Nomadic flow-state solidified soil production process method

Abstract

The invention provides a nomadic flow-state solidified soil production process method, and belongs to the technical field of flow-state solidification, and the nomadic flow-state solidified soil production process method provided by the invention comprises the following steps: firstly, determining an optimal feeding ratio through a feeding parameter optimization function, and then adding water and feeding according to optimized parameters. In the stirring process, the stirring state is monitored in real time through the stirring working condition judgment function, and stirring parameters are dynamically adjusted through the rotating speed adjustment function. The flow and the pressure are accurately controlled by adopting a conveying control function in addition of the curing agent, and impurities are removed from the mixed slurry through a screening process optimized by a screening control function. Finally, stirring parameters are regulated and controlled through a secondary stirring control function, and the final product quality is ensured. An intelligent closed-loop control system is formed in the whole process, real-time and accurate regulation and control of the technological parameters are achieved, and the technical problem that in the prior art, the technological parameters in the nomadic type flow state solidified soil production process are difficult to regulate and control in real time and accurately is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of fluidized solidification, and in particular, relates to a nomadic fluidized solidified soil production process method. Background Art

[0002] Fluidized solidified soil technology is widely used in municipal engineering, foundation treatment, waste resource utilization and other fields. Its core lies in the formation of solidified soil with specific fluidity and strength through mixing. Traditional fluidized solidified soil production mainly relies on fixed production equipment, and uses empirical parameters for process control, including the setting of parameters such as water consumption per unit time, soil feed amount, and mixer speed. Although this fixed production mode has a relatively stable process, the equipment utilization rate is low and it is difficult to adapt to the needs of different construction sites. In order to improve equipment utilization and adaptability, nomadic fluidized solidified soil production technology came into being. This technology uses mobile equipment and can be flexibly arranged at different construction sites.

[0003] However, in the production process of nomadic fluidized solidified soil, due to the large fluctuations in the properties of raw materials and the complex and changeable process conditions, the traditional empirical parameter control method is difficult to meet the process requirements. Specifically, first, the feed parameters are difficult to optimize, and too large or too small feed amounts will affect product quality; second, the mixing process lacks real-time monitoring and evaluation methods, and it is impossible to accurately judge the mixing uniformity; third, the amount of curing agent added and the conveying parameters are difficult to accurately control, resulting in fluctuations in product performance; finally, the screening and secondary mixing processes lack intelligent control strategies, which affects the quality of the final product. The existing technology mainly uses manual experience adjustment or simple single-loop control methods, which cannot achieve coordinated optimization of process parameters.

[0004] The root of these problems lies in the lack of a systematic real-time optimization control method for process parameters in the nomadic fluidized solidification soil production process. The randomness of raw material properties, the complexity of the process, and the coupling relationship between the various processes make it difficult for traditional control methods to achieve precise control of process parameters, which directly affects the stability and consistency of product quality. Therefore, it is urgent to develop a production method that can achieve real-time and precise control of process parameters. Summary of the invention

[0005] In view of this, the present invention provides a nomadic fluidized solidified soil production process method, which can solve the technical problem in the prior art that the process parameters of the nomadic fluidized solidified soil production process are difficult to accurately control in real time.

