Soil cement mix design method and soil cement construction method

Abstract

The present invention provides a soil cement mix design method that uses artificial intelligence to set mix parameters that affect the physical properties of soil cement, thereby enabling the strength of the elastic body region to be expressed and preventing the soil cement from turning into mud, and a soil cement construction method that includes said method. [Solution] The present invention is a method for designing soil cement mixes that includes at least the steps of (A) implementing trained artificial intelligence, (B) outputting lead mix examples, (C) measuring the physical properties of test specimens, and (D) setting a nominal mix. The present invention also provides a soil cement method that includes at least (i) a soil and sand collection step, (ii) a material testing step using the soil and sand, (iii) a step of setting a designated mix ratio using the above-mentioned mix design method, and (iv) a test construction step using the designated mix ratio, and further includes, as an optional step, (v) a step of resetting the designated mix ratio using the above-mentioned mix design method if the target strength is not met in the test construction.

Description

[Technical Field]

[0001] The present invention relates to a soil cement mix design method that uses artificial intelligence (AI) to set a soil cement nominal mix or, if necessary, reset (correct) it, and a soil cement construction method that uses said method. [Background technology]

[0002] The recommended mix is ​​typically determined through laboratory mix testing, and the mix parameters include the type of cement, the unit cement content, and the test water content. As shown in Figure 1, the test water content includes the water content on the left side of the quadratic curve intersecting the line representing the target compressive strength (lower limit water content), the test water content representing the peak strength of the curve (peak strength) (peak strength water content), and the water content on the right side of the curve (upper limit water content). As shown in Figure 2, the curve is determined by varying the unit cement content and test water content based on the type and characteristics of the soil and sand generated at the construction site (hereinafter simply referred to as "soil") obtained through tests (hereinafter referred to as "material tests") on the physical and chemical properties of the soil and sand (hereinafter referred to as "soil").

[0003] The above-mentioned soil cement is a mixture of earth and sand mixed with cement, water, etc., or a solidified product thereof. The main construction methods using soil cement are the sabo soil cement method (which includes compaction and fluidization types) used to build sabo facilities, and the ground improvement method, which involves excavating soft ground and replacing it with soil cement. The two methods differ in the strength and workability of the soil cement they aim for (hereinafter sometimes referred to as "quality" or "physical properties"), as well as the construction method. Of these, the sabo soil cement method is a method in which sabo facilities, which were previously constructed with concrete, are constructed by solidifying them with soil cement. Both methods utilize the soil generated by excavation, etc., so there is no need to transport or dispose of the soil, and it is possible to reduce the costs, noise, vibrations, etc. that arise from transporting soil. In addition, sabo soil cement generally has a viscosity of 3.0 N / mm 2It has the strength of the elastic region (the region where strain and stress are proportional) and is excellent for workability such as laying and compacting.

[0004] The mix design for the soil cement method is carried out through (i) soil sampling, (ii) material testing, and (iii) setting of a nominal mix through indoor mix testing (see, for example, page 52 of Non-Patent Document 1).

[0005] However, the soil used in the above-mentioned indoor mix tests (e.g., soil from the surface of a construction site sampled by hand) and the soil collected during construction (e.g., soil from a mixture of surface and deep layers excavated extensively with construction machinery) may have different soil properties due to the different collection areas. These differences in soil properties affect the quality of the soil cement. To confirm this effect, a trial construction (hereinafter referred to as "trial construction") is carried out at the construction site based on the designated mix obtained from the mix design. If the trial construction results in the soil cement with the designated mix not meeting the required quality, such as strength, material tests and indoor mix tests are repeated using the soil to be used, and the designated mix is ​​then reset (corrected).

[0006] However, since the strength of soil cement is usually measured using the compressive strength of a 28-day-old specimen, construction is suspended until this strength is achieved. As mentioned above, the recommended mix proportions include a wide range of items, such as the type of cement, the unit cement content, and the test water content. As described above, in conventional soil cement construction methods, material testing and laboratory mix testing are repeatedly conducted for a wide range of mix parameters, making soil cement mix design complicated, time-consuming, and expensive.

