Soil intelligent classification automation system and method

Abstract

The invention discloses an intelligent soil classification automation system and method. Comprising a soil sample pretreatment module used for air drying, stirring and weighing treatment; the soil sample particle size grading screening module is used for carrying out step-by-step screening treatment to obtain a first-stage particle distribution result; the liquid-plastic limit automatic measurement module is used for performing measurement by using a conometer method to obtain a measurement result of the conometer method; the burning measurement module is used for carrying out burning treatment to obtain a corresponding burning measurement result; the densimeter measurement module is used for carrying out particle analysis through a densimeter method to obtain a second-stage particle distribution result; and the control analysis module is used for realizing automatic control of the intelligent soil classification automation system and performing data analysis to obtain the soil sample type and the corresponding optimal solidification method. According to the system disclosed by the invention, the whole-process automation and data processing of soil classification testing are realized, and the soil sample type and the corresponding optimal solidification method can be automatically obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of soil classification, and in particular to an automated system and method for intelligent soil classification. Background Art

[0002] my country boasts a vast territory, diverse landforms, and a complex array of soil types. In the construction of major infrastructure projects such as buildings, highways, railways, and water conservancy projects, soil engineering classification standards not only provide a basis for design and ensure construction safety, but also underpin the appropriate disposal and modification of heterogeneous soils. Natural soils vary significantly in particle composition, mineral content, organic matter content, and moisture conditions, resulting in highly heterogeneous engineering properties. According to the "GB / T 50123-2019 Standard for Soil Testing Methods," the traditional sieve analysis method combined with the liquid limit and plastic limit determination requires manual, step-by-step operation, with a single test taking up to 6-8 hours. This complex and variable soil sample presents common challenges, including cumbersome operation, low efficiency, and a lack of intelligent judgment, making it difficult to meet the dual demands of efficient and accurate classification and curing recommendations on-site.

[0003] This demonstrates that traditional soil classification methods suffer from heavy reliance on manual labor, resulting in broken data chains, poor equipment coordination, and inefficient and inconsistent testing. This is particularly true when dealing with specialized soil samples, such as those with high liquid limits, high organic matter content, and high water content. The lack of quantitative identification and adaptive solidification mechanisms often results in wasteful mixes and ineffective treatments. In sectors with demanding foundation requirements, such as transportation, rail, and water conservancy, this disconnect between classification and modification directly impacts construction safety and cost control.

[0004] Specifically:

[0005] On the one hand, traditional geotechnical material testing methods, such as sieve analysis, densimeters, and the combined liquid-plastic limit method, rely heavily on manual intervention, resulting in complex and time-consuming testing processes and difficulties in achieving efficient and real-time analysis. Due to the limitations of manual operation, these methods often lack real-time performance, accuracy, and efficiency. Furthermore, due to the uncertainty introduced by manual intervention, the stability of soil sample types is often difficult to guarantee.

[0006] On the other hand, existing experimental methods are typically based on independent operations, lacking data sharing and collaborative working mechanisms between devices. This information silo phenomenon not only increases the complexity of data integration but can also lead to omissions or contradictions in key information, increasing the difficulty of data processing and the potential risk of errors. This makes it particularly difficult to accurately identify the engineering risks and curing adaptability of special soils such as those with high liquid limits, high organic matter content, and high water content. This significantly limits the application value of geotechnical material testing methods in complex engineering environments. Summary of the Invention

[0007] To address the shortcomings of existing technologies, the present invention provides an automated intelligent soil classification system and method. The goal is to efficiently, accurately, and intelligently classify complex and highly diverse natural soil samples and provide differentiated recommendations for their curing. The system combines machine learning technology with intelligent control algorithms to systematically collect and analyze multi-dimensional soil parameters, including particle size distribution, moisture content, density, liquid limit, plastic limit, and organic matter content.

[0008] The technical solution adopted in the present invention is as follows:

[0009] 1. A soil intelligent classification automation system includes:

[0010] Soil sample pretreatment module, used for air-drying, stirring and weighing the soil samples to be processed;

[0011] The soil sample particle size classification and screening module is used to screen the pre-treated soil sample step by step to obtain the first-level particle distribution result, and obtain the coarse-screened soil sample and the fine-screened soil sample from the screened soil sample;

[0012] The liquid and plastic limit automatic measurement module is used to prepare the coarsely screened soil sample into soil pastes with different moisture contents, and measure them using the cone oscillator method to obtain the cone oscillator method measurement results;

[0013] The burning measurement module is used to burn the soil sample to be processed and obtain the corresponding burning measurement results;

[0014] Densitometer measurement module, used to perform particle analysis on finely sieved soil samples using the densitometer method to obtain the second-level particle distribution results;

[0015] The control and analysis module is used to realize the automatic control of the soil intelligent classification automation system, receive the first-level particle distribution results, the cone instrument measurement results, the ignition measurement results and the second-level particle distribution results, perform data analysis, and obtain the soil sample type and the corresponding optimal solidification method.

[0016] Specifically, when the control analysis module implements data analysis, the following steps are performed:

[0017] The estimated moisture content is obtained based on the burning measurement results corresponding to the pretreated soil samples;

[0018] The estimated liquid and plastic limits are calculated based on the estimated moisture content and the cone instrument measurement results;

[0019] Determine the water volatilization time and organic matter volatilization time based on the estimated liquid and plastic limits, and then obtain the moisture content and organic matter content of the soil sample to be treated;

[0020] The liquid and plastic limits are calculated based on the moisture content of the soil sample to be treated and the results of cone oscillator method measurement;

[0021] The particle grading parameters are obtained according to the first-level particle distribution results and the second-level particle distribution results;

[0022] According to the particle gradation parameters, moisture content, organic matter content and liquid limit and plastic limit, the soil sample type is obtained; the soil sample types include conventional type, high liquid limit and high plasticity type, high organic matter type, gradation adjustable type and high moisture content type;

[0023] Output soil sample type and corresponding optimal curing method.

[0024] Specifically, the soil sample pretreatment module includes an air-drying bin, a stirring bin and a weighing bin; the soil sample to be processed is air-dried in the air-drying bin, stirred in the stirring bin and weighed in the weighing bin in turn; a number of fans are provided in the air-drying bin, a first stirrer and a camera are provided in the stirring bin, and a first weighing device is provided in the weighing bin; the fan, camera, first stirrer and first weighing device are all electrically and / or communicatively connected to the control module; the control and analysis module can control the stirrer according to the real-time image collected by the camera, and use the internally stored air-drying control model to process the initial parameters of the soil sample to be processed, predict the optimal air-drying parameters, and then control the fan according to the optimal air-drying parameters; the initial parameters include initial humidity, particle distribution and particle adsorption parameters.

[0025] The soil sample particle size grading screening module includes a screen, a screening instrument casing, a vibration unit and a soil sample conveyor belt; a cavity is provided in the screening instrument casing as a screening chamber, a vibration unit is provided in the center of the screening chamber to separate the screening chamber into a coarse screening chamber and a fine screening chamber, and a group of screening components and a soil sample conveyor belt are arranged above and below the coarse screening chamber and the fine screening chamber, respectively, the screening component is mainly composed of a number of screens with different apertures, and the various screens are arranged in sequence from top to bottom, and the apertures decrease in sequence; the vibration unit includes a shock-absorbing spring, a visual sensor, a vibration partition and a vibration machine; the vibration machine is installed on the bottom inner wall of the screening chamber, and the output shaft of the vibration machine is connected to the vibration partition The lower end of the plate is connected, and the upper and lower ends of the vibrating partition are respectively connected to the top and bottom inner walls of the screening chamber through at least one shock-absorbing spring; a number of visual sensors are installed on the vibrating partition, and the number of the visual sensors and the screen is the same and one-to-one corresponding, and the visual sensors are arranged above the corresponding screen; the visual sensors and the vibration machine are electrically and / or communicatively connected to the control and analysis module; the control and analysis module can receive the real-time image collected by the visual sensor, extract the real-time screening state parameters from the real-time image, and use the screening control model to process it to obtain the predicted vibration state of the vibrating partition, and control the vibration machine according to the predicted vibration state of the vibrating partition.

[0026] The screening state parameters include the coverage area ratio of soil sample particles on the screen, the average particle size and the particle pass rate, and the particle pass rate is expressed as the relative value of the average particle size and the screen aperture; the predicted vibration state includes the predicted vibration amplitude, predicted frequency and predicted mode.