[0006] The present invention is implemented as follows: The present invention provides a nomadic fluidized solidified soil production process method, which includes the following steps: Collect the rated power value of the slurry mixer and the volume value of the slurry mixer, measure the original soil density value, input the feeding parameter optimization function, and obtain the water consumption per unit time, the soil material feeding amount per unit time, and the initial rotation speed value of the slurry mixer; Pump water into the mixing tank of the slurry mixer according to the water consumption per unit time, and start the slurry mixer according to the initial rotation speed value of the slurry mixer; Convey the original soil into the slurry mixer by an excavator according to the soil material feeding amount per unit time; Collect the real-time power value of the slurry mixer, the real-time temperature value of the slurry mixer, and the real-time torque value of the slurry mixer, input the mixing condition determination function, and obtain the mixing uniformity value and the slurry viscosity change rate value; Input the mixing uniformity value and the slurry viscosity change rate value into the rotation speed adjustment function to adjust the operating parameters of the slurry mixer; Collect the rotation speed value of the curing agent pulping machine and the production value of the curing agent pulping machine, input the conveying control function, and obtain the slurry pump pressure value and the slurry pump flow value; Pump the curing agent slurry into the slurry mixer according to the slurry pump pressure value and the slurry pump flow value for mixing and stirring; Collect the feeding rate value of the slurry after mixing and stirring and the sieve residue content value, input the screening control function, and obtain the rotating screen speed value of the rotary screen for screening; Measure the fluidity value of the slurry after screening, collect the feeding rate value of the slurry after screening, input the secondary mixing control function, and obtain the rotation speed value of the secondary mixer and the secondary mixing time value; Perform secondary mixing according to the rotation speed value of the secondary mixer and the secondary mixing time value to obtain the fluidized solidified soil.

[0007] Among them, the feeding parameter optimization function is based on the principle of maximizing production efficiency, considering the equipment capacity limit and the material flow characteristics. The inputs include the rated power value of the slurry mixer, the volume value of the slurry mixer, and the original soil density value, and the outputs include the water consumption per unit time, the soil material feeding amount per unit time, and the initial rotation speed value of the slurry mixer; The feeding parameter optimization function is implemented based on the genetic algorithm, with a population size of 100, an iteration number of 500, a crossover probability of 0.8, and a mutation probability of 0.1.

[0008] Among them, the mixing condition determination function is based on the principle of slurry uniformity determination, considering the mixing energy consumption and the material state. The inputs include the real-time power value of the slurry mixer, the real-time temperature value of the slurry mixer, and the real-time torque value of the slurry mixer, and the outputs include the mixing uniformity value and the slurry viscosity change rate value; The mixing condition determination function is implemented based on the support vector machine algorithm, using the radial basis kernel function, with a penalty factor of 100 and a kernel function parameter of 0.1.

[0009] Among them, the transportation control function is based on the principle of pipeline transportation dynamics, taking into account the flow characteristics of the curing agent slurry. The input includes the speed value of the curing agent pulping machine and the output value of the curing agent pulping machine, and the output includes the mud pump pressure value and the mud pump flow value. The transportation control function solves the optimal transportation parameters based on the particle swarm optimization algorithm, with a population size of 50, a maximum number of iterations of 200, and a learning factor of 2.

[0010] Among them, the screening control function is based on the particle classification principle, taking into account the material dispersion and equipment screening efficiency. The input includes the slurry feed rate value and the screen content value, and the output includes the drum screen speed value. The screening control function is implemented based on the Kalman filter algorithm, and the system characteristics are determined through the state estimation matrix and the observation matrix.

[0011] Among them, the secondary mixing control function is based on the performance optimization principle of fluidized solidified soil, taking into account the final fluidity requirements of the slurry. The input includes the fluidity value of the slurry after screening and the feed rate value of the slurry after screening. The output includes the secondary mixer speed value and the secondary mixing time value. The secondary mixing control function is implemented based on the model predictive control algorithm, with a prediction time domain of 10 and a control time domain of 3.

[0012] Among them, the stirring uniformity value ranges from 0 to 1, and when it is greater than 0.85, it is considered that a uniform stirring state is achieved; when the slurry viscosity change rate value is less than 0.05 per second, it indicates that the slurry properties tend to be stable.

[0013] The adjustment range of the real-time speed value of the mud mixer is ±30% of the initial speed, and the speed change rate does not exceed 10 revolutions per minute per second; the mixing time value of the mud mixer is between 10 minutes and 30 minutes.

[0014] Among them, the speed range of the drum sieve is 20 to 60 rpm, and the speed adjustment accuracy is ±1 rpm; the speed range of the secondary mixer is 40 to 120 rpm, and the secondary mixing time is between 5 minutes and 15 minutes.