[0007] Therefore, in order to improve the efficiency of soil cement mix design, Patent Document 1 proposes a method for determining whether soil and sand can be used, and Patent Document 2 proposes a method for soil cement mix design. Of these, the method of Patent Document 1 includes a step of determining the amount of microorganisms, the amount of humic acid, and the water content contained in the soil and sand, and comparing them with criteria (threshold values) for determining whether the soil or sand can be used, and a step of determining that the soil or sand can be used if each of the amounts is below the threshold value.

[0008] The method of Patent Document 2 includes steps of collecting soil and sand from a construction site; A step of determining a target strength of a construction object using soil cement; Calculating the fine particle content of the on-site soil and sand used in the soil cement mix; a step of determining whether or not a target strength determined from the fine particle content calculated from the relationship between the fine particle content, the amount of cement added, and the compressive strength is met; A step of determining a type of soil cement that satisfies the target strength when it is determined that the target strength is met; A step of determining the amount of cement to be added based on the relationship between the fine particle content, the type, and the amount of cement to be added; determining the amount of water to be added based on the relationship between the fine particle content and the amounts of cement and water to be added; Based on the determined amounts of cement and water added, a laboratory mix test is carried out. The method includes:

[0009] However, the method of Patent Document 1 is limited to determining whether or not soil and sand can be used, and the method of Patent Document 2 is complicated and involves many steps. [Prior art documents] [Non-patent literature]

[0010] [Non-Patent Document 1] Sabo and Landslide Technical Center, "Sabo Soil Cement Construction Handbook," revised second edition published on September 30, 2016 [Patent documents]

[0011] [Patent Document 1] JP 2019-74485 A [Patent Document 2] Japanese Patent Application Laid-Open No. 2016-218031 Summary of the Invention [Problem to be solved by the invention]

[0012] As mentioned above, sabo soil cement needs to have long-term durability and strength in the elastic range, and unlike the soil cement used in the above-mentioned ground improvement methods, it is essential that the soil and sand be solidified by hydration of the cement. Furthermore, because cement hydration is affected by various factors, the mix design for sabo soil cement is particularly complicated, time-consuming, and costly. Therefore, the present invention aims to provide a soil cement mix design method that uses artificial intelligence to set the above-mentioned mix parameters that affect the physical properties of soil cement, such as the hydration reaction of cement, thereby enabling the strength of the elastic body region to be expressed and preventing the soil cement from turning into mud, and a soil cement construction method that includes said method. [Means for solving the problem]

[0013] As a result of extensive research to achieve the above object, the present inventors have completed the following inventions [1] to [7].