[0027] The soil sample particle size grading and screening module also includes a coarse screen soil inlet, a fine screen soil inlet, a sieve soil collection box and a controllable gate; the tops of the coarse screen chamber and the fine screen chamber are respectively provided with a coarse screen soil inlet and a fine screen soil inlet; the inner sides of the screen and the soil sample conveyor belt are connected to the vibrating partition, and the outer sides are connected to the inner wall of the screening chamber, and controllable gates are correspondingly arranged at the connection between the screen / soil sample conveyor belt and the inner wall of the screening chamber, and a sieve soil collection box is correspondingly arranged on the outer side of each controllable gate, and the sieve soil collection box is installed on the outer side wall of the screening instrument casing; a second weighing device is installed on the sieve soil collection box, and the second weighing device is communicatively connected to the control and analysis module.

[0028] The liquid-plastic limit automated measurement module includes a mixing drum, an aluminum box conveyor belt, multiple aluminum boxes, a scraping blade group, and a photoelectric combined measuring instrument; the mixing drum is used to gradually add water to the finely screened soil sample to obtain soil paste with different moisture contents; the aluminum box is used to fill the soil paste; the loading end of the aluminum box conveyor belt is used to place the aluminum box, and the discharging end is sequentially arranged with a scraping blade group and a photoelectric combined measuring instrument along the conveying direction. The photoelectric combined measuring instrument is equipped with multiple cone meters, the number of which is the same as the number of aluminum boxes and corresponds one to one. The photoelectric combined measuring instrument is equipped with a sensor. When the aluminum box reaches a preset position, the sensor automatically senses and triggers the conveyor belt to decelerate to a stop, causing the aluminum box to move directly below the corresponding cone meter.

[0029] The burning measurement module includes a high-precision scale, an insulating stone, a burning dish and a flame head; the densitometer measurement module includes a soil material nozzle, a pure water nozzle, a dispersant nozzle, a second agitator, a densitometer, a thermometer and a soil material container; the soil material container is placed on a conveying device, and a bracket is provided above the conveying device, and the bracket is telescopically connected to the soil material nozzle, the pure water nozzle, the dispersant nozzle, the second agitator, the densitometer and the thermometer in sequence along the conveying direction.

[0030] 2. An automated method for intelligent soil classification

[0031] The method comprises the following steps:

[0032] S1) using a soil sample pretreatment module to air-dry, stir, and weigh the soil sample to be treated, to obtain the pretreated soil sample and its total weight;

[0033] S2) using the soil sample particle size classification and screening module to perform step-by-step screening on the pretreated soil sample, obtain the first-level particle distribution result and transmit it to the control and analysis module, take the last level soil sample output by the coarse screening chamber as the coarse screened soil sample, and take the last level soil sample output by the fine screening chamber as the fine screened soil sample;

[0034] S3) preparing the coarsely screened soil sample into soil pastes with different moisture contents by the liquid-plastic limit automated measurement module, measuring the soil pastes using the cone oscillator method, obtaining the cone oscillator measurement results, and transmitting them to the control and analysis module;

[0035] S4) burning the soil sample to be processed using the burning measurement module until the time rate of change of mass is less than a preset threshold and lasts for a preset time period, obtaining a mass-time curve as the burning measurement result, and transmitting it to the control and analysis module;

[0036] S5) using a control analysis module to analyze the mass-time curve obtained in step S4, and calculating an estimated moisture content based on the total mass loss; calculating the estimated moisture content of each level of soil paste prepared in step S3 based on the estimated moisture content, and combining the cone instrument measurement results to calculate the estimated liquid limit and plastic limit; determining the water volatilization time and organic matter volatilization time based on the estimated liquid limit, thereby obtaining the moisture content and organic matter content of the soil sample to be processed;

[0037] S6) calculating the moisture content of each level of soil paste in step S3 according to the moisture content of the soil sample to be processed, and combining the measurement results of the cone oscillator method to calculate the liquid limit and plastic limit of the soil sample to be processed;

[0038] S7) using a density meter measurement module to perform particle analysis on the finely screened soil sample using a density meter method to obtain a second-level particle distribution result, which is transmitted to a control and analysis module; the control and analysis module obtains a particle grading parameter based on the first-level particle distribution result and the second-level particle distribution result;

[0039] S8) using the control analysis module to obtain the soil sample type according to the particle gradation parameters, moisture content, organic matter content and liquid limit of the soil sample to be processed;

[0040] S9) The control analysis module outputs the soil sample type and the corresponding optimal solidification method.

[0041] The method further includes: realizing automatic control of the soil intelligent classification automation system through the control and analysis module. This step is specifically as follows:

[0042] In step S1, before air-drying, the initial humidity, particle distribution, and particle adsorption parameters of the soil sample to be processed are first input into the air-drying control model of the control and analysis module. The air-drying control model predicts and outputs the optimal air-drying parameters. The control and analysis module converts the optimal air-drying parameters into instructions and outputs them to the fan. During the stirring process, the control and analysis module processes the real-time image captured by the camera through an image processing algorithm to obtain the uniformity parameter of the soil sample. When the uniformity parameter of the soil sample reaches a preset threshold, the stirring ends. After weighing, the first weigher transmits the total weight of the pre-treated soil sample to the control and analysis module.

[0043] In step S2, the real-time image collected by the visual sensor is input into the control and analysis module, the real-time screening state parameters are extracted from the real-time image, and the screening control model is used to process the real-time image to obtain the predicted vibration state of the vibration partition. The control and analysis module controls the vibration machine according to the predicted vibration state of the vibration partition.

[0044] In summary, the innovation of the present invention lies in proposing a soil intelligent classification automation system process method that combines machine learning and intelligent control. It not only opens up the experimental data chain of sieving method, densitometer method and liquid limit and plastic limit joint determination, but also integrates differential soil identification and solidification parameter recommendation modules. The system automatically collects and analyzes multi-dimensional soil parameters (such as liquid limit and plastic limit, moisture content, particle grading, organic matter content, etc.), builds a classification-adaptation integrated model based on machine learning algorithms, automatically identifies high-risk parameters (such as liquid limit>60%, organic matter>4.3%, moisture content exceeds the limit) and matches the corresponding differential solidification strategies (such as lime-fly ash modification, cement-gypsum composite solidification, coarse particle deployment, etc.), and realizes the automation of the whole process from "intelligent classification" to "directional modification". Different from the problems of fragmented operation of existing equipment and inconsistent experimental processes, the present invention realizes the integrated operation of testing, analysis, optimization and recommendation through a unified intelligent control platform. The system can adjust parameter settings based on real-time data streams, improve test efficiency, reduce errors, and achieve a truly intelligent, automated, and multi-scenario adaptable soil classification and solidification integrated solution.

[0045] The beneficial effects of the present invention are:

[0046] (1) The present invention has a higher level of automation integration: By constructing an integrated automated control system, the present invention connects traditional classification processes such as soil sample pretreatment, screening, density testing, and liquid limit and plastic limit determination, achieving full automation of the soil classification test process. Compared with the manual segmented operation and independent equipment operation mode commonly found in the prior art, the present invention can significantly reduce manual intervention, unify the operation process, and effectively improve operational efficiency and control accuracy.

[0047] (2) The present invention improves test accuracy and consistency: The system automatically performs multiple tests, eliminating errors caused by manual operation. The automated control system can adjust key test parameters (such as temperature, humidity, and pressure) through real-time monitoring and sensor feedback, ensuring precise control during the test process, thereby ensuring the consistency and accuracy of each test result.

[0048] (3) The present invention enables fully automated closed-loop execution: Unlike existing technologies that still require some manual operations, the present invention's automated soil classification system achieves full automation. All testing steps (including soil sample pretreatment, screening, densimeter method, and liquid-plastic limit method) are controlled by an automated control system. The system optimizes the coordination and scheduling of each unit through specific algorithms, ensuring that each step is executed in an optimal state.

[0049] (5) Based on the automated preprocessing and multi-parameter testing of soil samples (including liquid limit, plastic limit, particle grading, moisture content, organic matter content, etc.), the present invention determines the soil type based on the collected multi-dimensional soil parameters, and simultaneously calls and outputs the pre-set differentiated solidification method corresponding to each soil type. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Attachment Figure 1 2. It is a plan view of the soil intelligent classification automation system according to an embodiment of the present invention;

[0051] Attachment Figure 2 1 is a plan view of a soil sample pretreatment module according to an embodiment of the present invention;

[0052] Attachment Figure 3 yes Figure 2 A magnified diagram of the section A in the middle: soil particles are slightly broken up by stirring, and stuck-together particles are simply broken up. Nearly half of the soil sample is then used for subsequent experiments.

[0053] Attachment Figure 4 yes Figure 2 Enlarged schematic diagram of the part B in the middle: accelerated air drying of soil particles;

[0054] Attachment Figure 5 Schematic diagram of a soil sample particle size classification and screening module according to an embodiment of the present invention;

[0055] Attachment Figure 6 Schematic diagram of the liquid-plastic limit automated measurement module in an embodiment of the present invention;

[0056] Attachment Figure 7 Schematic diagram of a density meter measurement module and a burning measurement module in an embodiment of the present invention;

[0057] Attachment Figure 8 Schematic diagram of a control and analysis module in an embodiment of the present invention.

