will not have a significant impact on the continuous construction process, and the construction party is willing to bear the additional inspection costs incurred by expanding the sampling inspection sample size n, the second statistical model that takes into account the acceptance risk control of both the construction party and the handover party is adopted. The second statistical model is specifically shown as follows: In the formula, n is the sample size in the sampling plan, that is, the number of discrete sampling points, m is the maximum number of unqualified test points allowed in the sampling plan. N is the number of independent inspection units divided in the construction warehouse. T β is the sampling failure rate q in the construction warehouse that is acceptable to the handover party β The estimated number of unqualified units to be inspected, T α is the sampling failure rate q used by the construction party for self-inspection control in the construction warehouse α The estimated number of unqualified units to be inspected, (n, m) means that the maximum number of unqualified points allowed in the n discrete sampling points randomly selected in the sampling plan is m. β is the probability of a pseudo-error in sampling detection, β0 is the control probability of pseudo-error in sampling detection, α is the probability of false negatives in sampling detection, α0 is the control probability of false positive errors in sampling detection, γ is the risk self-control reduction coefficient of the construction party, q β is the sampling failure rate acceptable to the transferor, q α The sampling failure rate used by the construction party for self-inspection control. is the number of combinations of n discrete sampling point units randomly selected from the N units to be inspected in the construction warehouse surface, T from the construction warehouse β The number of combinations of t discrete sampling point units randomly selected from the unqualified units to be inspected, T from the construction warehouse α The number of combinations of t discrete sampling point units randomly selected from the unqualified units to be inspected, To NT from the construction warehouse β The number of combinations of nt discrete sampling point units randomly selected from the qualified units to be inspected, To NT from the construction warehouse α The number of combinations of nt discrete sampling point units randomly selected from the qualified units to be inspected, floor(·) is the floor rounding function; S2) Based on the selected first statistical model or second statistical model, a Monte Carlo method is used to perform simulation and solve, to obtain specific values ​​of the sample capacity n in the sampling plan and the maximum number m of unqualified measurement points allowed in the sampling plan, and to obtain the optimal sampling plan (n, 0) or (n, m) based on different statistical models respectively; S3) Based on the determined optimal sampling scheme (n, 0) or (n, m), n discrete sampling points are randomly selected from the N independent units to be inspected in the construction warehouse surface, and the spatial coordinates (x i ,y i ) to generate a spatial distribution pattern map of discrete sampling points that meets the requirements of spatial aggregation state control, which is used to guide sampling inspection at earth-rock dam construction sites.

2. The method for determining the optimal sampling scheme for compaction quality of earth-rock dam based on counting sampling principle according to claim 1 is characterized by: In step S2), the specific steps of using the Monte Carlo method to solve the first statistical model are as follows: S2A1) Determine the sampling failure rate q of the compaction quality on the surface of the construction bin to be inspected that is acceptable to the handover party β , the entire construction warehouse is discretized into N independent inspection units, random numbers are generated by software, and T is randomly assigned in the N independent inspection units. β Units are defective and unqualified units; S2A2) specifying a set of retrieval number pairs (n, 0) as an initial sampling scheme; S2A3) randomly select n discrete sampling points from N independent units to be inspected in the construction warehouse surface. When the number of units to be inspected with defects in the n discrete sampling points is equal to 0, it is considered that the group of sampling results can pass the inspection; S2A4) repeating step S2A3) for no less than 10,000 times, counting the frequency of occurrence of the sampling results that pass the test, and using the frequency as the probability β of a pseudo-error in the sampling test; S2A5) setting the control probability β0 of the occurrence of nano-pseudo errors in sampling detection, when the probability of the occurrence of nano-pseudo errors in sampling detection obtained by simulation β>β0, the sample capacity n is increased by 1 sampling unit; S2A6) Repeat steps S2A3) to S2A5) until the obtained sampling scheme satisfies the condition β≤β0. The sampling scheme at this time is the optimal sampling scheme (n, 0) under the principle of giving priority to protecting the transferred party.