[0014] [1] A method for designing a soil-cement mix, comprising at least the following steps (A) to (D): (A) An implementation process of the trained artificial intelligence by using training data including at least the soil quality, cement type, unit cement amount, test water content, unit volume mass of the soil cement, compressive strength, and mixing ratio of the improvement material to the soil and sand, and then implementing the trained artificial intelligence. (B) A lead mix example output process in which the trained artificial intelligence inputs at least the soil quality of the soil and sand at the construction site, the type of cement used in the soil cement, the unit cement amount, and the test water content, and the trained artificial intelligence outputs a lead mix example (a mix example that serves as a starting point for deriving (leading) a formula mix) that includes at least the unit cement amount and the test water content. (C) A step of measuring the physical properties of a soil cement specimen based on the output lead mix example, and measuring at least the short-term strength and mass of the specimen. (D) A process of setting a recommended mix by inputting the measured physical property values ​​of the test specimen into the trained artificial intelligence, which then predicts at least the unit cement amount, test water content, and long-term strength of the soil cement and outputs a mix example, and selecting a mix example from the mix examples that satisfies the target long-term strength as the recommended mix. (However, the compressive strength described in step (A) above includes short-term strength and long-term strength.) [2] (E) If the long-term strength of the specimen prepared based on the designated mix ratio set in the above (D) step does not satisfy the target long-term strength, the above (B) to (D) steps are repeated to reset the designated mix ratio. ,difference The method for designing a soil cement mix according to [1] above, further comprising: [3] The method for designing a soil cement mix according to [1] above, wherein the step (D) includes at least the following steps (D1) to (D5): (D1) A quadratic regression analysis process in which, for each unit cement amount, a regression analysis is performed using the test water content as an explanatory variable and the predicted value of the long-term strength as a response variable, and a quadratic regression equation is obtained for each unit cement amount. (D2) A process for predicting peak intensity, etc., using the above quadratic regression equation to calculate and predict peak intensity, peak intensity water content, upper limit water content and lower limit water content, and upper and lower limit water content intensity based on the upper limit water content and lower limit water content. (D3) A first linear regression analysis process in which a regression analysis is performed using the unit cement amount as an explanatory variable and the peak intensity as a response variable to obtain a first linear regression equation. (D4) A process of calculating the unit cement amount by performing a regression analysis using the unit cement amount as an explanatory variable and the upper and lower limit water content strengths as objective variables to obtain a second linear regression equation, and then substituting the target strength value into the second linear regression equation to calculate the unit cement amount value to be described in the specified mix proportion. (D5) A process for calculating the design water content, in which a regression analysis is performed using the unit cement amount as an explanatory variable and the peak strength water content as a target variable to obtain a third linear regression equation, and then the value of the unit cement amount described in the specified mix is ​​substituted into the third linear regression equation to calculate the value of the peak strength water content, and the value of the peak strength water content is set as the design water content described in the specified mix. [4] The soil cement mix design method according to [3] above, wherein the quadratic regression equation and the linear regression equation are the following equation (1): y=ax 2 +bx+c (1) In equation (1), x is an explanatory variable, y is a response variable, and a, b, and c are regression coefficients. In particular, a<0 ​​when x is the test water content and y is the long-term strength, a=0 when x is the unit cement content and y is the peak strength or upper / lower limit water content strength, and a=0 when x is the unit cement content and y is the peak strength water content. [5] The method for designing a soil cement mix according to [1] above, wherein the soil quality is one or more selected from the particle size obtained in a soil particle size test, the density obtained in a soil particle density test, the bone dry specific gravity and water absorption rate obtained in a coarse aggregate density and water absorption rate test, the bone dry specific gravity and water absorption rate obtained in a fine aggregate density and water absorption rate test, the maximum dry density and optimum water content obtained in a soil compaction test by tamping, the water content obtained in a soil natural water content test, and the color of the soil and sand obtained in an organic impurity test. [6] A method for designing a soil cement mix according to [1] above, wherein the short-term strength is the compressive strength at or before the age of 7 days, and the long-term strength is the compressive strength at or after the age of 28 days. [7] (i) a process of collecting soil and sand, (ii) a material testing process using the soil and sand, (iii) a process of setting a designated mix ratio using the mix design method described in any one of [1] to [6] above, and (iv) a test construction process using the designated mix ratio. 、 ( v) If the target strength is not met in the above test construction, the process of resetting the designated mix using the above mix design method will be carried out. ,moreover Including soil cement construction method. [Effects of the Invention]

[0015] The mix design method of the present invention uses artificial intelligence to easily and accurately set or reset the designated mix proportions. Furthermore, the soil cement construction method of the present invention includes a step of setting or resetting the designated mix proportions using the mix design method, so the soil cement construction method can be implemented early. [Brief explanation of the drawings]