[0058] In the figure, 1. soil sample pretreatment module, 11. air drying chamber, 12. mixing chamber, 13. weighing chamber, 14. fan, 15. bottom plate, 16. first stirrer, 17. first weigher, 18. soil sample outlet, 2. soil sample particle size classification and screening module, 21. coarse screen soil inlet, 22. fine screen soil inlet, 23. screen, 24. sieve analyzer housing, 25. soil collection box after screening, 26. shock-absorbing spring, 27. visual sensor, 28. vibration machine, 29. conveyor belt, 210. Control gate, 3. Liquid and plastic limit automatic measurement module, 31. Conveyor belt, 32. Aluminum box, 33. Scraping blade assembly, 34. Photoelectric combined measuring instrument, 4. Density meter measurement module, 41. Soil material nozzle, 42. Dispersant nozzle, 43. Pure water nozzle, 44. Second agitator, 45. Density meter, 46. Thermometer, 47. Soil material container, 5. Burning measurement module, 51. High-precision scale, 52. Insulation stone, 53. Burning dish, 54. Flame head, 6. Control and analysis module. DETAILED DESCRIPTION

[0059] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] The first aspect of the present invention provides an automated system for intelligent soil classification. The system of the present invention significantly improves the automation and intelligence level of the classification process by integrating sieve analysis, densitometer method and liquid limit and plastic limit combined determination method. During the experiment, the machine learning algorithm analyzes the multi-dimensional data collected in real time (such as particle size distribution, moisture content, liquid limit and plastic limit, organic matter content, etc.), dynamically optimizes the experimental parameters and adjusts the classification process, which not only realizes the dynamic optimization of soil classification, but also accurately identifies the key indicators that affect the curing performance, and assists in generating differential curing strategy recommendations based on the soil sample type. Through real-time monitoring and closed-loop feedback mechanism, the system can simultaneously optimize test efficiency and classification accuracy, realize intelligent closed-loop optimization of the entire process from data collection, classification judgment to differentiated curing recommendations, and provide highly adaptable solutions for complex engineering scenarios.

[0061] like Figure 1 As shown, the automation system of the present invention includes:

[0062] Soil sample pretreatment module 1, used for air-drying, stirring and weighing the soil sample to be processed;

[0063] Soil sample particle size classification and screening module 2 is used to screen the pre-treated soil sample step by step to obtain the first-level particle distribution result, and obtain the coarse-screened soil sample and the fine-screened soil sample from the screened soil sample for cone oscillator measurement and density meter analysis respectively;

[0064] The liquid and plastic limit automatic measurement module 3 is used to prepare the coarsely screened soil sample into soil pastes with different moisture contents, and measure them using the cone meter method to obtain the cone meter method measurement results;

[0065] The burning measurement module 5 is used to burn the soil sample to be processed and obtain the corresponding burning measurement result;

[0066] Densitometer measurement module 4, used to perform particle analysis on the finely sieved soil sample using a densitometer method to obtain a second-level particle distribution result;

[0067] The control and analysis module 6 is used to realize the automatic control of the soil intelligent classification automation system, receive the first-level particle distribution results, cone instrument measurement results, ignition measurement results and second-level particle distribution results, perform data analysis, and obtain the soil sample type and the corresponding optimal solidification method.

[0068] Specifically, if Figure 8 The control analysis module 6 shown implements data analysis by executing the following steps:

[0069] The estimated moisture content is obtained based on the burning measurement results corresponding to the pretreated soil samples;

[0070] The estimated liquid and plastic limits are calculated based on the estimated moisture content and the cone instrument measurement results;

[0071] Determine the water volatilization time and organic matter volatilization time based on the estimated liquid and plastic limits, and then obtain the moisture content and organic matter content of the soil sample to be treated;

[0072] The liquid and plastic limits are calculated based on the moisture content of the soil sample to be treated and the results of cone oscillator method measurement;

[0073] The particle grading parameters are obtained according to the first-level particle distribution results and the second-level particle distribution results;

[0074] According to the particle grading parameters, moisture content, organic matter content and liquid limit and plastic limit, the soil sample type is obtained; the soil sample types include conventional type, high liquid limit and high plasticity type, high organic matter type, gradation adjustable type and high moisture content type;

[0075] Output soil sample type and corresponding optimal curing method.

[0076] Specifically, if Figures 2 to 4As shown, the soil sample pretreatment module 1 includes an air-drying bin 11, a stirring bin 12 and a weighing bin 13; the soil samples to be processed are air-dried in the air-drying bin 11, stirred in the stirring bin 12 and weighed in the weighing bin 13 in sequence; a plurality of fans 14 are provided in the air-drying bin 11 for air-drying the soil samples to be processed in the air-drying bin 11, a first stirrer 16 and a camera 19 are provided in the stirring bin 12, the first stirrer 16 is used to stir the air-dried soil samples in the stirring bin 12, the camera 19 is used to collect images of the air-dried soil samples in the stirring bin 12 in real time during the stirring process, and a first weighing device 17 is provided in the weighing bin 13 for weighing the stirred soil samples; the fan 14, the camera 19, the first stirrer 16 and the first weighing device 17 are all electrically connected and / or communicatively connected to the control module 7; the real-time image data collected by the camera 19 and the soil sample mass collected by the first weighing device 17 are all transmitted to the control and analysis module 6. The control and analysis module 6 can control the agitator 16 according to the real-time image captured by the camera 19, and use the internally stored air-drying control model to process the initial parameters of the soil sample to be processed, predict the optimal air-drying parameters, and then control the fan 14 according to the optimal air-drying parameters; the initial parameters include initial humidity, particle distribution and particle adsorption parameters.

[0077] In some optional embodiments, the air-drying bin 11 , the stirring bin 12 and the weighing bin 13 are arranged in sequence from top to bottom.

[0078] In the above-mentioned optional embodiment, the top of the air-drying chamber 11 is provided with an opening serving as a feed port, and a plurality of fans 14 are mounted on the side walls for air-drying the soil sample to be processed. The bottom plate 15 is an openable structure. After air-drying is completed, the bottom plate 15 is opened, connecting the air-drying chamber 11 with the mixing chamber 12, and the air-dried soil sample in the air-drying chamber 11 falls into the mixing chamber 12 under the action of gravity.

[0079] In this optional embodiment, a camera 19 is mounted on the inner sidewall of the mixing chamber 12 for real-time monitoring of the soil sample's condition. A first agitator 16 is located at the bottom for uniformly mixing the soil sample. A retractable chassis is located below the first agitator 16. After mixing is complete, the retractable chassis is withdrawn, allowing the stirred soil sample to fall under gravity onto the weighing pan of the first scale 17 for weighing. After weighing, the stirred soil sample is output from the soil sample pretreatment module 1 through the soil sample outlet 18.

[0080] In some optional embodiments, the control and analysis module 6 processes the real-time image captured by the camera 19 through an image processing algorithm to obtain a uniformity parameter of the soil sample. When the uniformity parameter of the soil sample reaches a preset threshold, the stirring ends.

[0081] Specifically, if Figure 5As shown, the soil sample particle size grading screening module 2 includes a screen 23, a screening instrument housing 24, a vibration unit and a soil sample conveyor belt 29; a cavity is provided in the screening instrument housing 24 as a screening chamber, and a vibration unit is provided in the center of the screening chamber. The vibration unit divides the screening chamber into a coarse screening chamber and a fine screening chamber. A group of screening components and a soil sample conveyor belt 29 are arranged above and below the coarse screening chamber and the fine screening chamber respectively. The soil sample conveyor belt 29 is located below the screening component. The screening component is mainly composed of a number of screens 23 with different apertures. In the same group of screening components, the screens 23 are arranged in sequence from top to bottom, and the apertures decrease in sequence; the vibration unit includes a shock-absorbing spring 26, a visual sensor 27, a vibration partition and a vibrator 28; the vibrator 28 is installed at the center of the bottom inner wall of the screening chamber, and the vibration machine 28 is arranged at the bottom inner wall of the vibrator 28. The output shaft is connected to the lower end of the vibrating partition, and the upper and lower ends of the vibrating partition are respectively connected to the top and bottom inner walls of the screening chamber through at least one shock-absorbing spring 26; a number of visual sensors 27 are installed on the vibrating partition, and the number of visual sensors 27 and the screen 23 are the same and one-to-one corresponding. The visual sensor 27 is arranged above the corresponding screen 23 and is used to collect the distribution image of the particles on the screen 23 in real time; the visual sensor 27 and the vibration machine 28 are both electrically and / or communicatively connected to the control and analysis module 6; the control and analysis module 6 can receive the real-time image collected by the visual sensor 27, extract the real-time screening state parameters from the real-time image, and use the screening control model to process it to obtain the predicted vibration state of the vibrating partition, and control the vibration machine 28 according to the predicted vibration state of the vibrating partition.