[0015] Among them, the pressure value of the mud pump is between 0.5 MPa and 2 MPa; the flow value of the mud pump ranges from 0.5 cubic meters per hour to 5 cubic meters per hour; and the mesh size of the screen is from 2 mm to 5 mm.

[0016] Compared with the prior art, the present invention provides a nomadic fluidized solidified soil production process method. The present invention proposes a nomadic fluidized solidified soil production process method based on multiple intelligent algorithms, which realizes real-time and precise control of process parameters by constructing a feed parameter optimization function, a mixing condition determination function, a speed adjustment function, a conveying control function, a screening control function, and a secondary mixing control function. The method fully considers the coupling relationship between each process, adopts a closed-loop control strategy, and ensures the stability of the production process.

[0017] The present invention solves the problems existing in traditional technologies: the optimal configuration of feeding parameters is achieved through genetic algorithms, avoiding the blindness of empirical parameter setting; a stirring condition judgment model is established using a support vector machine algorithm, realizing real-time monitoring of the stirring process; a speed regulation method based on a deep neural network improves stirring efficiency; a particle swarm optimization algorithm ensures the precise delivery of the curing agent; and Kalman filtering and model predictive control ensure the controllability of the screening and secondary stirring processes.

[0018] The method of the present invention can solve the problem that process parameters are difficult to accurately control in real time because a systematic intelligent control strategy is adopted. Each control function is based on a physical model, and parameter adjustment is achieved through an online optimization algorithm. An information feedback mechanism is formed between various control links, so that the entire production process forms an adaptive optimization system. At the same time, the selection of the control algorithm fully considers the characteristics of different processes, ensuring the reliability of the control effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0021] like Figure 1 FIG. 1 is a flow chart of a nomadic fluidized solidified soil production process provided by the present invention, and the method comprises the following steps:

[0022] S1, collecting the rated power value and volume value of the mud mixer, measuring the original soil density value, inputting the rated power value, volume value and original soil density value of the mud mixer into the feed parameter optimization function, and obtaining the water consumption per unit time, soil material feed per unit time and initial speed value of the mud mixer;

[0023] S2, pumping water into the mixing cylinder of the mud mixer according to the water consumption per unit time, and starting the mud mixer according to the initial speed value of the mud mixer;

[0024] S3, conveying the original soil into the mud mixer through the excavator according to the soil material feed rate per unit time;

[0025] S4. During the operation of the mud mixer, real-time power value, real-time temperature value and real-time torque value of the mud mixer are collected, and the real-time power value, real-time temperature value and real-time torque value of the mud mixer are input into a mixing condition determination function to obtain a mixing uniformity value and a slurry viscosity change rate value.

[0026] S5, inputting the stirring uniformity value and the slurry viscosity change rate value into the speed adjustment function, obtaining the real-time speed value and the stirring time value of the mud mixer, and adjusting the operating parameters of the mud mixer according to the real-time speed value and the stirring time value of the mud mixer;

[0027] S6, collecting the speed value and output value of the curing agent pulping machine, inputting the speed value and output value of the curing agent pulping machine into the conveying control function, and obtaining the pressure value and flow value of the mud pump;

[0028] S7, pumping the curing agent slurry into the mud mixer for mixing and stirring according to the mud pump pressure value and the mud pump flow value;

[0029] S8, collecting the slurry feed rate value and the oversize content value after mixing and stirring, inputting the slurry feed rate value and the oversize content value into the screening control function, and obtaining the drum rotating screen speed value;

[0030] S9, starting the drum sieve according to the drum sieve speed value to screen the mixed and stirred slurry;

[0031] S10, measuring the fluidity value of the slurry after screening, collecting the feed rate value of the slurry after screening, inputting the fluidity value of the slurry after screening and the feed rate value of the slurry after screening into the secondary stirring control function, and obtaining the speed value of the secondary mixer and the secondary stirring time value;

[0032] S11, starting the secondary mixer according to the secondary mixer speed value and the secondary mixing time value to obtain fluidized solidified soil;

[0033] The feed parameter optimization function is based on the principle of maximizing production efficiency, taking into account equipment capacity limitations and material flow characteristics. The input includes the mud mixer rated power value, mud mixer volume value, and original soil density value. The output includes the water consumption per unit time, the soil material feed per unit time, and the initial speed value of the mud mixer.