[0016] [Figure 1] 1 is a conceptual diagram showing the relationship between the test water content of soil and the compressive strength of soil cement. [Figure 2] This graph plots the measured test water content of soils of different soil types and the measured compressive strength of soil-cement at 28 days of age, broken down by unit cement content (C-80 to C-200). The soil types, from left to right on the quadratic curves, are those with high soil-cement suitability (e.g., crushed run), medium suitability (e.g., sandy soil), and low suitability (e.g., clayey soil). The higher the suitability of the soil and the higher the unit cement content, the larger the regression coefficient (absolute value) of the quadratic term in the quadratic function, and the more pointed the quadratic curve becomes. The horizontal line indicates the target strength (mix strength, 4.5 N / mm2) in the laboratory mix test. [Figure 3] This graph plots the measured test water content and the measured compressive strength of soil-cement at 28 days old for three different unit cement amounts. The unit cement amounts for soil-cement are 50kg / m3, 80kg / m3, and 110kg / m3, respectively, from the bottom of the quadratic curve. The horizontal line indicates the target strength at the site (mix strength, 3.0N / mm2). [Figure 4] This is a graph plotting the unit cement content, peak strength (◯), and upper and lower limit water content strength (●) from Figure 3. The unit cement content indicated by the intersection of the dashed line and the target strength at the site (horizontal line, 3.0 N / mm2) is the optimal unit cement content (131.9 kg / m3) for the specified mix proportion. [Figure 5]This is a graph plotting the unit cement content and peak strength water content of Figure 3. The peak strength water content indicated by the above optimum cement content (131.9 kg / m3) is the optimum test water content (15.9%) for the designated mix proportion. DETAILED DESCRIPTION OF THE INVENTION

[0017] The soil cement mix design method of the present invention includes at least (A) a process of implementing trained artificial intelligence, (B) a process of outputting a lead mix example, (C) a process of measuring the physical properties of a test specimen, and (D) a process of setting a nominal mix. In addition, the soil cement method of the present invention includes at least (i) a soil and sand collection process, (ii) a material testing process using the soil and sand, (iii) a process of setting a designated mix ratio using the above-mentioned mix design method, and (v) a test construction process using the designated mix ratio. The soil cement mix design method and soil cement construction method of the present invention will be described in detail below.

[0018] 1. Soil-cement mix design method (A) Implementation process of trained AI The above-mentioned process (A) is a process of implementing the trained artificial intelligence by machine learning using training data that includes at least the soil quality, cement type, unit cement amount, test water content, unit volume mass of the soil cement, compressive strength, and mixing ratio of improvement material to the soil and sand used in the soil cement.

[0019] As shown in Table 1, the above soil quality is one or more selected from the particle size obtained in a soil particle size test, the density obtained in a soil particle density test, the bone dry specific gravity and water absorption rate obtained in a coarse aggregate density and water absorption rate test, the bone dry specific gravity and water absorption rate obtained in a fine aggregate density and water absorption rate test, the maximum dry density and optimum water content obtained in a soil compaction test by tamping, the water content obtained in a soil natural water content test, and the color of the soil and sand obtained in an organic impurity test.

[0020] The type of cement is not particularly limited and may be one or more selected from Portland cements such as blast furnace cement type A, blast furnace cement type B, blast furnace cement type C, and ordinary Portland cement, silica cement, fly ash cement, ecocement, and cement-based solidification materials. The cement-based solidification material is a composite material with cement as the base material, and the other solidification components and their amounts are determined based on the degree of difficulty of solidification (e.g., for general soft soil, special soil (general-purpose type), or high organic soil), as well as the conditions of the construction site. Examples of commercially available cement-based solidification materials include Geoset (registered trademark, manufactured by Taiheiyo Cement Corporation) 200 for general-purpose use, and Geoset 225 (manufactured by Taiheiyo Cement Corporation) and U-Stabilizer (manufactured by UBE Mitsubishi Corporation) for high organic soil. Considering cost and solidification performance, the solidification material is preferably blast furnace cement type B or a cement-based solidification material.

[0021] (B) Output process of lead blend example The above-mentioned (B) step is a step in which the trained artificial intelligence inputs at least the soil quality of the soil and sand at the construction site, the type of cement used in the soil cement, the unit cement amount, and the test moisture content, and the trained artificial intelligence outputs a lead mix example that includes at least the unit cement amount and the test moisture content.