3. The method for determining the optimal sampling scheme for compaction quality of earth-rock dam based on counting sampling principle according to claim 2 is characterized by: In step S2), the specific steps of using the Monte Carlo method to solve the second statistical model are as follows: S2B1) Determine the sampling failure rate q used by the construction party for self-inspection control α , the entire construction warehouse is discretized into N independent inspection units, random numbers are generated by software, and T is randomly assigned in the N independent inspection units. α Units are defective and unqualified units; S2B2) using the sampling scheme obtained by solving the first statistical model using the Monte Carlo method as a preliminary sampling scheme (n, m) for solving the second statistical model; S2B3) randomly select n discrete sampling points from N independent units to be inspected in the construction warehouse surface. When the number of units to be inspected with defects among the n discrete sampling points selected is less than or equal to m, it is considered that the group of sampling results can pass the inspection; S2B4) repeating step S2B3) for no less than 10,000 times, counting the frequency of occurrence of sampling results that fail the test, and using the frequency as the probability α of a true rejection error in the sampling test; S2B5) setting the control probability of sampling detection to reject the true error α0, when the probability of sampling detection to reject the true error obtained by simulation α>α0, the maximum number m of unqualified measurement points allowed in the sampling plan is increased by 1 sampling unit; S2B6) using the sampling plan adjusted in step S2B5) as the initial sampling plan (n, m); S2B7) repeating step S2B3) for no less than 10,000 times, counting the frequency of occurrence of the sampling results that pass the test, and using the frequency as the probability β of a pseudo-error in the sampling test; S2B8) is compared with the set control probability β0 of the occurrence of nano-pseudo-error in the sampling detection. When the simulated probability of the occurrence of nano-pseudo-error in the sampling detection β>β0, the sample capacity n is increased by 1 sampling unit; S2B9) repeating steps S2B7) to S2B8) until the obtained sampling plan satisfies the condition β≤β0, and using the obtained sampling plan as the adjusted primary sampling plan (n, m) for solving the second statistical model; S2B 10 ) Repeat steps S2B3)~S2B9) until the obtained sampling plan satisfies the conditions of β≤β0 and α≤α0 at the same time. The sampling plan at this time is the optimal sampling plan (n, m) that takes into account the acceptance risk control of both the construction party and the handover party.

4. The method for determining the optimal sampling scheme for compaction quality of earth-rock dam based on counting sampling principle according to claim 1 is characterized by: In step S3), the specific steps of obtaining the spatial distribution pattern diagram of discrete sampling points that meets the spatial aggregation state control requirements are as follows: S31) According to the shape of the actual construction warehouse surface, the N independent units to be inspected are numbered one by one according to a certain sorting principle; S32) Based on the optimal sampling scheme (n, 0) or (n, m) determined in step S2), n discrete sampling point units are randomly selected from the N independent units to be inspected in the construction warehouse surface, and the spatial coordinates (x ) of the center positions of the selected n discrete sampling points are determined according to the numbers of the discrete sampling point units. i ,y i ); S33) Using the determined spatial coordinates (x) of the center position of the discrete sampling point i ,y i ), calculate the nearest point index R of the optimal sampling scheme (n, 0) or (n, m) in the spatial distribution within the construction warehouse; S34) judging whether the optimal sampling scheme (n, 0) or (n, m) meets the control requirements of the spatial aggregation state according to the size of the nearest point index R; the calculation model of the nearest point index R is as follows: In the formula, r e is the average distance between n discrete sampling points, r T is the theoretical standard distance considering the regional boundary effect, r i is the distance between the ith discrete sampling point and the nearest discrete sampling point, n is the sample size in the optimal sampling scheme, N is the number of independent inspection units discretely separated from the construction warehouse surface. D is the theoretical density of n discrete sampling points evenly distributed within the construction warehouse. B is the perimeter of the construction warehouse surface, A is the area of ​​the construction warehouse; S35) If the nearest point index R≤0.5, the n discrete sampling points are concentratedly distributed and do not meet the control requirements of the spatial aggregation state. Steps S32) to S34) need to be repeated until R>0.5 is satisfied and the n discrete sampling points are randomly or uniformly distributed. The sampling scheme at this time is the optimal sampling scheme after spatial layout optimization considering the distribution pattern of discrete sampling points.

Images