[0082] Specifically, the screening state parameters include the coverage area ratio of soil sample particles on the sieve, the average particle size and the particle pass rate. The particle pass rate is expressed as the relative value of the average particle size and the sieve aperture.

[0083] Specifically, predicting the vibration state includes predicting the vibration amplitude, predicting the frequency, and predicting the mode.

[0084] Specifically, the soil sample particle size grading and screening module 2 also includes a coarse screen soil inlet 21, a fine screen soil inlet 22, a sieved soil collection box 25 and a controllable gate 210; the coarse screen chamber and the fine screen chamber are respectively provided with a coarse screen soil inlet 21 and a fine screen soil inlet 22 on the top; the inner side of the screen 23 and the soil sample conveyor belt 29 (i.e., the side close to the vibrating partition) are connected to the vibrating partition, and the outer side (i.e., the side away from the vibrating partition) is connected to the inner wall of the screening chamber, and a controllable gate 210 is correspondingly arranged at the connection between the screen 23 / soil sample conveyor belt 29 and the inner wall of the screening chamber, and a sieved soil collection box 25 is correspondingly arranged on the outside of each controllable gate 210, and the sieved soil collection box 25 is installed on the outer wall of the screening instrument housing 24; a second weighing device is installed on the sieved soil collection box 25, and the second weighing device is communicatively connected to the control and analysis module 6.

[0085] In some optional embodiments, the controllable gate 210 is electrically connected or in communication connection with the control and analysis module 6. The control and analysis module 6 controls the opening or closing of the controllable gate 210 by sending instructions to the controllable gate 210.

[0086] In some optional embodiments, the soil sample conveyor belt 29 is electrically connected or in communication connection with the control and analysis module 6. The control and analysis module 6 controls the soil sample conveyor belt 29 to be turned on or off by sending instructions to the soil sample conveyor belt 29.

[0087] In some optional embodiments, the side of the vibrating partition is connected to the inner wall of the sieving chamber through a guide structure.

[0088] In the above optional embodiments, the guide structure includes but is not limited to a slide rail, a roller guide device, etc.

[0089] In some optional embodiments, feed sensors can be installed at the coarse screen inlet 21 and the fine screen inlet 22 respectively to monitor the feed rate and material flow status in real time to avoid blockage or overload and ensure that the material enters the screening system evenly and continuously.

[0090] In the above optional embodiments, the feed sensor includes but is not limited to an industrial camera, a photoelectric sensor, a capacitive sensor, a vibration sensor, and the like.

[0091] Specifically, if Figure 6 As shown, the automated liquid-plastic limit measurement module 3 includes a mixing drum, an aluminum box conveyor belt 31, multiple aluminum boxes 32, a scraping blade assembly 33, and a photoelectric combined measuring instrument 34. The photoelectric combined measuring instrument 4 uses electromagnetic control to drop the cone gauge, and a sensor reads the reading. The mixing drum is used to gradually add water to the finely screened soil sample to obtain soil pastes with different moisture contents; the aluminum boxes 32 are used to fill the soil paste; the loading end of the aluminum box conveyor belt 31 is used to place the aluminum boxes 32, and the scraping blade assembly 33 and the photoelectric combined measuring instrument 34 are arranged in sequence along the conveying direction at the discharge end. The photoelectric combined measuring instrument 34 is equipped with multiple cone gauges, the same number as the aluminum boxes 32, and a one-to-one correspondence. The photoelectric combined measuring instrument 34 is equipped with a sensor. When the aluminum box 32 reaches a preset position, the sensor automatically senses and triggers the conveyor belt to slow down to a stop, causing the aluminum box 32 to move directly below the corresponding cone gauge.

[0092] In some optional embodiments, the sensor is a position sensor or a proximity sensor.

[0093] Specifically, if Figure 7As shown, the burning measurement module 5 includes a high-precision scale 51, an insulating stone 52, a burning dish 53 and a flame head 54 arranged in sequence from bottom to top; the density meter measurement module 4 includes a soil material nozzle 41, a dispersant nozzle 42, a pure water nozzle 43, a second agitator 44, a density meter 45, a thermometer 46 and a soil material container 47; the soil material container 47 is placed on a conveying device, and a bracket is provided above the conveying device, on which the soil material nozzle 41, the dispersant nozzle 42, the pure water nozzle 43, the second agitator 44, the density meter 45 and the thermometer 46 are telescopically connected in sequence along the conveying direction.

[0094] A second aspect of the present invention provides an automated method for intelligent soil classification.

[0095] The method of the present invention comprises the following steps:

[0096] S1) using the soil sample pretreatment module 1 to air-dry, stir and weigh the soil sample to be treated, to obtain the pretreated soil sample and its total weight;

[0097] S2) using the soil sample particle size classification and screening module 2 to perform step-by-step screening on the pre-treated soil sample, obtain the first-level particle distribution result and transmit it to the control and analysis module 6, take the last level soil sample output by the coarse screening chamber as the coarse screened soil sample, and take the last level soil sample output by the fine screening chamber as the fine screened soil sample;

[0098] S3) The coarsely screened soil sample is prepared into soil pastes with different moisture contents by the liquid-plastic limit automated measurement module 3, and the soil pastes are measured using the cone oscillator method. The cone oscillator measurement results are obtained and transmitted to the control and analysis module 6;

[0099] S4) burning the soil sample to be processed using the burning measurement module 5 until the time rate of change of mass is less than a preset threshold and lasts for a preset time period, obtaining a mass-time curve as a burning measurement result, and transmitting it to the control and analysis module 6;

[0100] S5) using the control analysis module 6 to analyze the mass-time curve obtained in step S4, and calculate the estimated moisture content based on the total mass loss;

[0101] The estimated liquid and plastic limits are calculated based on the estimated moisture content of each grade of soil paste prepared in step S3 and the results of cone oscillator measurement;

[0102] Determine the water volatilization time and organic matter volatilization time based on the estimated liquid and plastic limits, and then obtain the moisture content and organic matter content of the soil sample to be treated;

[0103] S6) calculating the moisture content of each level of soil paste in step S3 according to the moisture content of the soil sample to be processed, and combining the measurement results of the cone oscillator method to calculate the liquid limit and plastic limit of the soil sample to be processed;

[0104] S7) using the density meter measurement module 4 to perform particle analysis on the finely screened soil sample using a density meter method to obtain a second-level particle distribution result, which is transmitted to the control and analysis module 6;

[0105] The control and analysis module 6 obtains the particle gradation parameters according to the first-level particle distribution results and the second-level particle distribution results;

[0106] S8) using the control and analysis module 6 to obtain the soil sample type according to the particle gradation parameters, moisture content, organic matter content and liquid limit of the soil sample to be processed;

[0107] S9) The control analysis module 6 outputs the soil sample type and the corresponding optimal solidification method.

[0108] Furthermore, the method of the present invention also includes realizing automatic control of the soil intelligent classification automation system through the control analysis module 6. The specific process of automatic control is:

[0109] In step S1, before air-drying, the initial humidity, particle distribution, and particle adsorption parameters of the soil sample to be processed are first input into the air-drying control model of the control and analysis module 6. The air-drying control model predicts and outputs the optimal air-drying parameters. The control and analysis module 6 converts the optimal air-drying parameters into instructions and outputs them to the fan 14. During the stirring process, the control and analysis module 6 processes the real-time image captured by the camera 19 through an image processing algorithm to obtain the uniformity parameter of the soil sample. When the uniformity parameter of the soil sample reaches a preset threshold, the stirring ends. After weighing, the first weigher 17 transmits the total weight of the pre-treated soil sample to the control and analysis module 6.

[0110] In step S2, the real-time image collected by the visual sensor 27 is input into the control and analysis module 6, the real-time screening state parameters are extracted from the real-time image, and the screening control model is used to process the real-time image to obtain the predicted vibration state of the vibration partition. The control and analysis module 6 controls the vibration machine 28 according to the predicted vibration state of the vibration partition.

[0111] Optionally, in step S5, the process of determining the water volatilization time and the organic matter volatilization time based on the estimated liquid and plastic limits, and then obtaining the moisture content and organic matter content of the soil sample to be processed can be achieved by any of the following steps:

[0112] a. Input the water volatilization time and the organic matter volatilization time into the mass-time curve obtained in step S4, and calculate the moisture content and organic matter content based on the mass corresponding to the water volatilization time and the organic matter volatilization time;

[0113] b. Burn the soil samples according to the water volatilization time and organic matter volatilization time to obtain the moisture content and organic matter content.

[0114] The specific embodiments of the present invention are as follows:

[0115] Example

[0116] like Figure 1 As shown, the soil intelligent classification automation system of this embodiment includes a soil sample pretreatment module 1, a soil sample particle size classification and screening module 2, a liquid limit and plastic limit automatic measurement module 3, a densitometer measurement module 4, an ignition measurement module 5 and a control and analysis module 6.