[0022] (C) Measurement process of the physical properties of the specimen The step (C) is a step of preparing a soil cement specimen based on the output lead mix example and measuring at least the short-term strength and mass of the specimen. The strength measurement should be performed in accordance with JIS A 1108-2006 "Testing Method for Compressive Strength of Concrete." Here, the short-term strength is the compressive strength of soil cement at an age of 7 days or less. In the field of soil cement, the compressive strength at an age of 7 days is generally used for the short-term strength. However, if the long-term strength can be predicted based on the strength at a shorter age, the short-term strength of the test specimen can be obtained in a shorter period of time, so the shorter the short-term age, the better, as long as high prediction accuracy can be maintained.

[0023] (D) Setting the formula composition The above-mentioned (D) step is a step in which the measured physical property values ​​of the test specimen are input into the trained artificial intelligence, and the trained artificial intelligence predicts at least the unit cement amount, test water content, and long-term strength of the soil cement, outputs a mixture example, and sets the mixture example that satisfies the target strength from among the mixture examples as the nominal mixture. Here, the long-term strength is the compressive strength of soil cement after 28 days. However, in the field of soil cement, the compressive strength at 28 days is generally used for long-term strength. If the long-term strength of the specimens prepared based on the designated mix proportions set in step (D) above does not satisfy the target strength, the method may further include an optional step of resetting the designated mix proportions by repeating steps (B) to (D) above. By performing this optional step, predicted strengths exceeding the target strength can be easily obtained simply by changing and inputting the unit cement content and test water content. Therefore, the present invention, which utilizes artificial intelligence, eliminates the need for the time-consuming and labor-intensive 28-day compressive strength test, thereby speeding up and simplifying mix design.

[0024] More specifically, the above step (D) is a step including at least the following steps (D1) to (D5): (D1) Quadratic regression analysis process The (D1) process is a process of performing a regression analysis for each unit cement content, using the test water content as the explanatory variable and the predicted value of the long-term strength as the objective variable, to find a quadratic regression equation for each unit cement content. Figures 2 and 3 show graphs of the (D1) process.

[0025] (D2) Prediction process of peak intensity, etc. The above process (D2) uses the above quadratic regression equation to calculate and predict peak strength, peak strength moisture content, upper and lower limit moisture content, and upper and lower limit moisture content strengths based on the upper and lower limit moisture content. Because the above quadratic regression equation is symmetrical around the peak strength moisture content, the compressive strength on the quadratic curve indicated by the upper limit moisture content and the compressive strength on the quadratic curve indicated by the lower limit moisture content are the same value, and the upper and lower limit moisture content strengths based on the upper and lower limit moisture content are the same value. The above upper and lower limit moisture content can be substituted by the peak strength moisture content ±2-3%. The reason for substituting the peak strength moisture content ±2-3% for the upper and lower limit moisture content is that it is difficult to manage construction using a single peak strength moisture content value, allowing for a certain range. A conceptual diagram of the (D2) process is shown in Figure 1. Incidentally, the peak intensity (maximum value) can be determined by differentiating the above quadratic regression equation. Here, the upper and lower limit moisture content are substituted with the peak strength moisture content ±2 to 3%, as it is difficult to manage construction using a single value for the peak strength moisture content, and a certain range is provided.

[0026] (D3) First linear regression analysis process The (D3) step is a step of performing a regression analysis using the unit cement amount as an explanatory variable and the peak intensity as a response variable to find a first linear regression equation.

[0027] (D4) Calculation process for unit cement amount The (D4) step is a step of performing a regression analysis using the unit cement amount as an explanatory variable and the upper and lower limit water content strengths as objective variables to obtain a second linear regression equation, and then substituting the target strength value into the second linear regression equation to calculate the unit cement amount value to be written in the designated mix proportion.

[0028] (D5) Calculation process for design moisture content The above (D5) process is a process of performing regression analysis using the unit cement amount as the explanatory variable and the peak strength water content as the target variable to obtain a third linear regression equation, then substituting the value of the unit cement amount listed in the designated mix proportion into the third linear regression equation to calculate the value of the peak strength water content, and setting the value of the peak strength water content as the design water content to be listed in the designated mix proportion.