[0034] The stirring condition determination function is based on the slurry uniformity determination principle, taking into account the stirring energy consumption and material state, and the input includes the real-time power value of the mud mixer, the real-time temperature value of the mud mixer, and the real-time torque value of the mud mixer, and the output includes the stirring uniformity value and the slurry viscosity change rate value;

[0035] The speed adjustment function is based on the principle of slurry rheological properties, taking into account the shear stress distribution inside the slurry. The input includes the stirring uniformity value and the slurry viscosity change rate value, and the output includes the real-time speed value of the mud mixer and the stirring time value of the mud mixer.

[0036] The transport control function is based on the pipeline transport dynamics principle, taking into account the flow characteristics of the curing agent slurry, and the input includes the curing agent slurry machine speed value and the curing agent slurry machine output value, and the output includes the mud pump pressure value and the mud pump flow value;

[0037] The screening control function is based on the particle classification principle, taking into account the material dispersion and equipment screening efficiency. The input includes the slurry feed rate value and the screen content value, and the output includes the drum screen speed value.

[0038] The secondary mixing control function is based on the performance optimization principle of fluidized solidified soil and takes into account the final fluidity requirements of the slurry. The input includes the fluidity value of the slurry after screening and the feed rate value of the slurry after screening. The output includes the secondary mixer speed value and the secondary mixing time value.

[0039] The specific implementation methods of the above steps are described in detail below. The specific implementation methods of step S1 are as follows: first, the rated power value of the mud mixer is obtained by using a power detection module, which is generally between 15 kW and 45 kW; then the volume value of the mud mixer is obtained by a volume measuring device, and the conventional volume value is 2 cubic meters to 6 cubic meters; then the density value of the original soil is measured using a density meter, which is generally between 1.6 grams per cubic centimeter and 2.2 grams per cubic centimeter. The feed parameter optimization function is based on a genetic algorithm, which iteratively optimizes the feed parameters by simulating the natural evolution process. Set the population size to 100, the number of iterations to 500, the crossover probability to 0.8, and the mutation probability to 0.1. The optimization goal is to maximize production efficiency while meeting equipment capacity constraints and material fluidity requirements. The function outputs a water consumption per unit time of 0.8 cubic meters per hour to 2.4 cubic meters per hour, a soil material feed per unit time of 1.2 tons per hour to 3.6 tons per hour, and an initial speed value of the mud mixer of 60 to 180 revolutions per minute. The main purpose of this step is to determine the optimal feed parameter combination to provide basic parameter support for the subsequent mixing process.

[0040] The specific implementation method of step S2 is as follows: according to the water consumption per unit time obtained in step S1, a variable frequency water pump is used to deliver clean water to the mixing cylinder of the mud mixer. The working frequency of the water pump is automatically adjusted according to the required flow rate, generally between 20 Hz and 50 Hz; the rated head of the water pump is not less than 20 meters of water column to ensure that the water delivery is stable and controllable; a liquid level sensor is set in the mixing cylinder, and the water supply is automatically stopped when the water level reaches the set height; then the mixing motor is started according to the initial speed value of the mud mixer obtained, and a paddle is installed on the mixing shaft, and the paddle inclination angle is 30 degrees to 45 degrees to improve the mixing efficiency. The purpose of this step is to provide a suitable liquid phase environment for the subsequent addition of soil materials and establish initial flow field conditions.