[0117] 1. Soil sample pretreatment module 1

[0118] like Figure 2 As shown, the soil sample pretreatment module 1 includes an air drying chamber 11 , a stirring chamber 12 , a weighing chamber 13 , a fan 14 , a bottom plate 15 , a first stirrer 16 , a first weigher 17 and a soil sample outlet 18 .

[0119] The module places a predetermined mass (m0) of soil sample into the air-drying chamber and uses the control and analysis module 6 to predict the air-drying parameters, wind speed and drying time, to achieve uniform air-drying. Specifically, a machine learning algorithm is used to predict the optimal air-drying parameters based on the soil sample's initial moisture content, particle distribution, and particle adsorption properties, effectively improving drying efficiency and ensuring uniform moisture distribution.

[0120] In this embodiment, the air-drying control model specifically adopts a prediction model based on support vector regression, which takes the initial moisture content, particle distribution, and adsorption parameters of the soil sample as input and outputs the optimal combination of wind speed and air-drying time.

[0121] After air-drying is complete, the control and analysis module 6 precisely controls the opening of the openable bottom plate 15 of the air-drying chamber 11, transferring the air-dried soil sample to the mixing sampling layer. During the mixing process, the control and analysis module 6 adjusts the mixing duration in real time based on the soil sample's characteristics to disperse agglomerated particles while avoiding disruption of the particle structure. In this embodiment, the soil sample's characteristics are represented by its uniformity parameter.

[0122] The soil sample pretreatment module 1 simulates the traditional quartering method, dividing and mixing the sample multiple times to ensure its representativeness, and accurately transmits it to the high-precision weighing module. After weighing, the data is automatically recorded to provide high-quality sample support for subsequent experiments.

[0123] 2. Soil sample particle size classification and screening module 2

[0124] like Figure 3 As shown, the soil sample particle size grading and screening module 2 includes a coarse screen soil inlet 21, a fine screen soil inlet 22, a screen 23, a screening instrument housing 24, a screened soil collection box 25, a shock-absorbing spring 26, a visual sensor 27, a vibrator 28, a soil sample conveyor belt 29 and a controllable gate 210.

[0125] ① Coarse screen soil inlet 21 and fine screen soil inlet 22

[0126] The top of the sieve analyzer housing 24 features two feed ports: a coarse sieve inlet 21 and a fine sieve inlet 22, for adding coarse and fine materials, respectively. This design allows for preliminary separation based on the particle size characteristics of the soil sample, preventing particle size mixing. Furthermore, feed sensors can be installed at both the coarse sieve inlet 21 and the fine sieve inlet 22 to monitor the feed rate and material flow in real time, preventing blockages or overloads and ensuring uniform and continuous material entry into the screening system.

[0127] ② Screen 23 and soil sample conveyor belt 29

[0128] The core of the soil sample particle size classification and screening module 2 is the multi-layer screen 23. The apertures of each layer of screen 23 are different and gradually decrease from top to bottom, which can effectively separate soil samples of different particle sizes. The bottom layer of soil sample conveyor belt 29 is used to collect the bottom layer of soil samples.

[0129] In this embodiment, the pore sizes of the screen 23 in the left coarse screening chamber are 60 mm, 40 mm, 20 mm, 10 mm, 5 mm, and 2 mm, respectively, from top to bottom. The pore sizes of the screen 23 in the right coarse screening chamber are 2.0 mm, 1.0 mm, 0.5 mm, 0.25 mm, 0.1 mm, and 0.075 mm, respectively, from top to bottom.

[0130] like Figure 3 As shown, in this embodiment, the inclination angle of the screen is initially set to 45°, and the height of the outer side is lower than that of the inner side.

[0131] In this embodiment, the screen angle is dynamically adjusted within a range of 30° to 60° to optimize screening efficiency. Dynamic adjustment of the screen angle is achieved by dynamically controlling the up and down vibration of the vibrating diaphragm. The maximum and minimum screen angles correspond to the amplitude of the vibrating diaphragm.

[0132] ③ Shock-absorbing spring 26, visual sensor 27, vibration machine 28

[0133] To optimize the screening process, the system uses visual sensors 27 to collect real-time data on screening efficiency, fluidity, and particle pass rate, and combines it with a reinforcement learning model to dynamically adjust the screen angle and screening strategy, thereby improving classification accuracy and efficiency.

[0134] The vibrating machine 28 has a stable power output capability. Through the energy consumption optimization algorithm, the control and analysis module 6 dynamically adjusts the motor output to meet the screening requirements while reducing energy consumption.

[0135] In this embodiment, the visual sensor 27 is an image collector. The control and analysis module 6 uses the image collector to obtain a real-time distribution image of the particles on the screen, processes the image, and extracts the coverage area of ​​the particles on the screen, the average particle size, and its relative relationship with the screen aperture. In combination with real-time monitoring data, the system's control and analysis module 6 uses a supervised learning algorithm to dynamically adjust the vibration amplitude, vibration frequency (range: 10Hz to 50Hz), and mode (high frequency, low frequency, or periodic switching) based on the blockage risk and pass rate optimization goals to ensure maximum screening efficiency and the integrity of the soil sample particles.

[0136] In this embodiment, the screening control model specifically adopts a control model based on deep reinforcement learning.

[0137] ④Controllable gate 210, sieve soil collection box 25

[0138] The sieve analyzer housing 24 is equipped with multiple outlets on the side, corresponding to the particle size classification results. Coarse material is discharged through the left outlet, and fine material is collected through the right outlet. Each outlet is equipped with a controllable gate 210 and a sieve soil collection box 25. A high-precision weighing module and automatic recording device are also provided to measure the weight of the sample at each particle size classification in real time.

[0139] The coarse-screened soil sample discharged from the leftmost lower outlet is transported to the fine-screen soil inlet 22 for fine screening.

[0140] 3. Liquid and plastic limit automatic measurement module 3

[0141] After the system screens out soil samples with a particle size of less than 0.5 mm through the soil sample particle size classification and screening module 2, the sample is automatically collected into the mixing drum and uniformly stirred to form an initial soil paste, which is then squeezed into the aluminum box 32 below for subsequent testing.

[0142] During the test, water is added incrementally in set amounts to produce a multi-level soil paste. The evenly mixed soil sample is automatically loaded into a fixed aluminum box 32, which moves along an aluminum box conveyor belt 31 and passes through a flattening blade assembly 33 before entering the cone test area. The flattening blade assembly 33 consists of five independent scrapers, ensuring a smooth and even surface for the soil sample within the aluminum box 32. When the aluminum box 32 reaches the predetermined position within the cone tester, a sensor automatically senses and controls the aluminum box conveyor belt 31 to slow down to a stop, precisely positioning the aluminum box 32 below the cone tester for automated determination of the liquid and plastic limits. The system supports flexible adjustment of the automated process based on experimental requirements.

[0143] During the liquid and plastic limit measurement, the combined dry soil mass (m d ) remains constant, and the change in the total mass of the soil paste only comes from the incremental mass of water added each time. By monitoring and controlling the mass of water added each time (Δm ωi ) can accurately calculate the moisture content at each level (ωi The initial moisture content is determined by the burning method, and the initial mass of the soil sample (m1), the dry soil mass (m d ) and the initial moisture content (ω0) are known, so there is no need to measure the moisture content of the soil paste by burning it step by step. The moisture content of each level of soil paste can be directly calculated quickly using the following formula, thereby further improving the experimental efficiency.

[0144]

[0145] Where, ω x is the moisture content of the last level (i.e. level x) soil paste (%), Δm ωi is the mass of water added for the i-th time, m d is the mass of dry soil. Combined with the mass of water added successively (Δm ωi ) can be used to further calculate the moisture content of the first two levels.

[0146] The main advantage of this method is that in the process of preparing the soil paste, it is only necessary to add water to the mixing drum in set amounts. The subsequent determination of the liquid limit and plastic limit only requires one moisture content test, which significantly reduces the complexity of the experimental operation. The automated process also significantly reduces the random errors that may be introduced by manual operation, such as uneven water addition or non-standard scraping of the soil sample.

[0147] At the same time, the automation system can ensure high precision and repeatability during the experiment, improving experimental efficiency and data accuracy.

[0148] The system design can collect more data on the relationship between moisture content and cone sinking depth. This data provides rich support for the study of soil sample liquid plasticity and significantly enhances the reliability of experimental results and the accuracy of subsequent analysis. After being evenly mixed, the soil sample is automatically loaded into a fixed aluminum box. The aluminum box moves along the conveyor belt and enters the cone test area after passing through the scraping blade group 33. The scraping blade group 33 consists of five independent scrapers to ensure that the surface of the soil sample in the aluminum box is flat and uniform. When the aluminum box reaches the predetermined position of the cone instrument, the sensor automatically senses and controls the conveyor belt to slow down to a stop, accurately positioning the aluminum box under the cone instrument to complete the automated determination of the liquid limit and plastic limit. The system supports flexible adjustment of the automation process according to experimental needs.