[0029] The quadratic regression equation and the linear regression equation are the following equations (1). y=ax 2 +bx+c (1) In equation (1), x is an explanatory variable, y is a response variable, and a, b, and c are regression coefficients. In particular, a<0 ​​when x is the test water content and y is the long-term strength, a=0 when x is the unit cement content and y is the peak strength or upper / lower limit water content strength, and a=0 when x is the unit cement content and y is the peak strength water content.

[0030] 2. Soil cement method Next, the soil cement method of the present invention will be described. The soil cement method of the present invention includes at least (i) a soil and sand collection step, (ii) a material testing step using the soil and sand, (iii) a step of setting a designated mix ratio using the mix design method described in any one of [1] to [6] above, and (iv) a test construction step using the designated mix ratio, and further includes, as an optional step, (v) a step of resetting (correcting) the designated mix ratio using the mix design method if the target strength is not achieved in the test construction. The above steps (i) to (v) will be explained in detail below.

[0031] (i) Soil collection process The depth to which soil and sand can be excavated and the amount of soil and sand that can be extracted are investigated to understand the variation in the soil and sand. If the variation in the soil and sand is large, it is necessary to decide whether to separate the soil and sand and collect it, or whether to mix the soil and sand that has been collected separately.

[0032] (ii) Material testing process using the soil and sand The material test items are soil quality that affects the physical properties of soil cement, such as particle size, density, water absorption rate, compactibility, water content, and organic impurities. These test items should be tested in accordance with the test methods specified in Table 1 and other standards such as the Japanese Industrial Standards (JIS).

[0033] (iii) Setting the designated recipe The designated mix is ​​set using the mix design method of the present invention.

[0034] (iv) Test construction process Test specimens are prepared based on the specified mix ratio set above, and physical properties such as long-term strength are measured. As mentioned above, the soil used in the above-mentioned laboratory mix test and the soil to be used for construction collected during the construction stage may have different soil properties. Therefore, in order to confirm the effect that differences in soil properties have on the quality of soil cement, trial construction is carried out at the construction site based on the designated mix obtained from the laboratory mix test.

[0035] (v) Resetting the formula If the results of the test construction show that the soil cement specimen does not meet the required quality, the designated mix proportions are reset using the mix design method of the present invention. [Example]

[0036] The present invention will be described below with reference to examples, but the present invention is not limited to these examples.

[0037] (A) Implementation of trained artificial intelligence The soil quality used for soil cement (each item of the material test shown in Table 1 and Table 2), cement type, unit cement amount, test water content, unit volume mass (based on the mass of the test specimen, the volume of the test specimen is 1 m 3 The artificial intelligence was trained using training data including the mass of the soil when converted into a concrete mass, the compressive strength at 7 and 28 days, and the mixing ratio of the improvement material to the soil. The material test methods in Table 2 are the same as those listed in Table 1.

[0038] [Table 1]

[0039] [Table 2]

[0040] (B) Output of lead blending example Next, the soil quality values, cement type, unit cement amount, and test moisture content of the soil and sand at the construction site (alluvial fan) (hereinafter referred to as "sample soil") obtained by measuring using the test methods shown in Tables 1 and 2 were input into the trained AI. As a result, the trained AI output lead mix examples (2a, 2b, and 2c) consisting of the unit cement amount, test moisture content, and unit volume mass shown in Table 3. The lead mix examples are shown in Table 3.

[0041] [Table 3]

[0042] (C) Measurement of physical properties of the specimen Based on the lead blending examples (2a, 2b, and 2c) shown in Table 3, the above sample soil, which had been air-dried for 18 hours in a room at 20°C, a cement-based solidification material (product name: Blast-furnace cement type B, manufactured by Taiheiyo Cement Corporation), and water were mixed in a pan mixer to obtain the above three types of soil-cement mixtures. Next, the kneaded material was poured into a mold, sealed, and demolded after 7 days. Three test pieces with a diameter of 125 mm and a height of 250 mm were prepared for each blending example. The mass of each test piece was measured, and the volume of the test piece was then calculated based on the mass of the test piece. 3 The mass when converted to 1000g was calculated to calculate the unit volume mass. Furthermore, the compressive strength of the above specimens was measured at an age of 7 days in accordance with JIS A 1108-2006 "Testing Method for Compressive Strength of Concrete." The above unit volume mass and compressive strength are shown in Table 3. Note that the above unit volume mass and compressive strength are the average values ​​of three specimens for each mix example.