[0041] The specific implementation of step S3 is as follows: an excavator with a weighing sensor is used to transport the original soil into the mud mixer according to the soil feed rate per unit time determined in step S1, and the weighing sensor has an accuracy of ±0.5 kg; the excavator bucket capacity matches the feed rate, which is generally 0.5 cubic meters to 1 cubic meter; a fuzzy control algorithm is used in the feeding process to achieve accurate adjustment of the feed rate, and the fuzzy rule is established based on the feed rate deviation and the deviation change rate, and the output is the excavator action control amount; the feed rate deviation threshold is set to ±5%, and an alarm signal is triggered when it exceeds this range. The purpose of this step is to achieve quantitative addition of the original soil and provide a stable material basis for the subsequent mixing process.

[0042] The specific implementation method of step S4 is as follows: during the mud mixing process, the power detection module is used to collect the power value of the mixer in real time, and the sampling frequency is 10 Hz; the temperature value of the mixer is collected by the temperature sensor, and the temperature measurement range is 0 degrees Celsius to 100 degrees Celsius, with an accuracy of ±0.5 degrees Celsius; the torque value of the mixer is collected by the torque sensor, with a range of 0 to 1000 Newton meters and an accuracy of ±0.5% full scale. The mixing condition judgment function is implemented based on the support vector machine algorithm, using the radial basis kernel function, the penalty factor is 100, and the kernel function parameter is 0.1. Through real-time analysis of the operating parameters of the mixer, the mixing uniformity value is output, with a value range of 0 to 1. When it is greater than 0.85, it is considered to have reached a uniform mixing state; at the same time, the slurry viscosity change rate value is output to characterize the rheological properties of the slurry. When the value is less than 0.05 per second, it indicates that the slurry properties tend to be stable. The purpose of this step is to achieve real-time monitoring and evaluation of the mixing process.

[0043] The specific implementation method of step S5 is as follows: Based on the stirring uniformity value and the slurry viscosity change rate value obtained in step S4, a speed adjustment function is constructed using a deep neural network algorithm. The network structure is three-layer, the number of hidden layer neurons is 20, and the activation function uses a hyperbolic tangent function. By learning from historical operating data, a mapping relationship between stirring parameters and slurry performance is established; the function outputs the real-time speed value of the mud mixer, the adjustment range is ±30% of the initial speed, and the speed change rate does not exceed 10 revolutions per minute per second; at the same time, the mud mixer stirring time value is output, which is generally between 10 minutes and 30 minutes. The mixer control system adjusts the operating status in real time according to these parameters to ensure optimal control of the stirring process. The purpose of this step is to achieve dynamic optimization of stirring parameters and improve stirring efficiency and slurry quality.

[0044] The specific implementation method of step S6 is as follows: the speed sensor is used to collect the speed value of the curing agent pulping machine, the measurement range is 0 to 300 revolutions per minute, and the accuracy is ±1 revolution per minute; the output value of the curing agent pulping machine is measured by a flow meter, and the range is 0 to 10 cubic meters per hour. The transportation control function is constructed based on the pipeline transportation dynamics model. Considering the rheological properties of the slurry and the pipeline transportation loss, the particle swarm optimization algorithm is used to solve the optimal transportation parameters. The population size is 50, the maximum number of iterations is 200, and the learning factor is 2. The function outputs the mud pump pressure value, which is generally between 0.5 MPa and 2 MPa; at the same time, the mud pump flow value is output, and the value range is 0.5 cubic meters per hour to 5 cubic meters per hour. The purpose of this step is to ensure the stable transportation of the curing agent slurry and provide guarantee for subsequent mixing and stirring.