[0149] 4. Burning measurement module 5

[0150] The combustion measurement module 5 adopts the combustion method to volatilize the water and organic matter in the soil by heating and measuring the mass change, thereby efficiently determining the moisture content and organic matter content of the soil.

[0151] The burning measurement module 5 includes a high-precision scale 51, an insulating stone 52, a burning dish 53 and a flame head 54. The high-precision scale 51 is used to accurately weigh the quality of soil samples and other experimental materials, ensure the accuracy of the quality data at each experimental stage, and provide a reliable basis for subsequent density measurement and calculation. The insulating stone 52 is placed under the burning area to isolate the high temperature generated during burning and prevent the heat from adversely affecting the device below or the experimental environment. The burning dish 53 is used to hold experimental soil samples. During the burning process, the soil samples are heated to remove the moisture and organic matter therein, thereby weighing the weight of the moisture and organic matter. The flame head 54 is the core component of the burning device, which can accurately control the flame temperature and intensity to ensure uniform and efficient burning.

[0152] The working process of the burning measurement module 5 is as follows:

[0153] The soil sample to be processed is placed on a calcination dish 53 and calcined using a flame head 54. During the calcination process, a high-precision scale 51 monitors soil mass changes in real time. The detection interval is set to 1 second, based on experimental optimization results, to effectively capture subtle trends in mass change. When the mass change falls below a detection threshold (e.g., 0.01g) and persists for more than 30 seconds, the mass change is automatically determined to have ceased, and the calcination process is terminated to ensure the accuracy and consistency of the experimental operation.

[0154] During the testing process, a mass-time (MT) curve is automatically generated, recording the mass change trajectory of the soil sample during the combustion process. Analysis of experimental data shows that when the slope of the MT curve drops to 0.05 g / s, the water has essentially evaporated. At this point, the system automatically estimates the moisture content of the soil sample to be processed.

[0155] The estimated moisture content of the soil sample and the cone-shaped instrument measurement results are then processed using the liquid-plasticity experimental formula to obtain the estimated liquid limit and plastic limit. Based on the estimated liquid limit and plastic limit, the soil is initially classified. Once the soil classification is complete, the water volatilization time and organic matter volatilization time are determined based on the soil type and its moisture content characteristics. This step can be automated by the program.

[0156] In this embodiment, the results of the preliminary classification mainly include three categories: silt, silty clay, and clay. The water volatilization time of silt, silty clay, and clay is 30 seconds, 60 seconds, and 75 seconds, respectively. Experimental verification shows that within this time range, the moisture in the soil can be completely volatilized while avoiding the risk of changes in soil physical properties or premature decomposition of organic matter due to excessive heating. The volatilization time of organic matter is 105 seconds, 120 seconds, and 135 seconds, respectively. Experimental verification shows that within this time range, the relative error of the measurement results is less than 0.5%, and the accuracy meets the experimental requirements, providing a reliable basis for soil classification and analysis.

[0157] The highly integrated ignition measurement module 5 effectively reduces human error and significantly shortens experimental cycles through a fully automated process. Precise quality monitoring and time control ensure high accuracy of the ignition process and measurement results. Classification and calculation logic adapt to the experimental needs of different soil types. Automated data recording and analysis further reduce manual calculation errors, significantly improving experimental efficiency and data reliability.

[0158] 5. Density meter measurement module 4

[0159] The densitometer measurement module 4 includes a soil material nozzle 41 , a dispersant nozzle 42 , a pure water nozzle 43 , a second agitator 44 , a densitometer 45 , a thermometer 46 and a soil material container 47 .

[0160] The soil material nozzle 41, pure water nozzle 43, and dispersant nozzle 42 form a nozzle group to complete sample preparation and processing. The soil material nozzle 41 is used to add soil samples to the experimental system to ensure smooth entry of the soil sample into the system; the pure water nozzle 43 is used to inject a fixed amount of pure water into the sample to adjust the sample's moisture content to meet the requirements of the density measurement; the dispersant nozzle 42 is used to spray a 6% concentration of sodium hexametaphosphate solution (dispersant), which fully disperses the soil sample particles in the suspension to prevent particle aggregation and affect the accuracy of the density measurement.

[0161] The second stirrer 44 consists of a stirring rod with a wheel diameter of 50 mm, a hole diameter of about 3 mm and a length of about 400 mm and a rotating blade, and is used to fully mix the soil sample, purified water and dispersant to ensure uniform distribution of soil particles in the suspension.

[0162] Density meter 45 adopts Type A density meter, which is used to measure the density of soil samples in suspension. Its scale unit is expressed in grams of soil per 1000 mL of suspension at 20°C. The scale range is 5 to 50, and the graduation value is 0.5. It can accurately display the density changes of suspension in different time periods.

[0163] The thermometer 46 is used to monitor the temperature of the suspension during the experiment. The scale range is 0°C to 50°C, and the graduation value is 0.5°C. It ensures that the density measurement is carried out under standard temperature conditions and provides reliable experimental data.

[0164] The specific process of the density meter measurement module 4 performing particle analysis on the finely sieved soil sample using the density meter method is as follows:

[0165] Stage 1: Soil pretreatment

[0166] The moisture content (ω0) of the soil sample to be processed is determined by the burning method. Since the density meter method requires the dry mass of the sample (m d =30g) to calculate the mass of the soil sample to be processed (m0):

[0167] m0=m d (1+0.01ω0)

[0168] The second stage: suspension preparation stage

[0169] The finely sieved soil sample is added to a measuring cylinder, and the system adds about 10 mL of a 4% concentration of sodium hexametaphosphate solution to the measuring cylinder (soil container 47) through an automated device, and then injects pure water through a pure water nozzle 43 to make the total volume of the suspension in the measuring cylinder reach 1000 mL. Pure water provides a liquid phase environment for the suspension, and the dispersant (sodium hexametaphosphate solution) reduces the electrostatic attraction between the soil particles, prevents particle aggregation, and ensures the uniformity of the suspension. Subsequently, the second agitator 44 stirs back and forth up and down along the entire suspension depth in the measuring cylinder for about 1 minute, about 30 times per minute, through an automatic control system. The rotating blade design of the second agitator 44 forms a stable liquid circulation flow, ensuring that the soil particles in the suspension are evenly distributed. The agitator is equipped with a splash-proof structure to effectively prevent the liquid from splashing out of the measuring cylinder. At the same time, the stirring time and rotation speed can be set according to the experimental requirements to ensure that the processing quality of the suspension meets the requirements of density determination.

[0170] Phase 3: Determination

[0171] After the stirring is completed, the measuring cylinder is conveyed to the bottom of the Type A density meter. After it stabilizes, the system will automatically place the Type A density meter 45 gently into the suspension and start the stopwatch to record the reading. The density meter bubble always remains in the middle of the measuring cylinder to avoid being close to the cylinder wall or affected by liquid disturbance. According to the predetermined program, the system automatically completes the density meter reading at time points such as 0.5min, 1min, 2min, 5min, 15min, 30min, 60min, 120min, 180min and 1440min, and records the measurement data. The density meter reading must be based on the upper edge of the curved liquid surface, and the measurement accuracy must reach 0.5. After each reading is completed, the system will automatically take out the density meter and place it in a measuring cylinder filled with pure water for cleaning. At the same time, the system will measure the temperature of the suspension in the measuring cylinder (accurate to 0.5°C) and record the temperature data in the experimental results to correct the density changes caused by temperature fluctuations.

[0172] Phase 4: Data recording and analysis

[0173] After the experiment is completed, the system will combine the experimental formula to conduct a comprehensive analysis of the recorded density and temperature data. The experimental formula is as follows:

[0174]

[0175]

[0176] Where:

[0177] C s ——Soil particle specific gravity correction value; R1——Type A densitometer reading;

[0178] m T ——temperature correction value; n w ——Meniscus correction value;

[0179] C D ——dispersant correction value; ρ s ——Soil particle density (g / cm 3 );

[0180] ρ w20 ——Density of water at 20℃ (g / cm 3 ).

[0181] The particle size corresponding to different measurement times is obtained according to the following formula:

[0182]

[0183] Where:

[0184] d——particle size (mm);

[0185] η——dynamic viscosity of water (1×10 -6 kPa·s);

[0186] G wT ——Specific gravity of water at temperature T℃;

[0187] ρ w0 ——Density of water at 4℃ (g / cm 3 );

[0188] g——acceleration due to gravity;

[0189] L t ——the distance that soil particles settle within a certain time t (cm);

[0190] t——sedimentation time (s).