[0043] (D) Setting the formula composition The unit cement amount, test water content, unit volume mass, and compressive strength at 7 days for each of the mix examples (2a, 2b, and 2c) listed in Table 3 were input into the trained AI. As a result, the trained AI output the unit cement amount, test water content, unit volume mass, and compressive strength at 28 days for mix examples 1a to 1c and mix examples 3a to 3c in Table 4, and the test water content, unit volume mass, and compressive strength at 28 days for mix examples 2a to 2c. These output results are shown in Table 4.

[0044] [Table 4]

[0045] (D1) Quadratic regression analysis process For each of the unit cement amounts listed in Table 4, a regression analysis was performed using the test water content as the explanatory variable and the predicted value of compressive strength at 28 days as the objective variable, and a quadratic regression equation was obtained for each of the unit cement amounts. The results are shown in a graph in Figure 3. Incidentally, the unit cement amounts of white triangles (△), white circles (◯), and white squares (□) are 50 kg / m 3 , 80 kg / m 3 , and 110 kg / m 3 is.

[0046] (D2) Prediction process of peak intensity, etc. Next, the quadratic regression equation was used to predict the peak intensity, the peak intensity water content, the upper limit water content and the lower limit water content, and the upper and lower limit water content strengths based on the upper limit water content and the lower limit water content.

[0047] (D3) First linear regression analysis process A regression analysis was performed with the unit cement content as the explanatory variable and the peak strength as the objective variable to obtain the first linear regression equation. The upper and lower water content strength limits were the strengths at water contents within ±3% of the peak strength water content.

[0048] (D4) Calculation process for unit cement amount A regression analysis was performed using the unit cement amount as an explanatory variable and the upper and lower limit water content strength as an objective variable to obtain a second linear regression equation, and then the target strength value (3.0 N / mm 2 ) to obtain the unit cement amount (103.3 kg / m 3 ) was calculated.

[0049] (D5) Calculation process for design moisture content The unit cement amount was used as an explanatory variable, and the peak strength water content was used as a target variable. Regression analysis was performed to obtain a third linear regression equation, and the unit cement amount (103.3 kg / m) listed in the specified mix ratio was then applied to the third linear regression equation. 3 The peak strength water content (10.8%) was calculated by substituting the above formula, and the peak strength water content was used as the design water content to be recorded in the designated mix proportions. The designated mix proportions obtained in this way are shown in Table 5. The unit cement content was 5.0 kg / m 3 Rounded number (e.g., 103.3 kg / m 3 is 105 kg / m 3 ) and the design moisture content was rounded off by 0.1% (for example, 10.81% was 10.8%, and 10.85% was 10.9%). In addition, soil cement specimens were prepared according to the mix proportions shown in Table 5, and the compressive strength at 28 days was measured in accordance with JIS A 1108-2006 "Testing Method for Compressive Strength of Concrete." As a result, the compressive strength at 28 days was 1.5 N / mm 2 exceeding the minimum strength of 2.7N / mm 2 , minimum strength is 3.81N / mm 2 It was.