[0045] The specific implementation method of step S7 is as follows: according to the mud pump pressure value and flow value determined in step S6, the mud pump is started to transport the curing agent slurry to the mud mixer; the conveying pipeline is made of wear-resistant material, and the pipe diameter matches the flow rate, generally between 50 mm and 150 mm; the pipeline system is equipped with a pressure sensor and a flow sensor to monitor the conveying parameters in real time; when the pressure exceeds 120% of the set value, the pressure relief valve is automatically opened; when the flow deviation exceeds ±10% of the set value, an alarm signal is triggered. The mixing and stirring process adopts an adaptive control strategy to adjust the stirring parameters in real time according to the slurry properties. The purpose of this step is to achieve accurate addition and uniform mixing of the curing agent.

[0046] The specific implementation of step S8 is as follows: a speed sensor is used to measure the feed rate value of the slurry after mixing and stirring, and the measurement range is 0 to 10 cubic meters per hour; the content value of the screen material is determined by an online detection device, and the detection accuracy is ±0.5%. The screening control function is implemented based on the Kalman filter algorithm, which can effectively filter out the measurement noise and improve the control accuracy. The state estimation matrix and the observation matrix are determined according to the system characteristics, and the process noise covariance and the observation noise covariance are calibrated by experimental data. The function outputs the drum rotating screen speed value, which ranges from 20 rpm to 60 rpm, and the speed adjustment accuracy is ±1 rpm. The purpose of this step is to achieve efficient screening of the slurry and remove impurities and agglomerates.

[0047] The specific implementation of step S9 is as follows: according to the drum screen speed value obtained in step S8, the drum screen is started to screen the mixed and stirred slurry; the screen is made of stainless steel, and the mesh size is 2 mm to 5 mm; high-frequency vibration is used to assist in the screening process, and the vibration frequency is 30 Hz to 50 Hz; an automatic cleaning device is set, and the cleaning program is started when the screen blockage rate exceeds 30%; the screening efficiency is not less than 90%, and the content of particles with a particle size larger than the mesh size in the screened material does not exceed 1%. The purpose of this step is to ensure the uniformity of the particle size of the slurry and improve the quality of the final product.

[0048] The specific implementation method of step S10 is as follows: the fluidity value of the slurry after screening is measured by a fluidity meter, and the measuring range is 160 mm to 260 mm; the feed rate value of the slurry after screening is measured by a flow meter, and the accuracy is ±2%. The secondary stirring control function is implemented based on the model predictive control algorithm, with a prediction time domain of 10, a control time domain of 3, and a control cycle of 1 second. The model uses a state space form to describe the dynamic characteristics of the system, and updates the model parameters through online identification; the optimization objectives include fluidity control accuracy and minimization of energy consumption, and quadratic programming is used to solve the optimal control sequence. The function outputs the secondary mixer speed value, ranging from 40 rpm to 120 rpm; at the same time, it outputs the secondary stirring time value, which is generally between 5 minutes and 15 minutes. The purpose of this step is to optimize the final performance indicators of the slurry.

[0049] The specific implementation of step S11 is as follows: according to the secondary mixer speed value and mixing time value obtained in step S10, the secondary mixer is started for final modulation; the mixing barrel is designed with a special shape and is equipped with guide ribs inside to enhance the mixing effect; a multi-layer mixing paddle is installed on the mixing shaft, and the paddle shape and installation angle are optimized; during the mixing process, online monitoring is used to ensure that the fluidity of the slurry is maintained within the set range, and the allowable fluctuation range is ±10 mm; when the mixing time reaches the set value and the slurry performance index meets the requirements, the mixing is automatically stopped and the qualified product is output. The purpose of this step is to obtain fluidized solidified soil that meets the requirements and ensure the stability and consistency of product quality.

[0050] The functions or models involved in the present invention are described in detail below.