[0191] Finally, the mass distribution percentage of particles of different particle sizes is calculated, and the particle grading curve is drawn in combination with the first-level particle distribution results of the soil sample particle size grading and screening system to obtain the uniformity coefficient and curvature coefficient.

[0192] 5. Control Analysis Module 6

[0193] In this embodiment, after completing the collection and intelligent identification of key parameters such as the liquid and plastic limits, moisture content, particle size distribution, and organic matter content of the soil, the control and analysis module 6 can further combine the preset engineering adaptability classification method to obtain the soil type and match the differentiated solidification method suitable for the characteristics of this type of soil.

[0194] In this embodiment, typical soils are divided into five main types: conventional type, high liquid limit and high plasticity type, high organic matter type, gradation adjustable type and high water content type.

[0195] The classification conditions and corresponding solidification methods of various types of soil are shown in the following table:

[0196]

[0197] Among them, the parameter conditions of the conventional type are:

[0198] The optimal particle gradation is (0-5mm): (5-10mm): (10-25mm) = 40%: 20%: 40% or 50%: 40%: 10%;

[0199] Organic matter content <1%, ω L <50%, I P <26, moisture content is less than 25%.

[0200] 6. Control Analysis Module 6

[0201] (1) Experimental data collection and integration

[0202] The control and analysis module 6 receives the multi-dimensional data generated during the experiment and implements full-process intelligent control.

[0203] For example, during particle size classification and screening, the system automatically adjusts the screen inclination (e.g., 30° to 45°) and the vibration frequency (10Hz to 50Hz) based on the soil sample's particle size distribution and screening objectives. This program optimizes screening efficiency and accuracy in real time while minimizing the possibility of particle accumulation and screen clogging. High-precision weighing modules collect real-time soil sample quality data from each outlet and upload it to the system simultaneously.

[0204] During the combustion process, the system monitors mass changes every second through high-precision sensors, generates a mass-time (mt) curve, and automatically records the temperature and mass values ​​at key time points (such as when the slope drops to 0.05g / s or the mass change is less than 0.01g) based on slope change monitoring and weight stability judgment, providing support for the subsequent accurate calculation of moisture content and organic matter content.

[0205] In the liquid limit and plastic limit automatic measurement system, the system accurately calculates the liquid limit and plastic limit values ​​by monitoring the changes in soil paste quality, cone sinking depth and temperature values ​​at multiple time points.

[0206] It should be noted that all data are identified and eliminated for outliers through the system's built-in integration module, and are automatically normalized based on experimental conditions to ensure data accuracy and comparability between different experimental groups.

[0207] In addition, a data sharing interface is set up between modules to achieve collaborative processing of experimental data.

[0208] (2) Intelligent data preprocessing

[0209] During the experiment, the system monitors and processes data in real time through sensors and intelligent control modules to ensure the accuracy and efficiency of the experimental process.

[0210] For the first-level particle distribution results: the frequency and screening time of the vibration unit are adjusted through the system's intelligent control module, the soil sample quality data of each discharge port is collected in real time and a particle grading curve is generated. At the same time, the uniformity coefficient and curvature coefficient are calculated to evaluate the uniformity of the soil sample particle size distribution and the correlation between particle sizes, providing a basis for soil sample classification.

[0211] For moisture content and organic matter data, the system analyzes the mass-time (mt) curve generated by the ignition method, combined with intelligently controlled temperature recording and quality monitoring points (for example, when the slope drops to 0.05g / s, water evaporation is essentially complete), to accurately calculate the moisture content and organic matter content of the soil sample. For liquid and plastic limits estimation, the system automatically calculates the liquid limit and plastic limit values ​​by automatically collecting soil paste mass, cone sinking depth, and temperature values ​​at multiple time points, combined with the liquid and plastic calculation formula. The measurement accuracy reaches ±0.5%, fully meeting the needs of engineering applications and scientific research.

[0212] (3) Data analysis and dynamic visualization

[0213] The system features a variety of built-in data analysis tools, enabling efficient processing of experimental results and generating detailed analysis reports. Particle gradation analysis automatically generates a particle gradation curve based on real-time recorded screening mass and discharge data, and calculates key parameters such as the coefficient of uniformity and curvature, visually reflecting the particle size distribution characteristics of the soil sample. Mass-time curve analysis uses real-time collected soil mass and temperature data to calculate the soil sample's moisture content and organic matter content, and combines these analysis results to assess the impact of different experimental conditions.

[0214] In addition, the control and analysis module 6 can also support comparative analysis of different experimental groups, such as the impact of burning time or vibration screen frequency on soil sample type. After the experiment is completed, the system will dynamically generate mass-time curves, particle grading curves, liquid-plastic limit and moisture content distribution diagrams, etc., and support custom chart styles, so that the experimental results are more intuitive. The system automatically generates an experimental report, which includes experimental parameters (such as vibration frequency, screen inclination, burning time, etc.), soil sample type (such as particle grading classification, liquid-plastic limit) and analysis charts, providing a reliable basis for scientific research literature writing and engineering design decisions. For the recommended curing ratio, the system synchronously outputs the corresponding recommendation reasons and experimental basis (such as curing success rate, adaptability historical sample size, etc.), and can realize dynamic optimization of the curing ratio scheme based on subsequent user feedback, and gradually form an experience feedback closed-loop mechanism.

[0215] In summary, the system of the present invention not only enables rapid classification of geotechnical materials but also automatically matches differentiated curing strategies for special soil types, such as those with high liquid limit, high plasticity, and high organic matter content, based on the sensitivity of curing parameters. Through real-time data monitoring and intelligent feedback control, the system of the present invention dynamically optimizes experimental processes and curing ratios, improving soil classification accuracy and the ability to adapt to differentiated curing conditions. Ultimately, it achieves comprehensive intelligent, automated, and highly accurate identification of soil engineering properties and recommendation of improvement plans.

[0216] The above specific embodiments are used to illustrate the present invention rather than to limit the present invention. Any modifications and changes made to the present invention within the spirit of the present invention and the protection scope of the claims shall fall within the protection scope of the present invention.

[0217] The above description is only a preferred embodiment of the present invention. Therefore, any equivalent changes or modifications made according to the structure, characteristics and principles described in the scope of the patent application of the present invention are included in the scope of the patent application of the present invention.

Claims

1. A soil intelligent classification automation system, characterized in that: include: A soil sample pretreatment module (1) is used for air-drying, stirring and weighing the soil sample to be processed; The soil sample particle size classification and screening module (2) is used to perform step-by-step screening on the pre-treated soil sample to obtain a first-level particle distribution result, and obtain a coarse-screened soil sample and a fine-screened soil sample from the screened soil sample; The liquid-plastic limit automatic measurement module (3) is used to prepare the coarsely screened soil sample into soil pastes with different moisture contents, and measure them using the cone meter method to obtain the cone meter method measurement results; A burning measurement module (5) is used to burn the soil sample to be processed and obtain corresponding burning measurement results; A densitometer measurement module (4) is used to perform particle analysis on the finely sieved soil sample using a densitometer method to obtain a second-level particle distribution result; A control and analysis module (6) is used to realize the automatic control of the soil intelligent classification automation system, receive the first-level particle distribution results, the cone instrument method measurement results, the burning measurement results and the second-level particle distribution results, perform data analysis, and obtain the soil sample type and the corresponding optimal solidification method.

2. The soil intelligent classification automation system according to claim 1, characterized in that: When the control analysis module (6) performs data analysis, the following steps are performed: The estimated moisture content is obtained based on the burning measurement results corresponding to the pretreated soil samples; The estimated liquid and plastic limits are calculated based on the estimated moisture content and the cone instrument measurement results; Determine the water volatilization time and organic matter volatilization time based on the estimated liquid and plastic limits, and then obtain the moisture content and organic matter content of the soil sample to be treated; The liquid and plastic limits are calculated based on the moisture content of the soil sample to be treated and the results of cone oscillator method measurement; The particle grading parameters are obtained according to the first-level particle distribution results and the second-level particle distribution results; According to the particle gradation parameters, moisture content, organic matter content and liquid limit and plastic limit, the soil sample type is obtained; the soil sample types include conventional type, high liquid limit and high plasticity type, high organic matter type, gradation adjustable type and high moisture content type; Output soil sample type and corresponding optimal curing method.