[0050] [Table 5]

Claims

1. A method for designing a soil cement mix, comprising at least the following steps (A) to (D): (A) A process of implementing the trained artificial intelligence by using training data including at least the soil quality, cement type, unit cement amount, test water content, unit volume mass of the soil cement, compressive strength, and mixing ratio of improvement material to the soil and sand, to train the artificial intelligence to learn the trained artificial intelligence. (B) A lead mix example output process in which the trained artificial intelligence inputs at least the soil quality of the soil at the construction site, the type of cement used in the soil cement, the unit cement amount, and the test water content, and the trained artificial intelligence outputs a lead mix example (a mix example that serves as a starting point for deriving (leading) a formula mix) that includes at least the unit cement amount and the test water content. (C) A step of measuring the physical properties of a soil cement specimen based on the output lead mix example, and measuring at least the short-term strength and mass of the specimen. (D) The physical property values ​​of the measured specimen are input into the trained artificial intelligence, and the trained artificial intelligence predicts at least the unit cement amount, test water content, and long-term strength of the soil cement, and outputs a mix example. A mix example that satisfies the target long-term strength is set as the recommended mix from among the mix examples. (However, the compressive strength described in the above step (A) includes short-term strength and long-term strength.)

2. (E) A soil cement mix design method as described in claim 1, further comprising a step of resetting the designated mix by repeating steps (B) to (D) above if the long-term strength of the test specimen prepared based on the designated mix set in step (D) above does not satisfy the target long-term strength.

3. The soil cement mix design method according to claim 1, wherein the step (D) includes at least the following steps (D1) to (D5): (D1) a quadratic regression analysis process for performing a regression analysis for each unit cement amount using the test water content as an explanatory variable and the predicted value of the long-term strength as a response variable, and determining a quadratic regression equation for each unit cement amount. (D2) A process for predicting peak intensity, etc., using the above-mentioned quadratic regression equation to calculate and predict peak intensity, peak intensity water content, upper limit water content and lower limit water content, and upper and lower limit water content intensities based on the upper limit water content and lower limit water content. (D3) A first linear regression analysis step of performing a regression analysis using the unit cement amount as an explanatory variable and the peak intensity as a response variable to obtain a first linear regression equation. (D4) A process of calculating the unit cement amount, in which a regression analysis is performed using the unit cement amount as an explanatory variable and the upper and lower limit water content strengths as objective variables to obtain a second linear regression equation, and then the target strength value is substituted into the second linear regression equation to calculate the value of the unit cement amount to be described in the designated mix proportion. (D5) A process for calculating the design water content, in which a regression analysis is performed using the unit cement amount as an explanatory variable and the peak strength water content as a target variable to obtain a third linear regression equation, and then the value of the unit cement amount described in the designated mix proportion is substituted into the third linear regression equation to calculate the value of the peak strength water content, and the value of the peak strength water content is set as the design water content described in the designated mix proportion.

4. 4. The soil cement mix design method according to claim 3, wherein the quadratic regression equation and the linear regression equation are the following equation (1): y=ax ２ +bx+c ・・・(1) In equation (1), x is an explanatory variable, y is a response variable, and a, b, and c are regression coefficients. In particular, a<0 ​​when x is the test water content and y is the long-term strength, a=0 when x is the unit cement content and y is the peak strength or upper / lower limit water content strength, and a=0 when x is the unit cement content and y is the peak strength water content.

5. 2. The method for designing a soil cement mix according to claim 1, wherein the soil quality is one or more selected from the group consisting of particle size obtained in a soil particle size test, density obtained in a soil particle density test, bone dry specific gravity and water absorption rate obtained in a coarse aggregate density and water absorption rate test, bone dry specific gravity and water absorption rate obtained in a fine aggregate density and water absorption rate test, maximum dry density and optimum water content obtained in a soil compaction test by tamping, water content obtained in a soil natural water content test, and color tone of the soil or sand obtained in an organic impurity test.

6. 2. A method for designing a soil cement mix according to claim 1, wherein the short-term strength is the compressive strength at or before 7 days of age, and the long-term strength is the compressive strength at or after 28 days of age.

7. (i) a soil and sand collection step, (ii) a material testing step using the soil and sand, (iii) a step of setting a designated mix ratio using the mix design method according to any one of claims 1 to 6, and (iv) a test construction step using the designated mix ratio, (v) A soil cement method further including a step of resetting the designated mix using the mix design method if the target strength is not met in the test construction.

Images