3. The soil intelligent classification automation system according to claim 1, characterized in that: The soil sample pretreatment module (1) comprises an air-drying bin (11), a stirring bin (12) and a weighing bin (13); the soil sample to be treated is sequentially air-dried in the air-drying bin (11), stirred in the stirring bin (12) and weighed in the weighing bin (13); The air-drying chamber (11) is provided with a plurality of fans (14), the stirring chamber (12) is provided with a first stirrer (16) and a camera (19), and the weighing chamber (13) is provided with a first weighing device (17); the fans (14), the camera (19), the first stirrer (16) and the first weighing device (17) are all electrically connected and / or communicatively connected to the control module (7); The control and analysis module (6) is capable of controlling the stirrer (16) based on the real-time image captured by the camera (19), and processing the initial parameters of the soil sample to be processed using the internally stored air-drying control model, predicting the optimal air-drying parameters, and then controlling the fan (14) based on the optimal air-drying parameters; the initial parameters include initial humidity, particle distribution, and particle adsorption parameters.

4. The soil intelligent classification automation system according to claim 1, characterized in that: The soil sample particle size classification and screening module (2) comprises a screen (23), a sieve analyzer housing (24), a vibration unit and a soil sample conveyor belt (29); The sieve analyzer housing (24) is provided with a cavity as a sieve chamber, and a vibration unit is provided in the center of the sieve chamber to separate the sieve chamber into a coarse sieve chamber and a fine sieve chamber. A group of sieve components and a soil sample conveyor belt (29) are arranged above and below the coarse sieve chamber and the fine sieve chamber, respectively. The sieve components are mainly composed of a plurality of sieves (23) with different apertures. The sieves (23) are arranged in sequence from top to bottom with intervals and apertures decreasing in sequence. The vibration unit includes a shock-absorbing spring (26), a visual sensor (27), a vibration diaphragm and a vibration machine (28); the vibration machine (28) is installed on the bottom inner wall of the screening chamber, the output shaft of the vibration machine (28) is connected to the lower end of the vibration diaphragm, and the upper and lower ends of the vibration diaphragm are respectively connected to the top and bottom inner walls of the screening chamber through at least one shock-absorbing spring (26); A plurality of visual sensors (27) are installed on the vibration partition, and the number of the visual sensors (27) and the screen (23) is the same and corresponds one to one, and the visual sensors (27) are arranged above the corresponding screen (23); the visual sensors (27) and the vibration machine (28) are both electrically connected and / or communicatively connected to the control and analysis module (6); the control and analysis module (6) is capable of receiving the real-time image collected by the visual sensor (27), extracting the real-time screening state parameter from the real-time image, and processing it using the screening control model to obtain the predicted vibration state of the vibration partition, and controlling the vibration machine (28) according to the predicted vibration state of the vibration partition.

5. The soil intelligent classification automation system according to claim 4, characterized in that: The screening state parameters include the coverage area ratio of soil sample particles on the screen, the average particle size and the particle pass rate, and the particle pass rate is expressed as the relative value of the average particle size and the screen aperture; the predicted vibration state includes the predicted vibration amplitude, predicted frequency and predicted mode.

6. The soil intelligent classification automation system according to claim 5, characterized in that: The soil sample particle size classification and screening module (2) further comprises a coarse sieve soil inlet (21), a fine sieve soil inlet (22), a sieved soil collection box (25) and a controllable gate (210); the tops of the coarse sieve chamber and the fine sieve chamber are respectively provided with a coarse sieve soil inlet (21) and a fine sieve soil inlet (22); The inner sides of the sieve (23) and the soil sample conveyor belt (29) are connected to the vibrating partition, and the outer sides are connected to the inner wall of the sieving chamber. A controllable gate (210) is correspondingly arranged at the connection between the sieve (23) / soil sample conveyor belt (29) and the inner wall of the sieving chamber. A sieved soil collection box (25) is correspondingly arranged on the outer side of each controllable gate (210). The sieved soil collection box (25) is installed on the outer side wall of the sieve analyzer housing (24). A second weighing device is installed on the sieved soil collection box (25), and the second weighing device is communicatively connected to the control and analysis module (6).

7. The soil intelligent classification automation system according to claim 1, characterized in that: The liquid-plastic limit automatic measurement module (3) comprises a mixing drum, an aluminum box conveyor belt (31), a plurality of aluminum boxes (32), a scraping blade group (33) and a photoelectric combined measuring instrument (34); the mixing drum is used to add water to the finely screened soil sample step by step to obtain soil paste with different moisture contents; the aluminum box (32) is used to fill the soil paste; the sample loading end of the aluminum box conveyor belt (31) is used to place the aluminum box (32), and the scraping blade group (33) and the photoelectric combined measuring instrument (34) are sequentially arranged along the conveying direction at the discharge end; the photoelectric combined measuring instrument (34) is equipped with a plurality of cone meters, the number of cone meters being the same as the number of aluminum boxes (32) and corresponding to each other; the photoelectric combined measuring instrument (34) is equipped with a sensor; when the aluminum box (32) reaches a preset position, the sensor automatically senses and triggers the conveyor belt to decelerate to a standstill, so that the aluminum box (32) moves to the position directly below the corresponding cone meter.

8. The soil intelligent classification automation system according to claim 1, characterized in that: The burning measurement module (5) includes a high-precision scale (51), an insulating stone (52), a burning dish (53) and a flame head (54); the densitometer measurement module (4) includes a soil material nozzle (41), a pure water nozzle (43), a dispersant nozzle (42), a second stirrer (44), a densitometer (45), a thermometer (46) and a soil material container (47); the soil material container (47) is placed on a conveying device, and a bracket is provided above the conveying device, and the soil material nozzle (41), the pure water nozzle (43), the dispersant nozzle (42), the second stirrer (44), the densitometer (45) and the thermometer (46) are sequentially and telescopically connected to the bracket along the conveying direction.

9. A soil intelligent classification automation method, using the soil intelligent classification automation system according to any one of claims 1 to 8, characterized in that: The following steps are involved: S1) using the soil sample pretreatment module (1) to air-dry, stir and weigh the soil sample to be treated, thereby obtaining the pretreated soil sample and its total weight; S2) using the soil sample particle size classification and screening module (2) to perform step-by-step screening on the pre-treated soil sample, obtain the first-level particle distribution result and transmit it to the control and analysis module (6), take the last-level soil sample output by the coarse screening chamber as the coarse-screened soil sample, and take the last-level soil sample output by the fine screening chamber as the fine-screened soil sample; S3) The coarsely screened soil sample is prepared into soil pastes with different moisture contents by the liquid-plastic limit automated measurement module (3), and the soil pastes are measured by the cone oscillator method. The cone oscillator method measurement results are obtained and transmitted to the control and analysis module (6); S4) burning the soil sample to be processed using the burning measurement module (5) until the time rate of change of mass is less than a preset threshold value and lasts for a preset time period, obtaining a mass-time curve as a burning measurement result, and transmitting it to the control analysis module (6); S5) using the control analysis module (6) to analyze the mass-time curve obtained in step S4, and to calculate the estimated moisture content based on the total mass loss; The estimated liquid and plastic limits are calculated based on the estimated moisture content of each grade of soil paste prepared in step S3 and the results of cone oscillator measurement; Determine the water volatilization time and organic matter volatilization time based on the estimated liquid and plastic limits, and then obtain the moisture content and organic matter content of the soil sample to be treated; S6) calculating the moisture content of each level of soil paste in step S3 according to the moisture content of the soil sample to be processed, and combining the measurement results of the cone oscillator method to calculate the liquid limit and plastic limit of the soil sample to be processed; S7) using a density meter measurement module (4) to perform particle analysis on the finely screened soil sample using a density meter method to obtain a second-level particle distribution result, which is transmitted to a control and analysis module (6); the control and analysis module (6) obtains a particle grading parameter based on the first-level particle distribution result and the second-level particle distribution result; S8) using the control analysis module (6) to obtain the soil sample type according to the particle grading parameters, moisture content, organic matter content and liquid limit of the soil sample to be processed; S9) The control analysis module (6) outputs the soil sample type and the corresponding optimal solidification method.

10. The soil intelligent classification automation method according to claim 9, characterized in that: The method further includes realizing automatic control of the soil intelligent classification automation system through the control analysis module (6), specifically: In the step S1, before air-drying, the initial humidity, particle distribution and particle adsorption parameters of the soil sample to be processed are first input into the air-drying control model of the control analysis module (6), the air-drying control model predicts and outputs the optimal air-drying parameters, and the control analysis module (6) converts the optimal air-drying parameters into instructions and outputs them to the fan (14); during the stirring process, the control analysis module (6) processes the real-time image collected by the camera (19) through the image processing algorithm to obtain the uniformity parameter of the soil sample, and the stirring ends when the uniformity parameter of the soil sample reaches a preset threshold; after the weighing process, the first weigher (17) transmits the total weight of the pre-treated soil sample to the control analysis module (6); In step S2, the real-time image collected by the visual sensor (27) is input into the control analysis module (6), the real-time screening state parameters are extracted from the real-time image, and the screening control model is used for processing to obtain the predicted vibration state of the vibration partition. The control analysis module (6) controls the vibration machine (28) according to the predicted vibration state of the vibration partition.

Images