[0051] 1. The mathematical expression of the feed parameter optimization function is as follows:

[0052] Q water =k1P rated V m ρ s +α 1 ;

[0053] Q soil =k 2 P rated V m ρ s +α 2 ;

[0054] N init =k 3 P rated V m ρ s +α 3 ;

[0055] In the formula, Q water is the water consumption per unit time, in cubic meters per hour; Q soil is the soil material feed rate per unit time, in tons per hour; N init is the initial speed of the mud mixer, in revolutions per minute; P rated V is the rated power of the mud mixer in kilowatts; m is the volume of the mud mixer, in cubic meters; ρ s is the original soil density in grams per cubic centimeter; k 1 , k 2 , k 3 is the proportionality coefficient; α 1 , α 2 , α 3 is the correction factor.

[0056] Among them, the parameter acquisition method is: rated The power is directly obtained through the power detection module, with a detection accuracy of ±0.1 kilowatt; V m Obtained through a volume measurement device with a measurement accuracy of ±0.05 cubic meters; ρ s The density is measured with a density meter with a measurement accuracy of ±0.01 g / cm3; k 1 , k 2 , k 3 Obtained through genetic algorithm optimization, the fitness function is: F = ω 1 Q water +ω 2Q soil +ω 3 N init ; In the formula, ω 1 ,ω 2 ,ω 3 is the weight coefficient, which is determined by expert experience.

[0057] 2. The mathematical expression of the mixing condition judgment function is as follows:

[0058]

[0059] Where, H is the stirring uniformity value, dimensionless; R v is the slurry viscosity change rate, in units of second; P real is the real-time power value of the mud mixer, in kilowatts; T real is the real-time temperature value of the mud mixer, in degrees Celsius; M real is the real-time torque value of the mud mixer, in Nm; T max is the maximum allowable temperature, which is 100 degrees Celsius; M rated is the rated torque value, in Nm; β 1 , β 2 is the error correction term; f, g are the support vector machine mapping functions.

[0060] Parameter acquisition method: P real Real-time data acquisition through power detection module, sampling frequency 10 Hz; T real The temperature is collected by the temperature sensor with a sampling frequency of 10 Hz; M real The torque sensor is used to collect the data with a sampling frequency of 10 Hz. The support vector machine mapping function uses the radial basis kernel function: K(x i , x j ) = exp(-γ||x i -x j || 2 ), where γ is the kernel function parameter and its value is 0.1.

[0061] 3. The mathematical expression of the speed regulation function is as follows:

[0062] N real =φ 1 (H, R v )+λ 1 ;

[0063] T mix =φ 2 (H, R v )+λ 2 ;

[0064] Where N realis the real-time speed value of the mud mixer, in revolutions per minute; T mix is the mixing time of the mud mixer, in minutes; φ 1 ,φ 2 is the deep neural network mapping function; 1 ,λ 2 is the correction factor.

[0065] The mathematical expression of a deep neural network is:

[0066]

[0067] In the formula, is the output of the i-th neuron in the l-th layer; is the connection weight; is the bias term; σ is the activation function, using the hyperbolic tangent function:

[0068] 4. The mathematical expression of the conveying control function is as follows:

[0069] P pump =ψ 1 (N solid , Q solid )+μ 1 ;

[0070] Q pump =ψ 2 (N solid , Q solid )+μ 2 ;

[0071] Where P pump is the mud pump pressure value, in MPa; Q pump is the mud pump flow value, in cubic meters per hour; N solid Q is the speed of the curing agent pulping machine, in revolutions per minute; solid is the output value of the curing agent pulping machine, in cubic meters per hour; ψ 1 , ψ 2 is the function solved by the particle swarm optimization algorithm; μ 1 , μ 2 is the error term.

[0072] The speed and position update equations of the particle swarm optimization algorithm are:

[0073]

[0074] In the formula, is the velocity of the ith particle in the dth dimension at the kth iteration; is the position; ω is the inertia weight; c 1 , c2 is the acceleration constant; r 1 , r 2 is a random number; is the optimal position of an individual; is the global optimal position.

[0075] 5. The mathematical expression of the screening control function is as follows:

[0076] N screen =θ(Q feed , C over )+δ;

[0077] Where N screen is the rotating speed of the drum screen, in revolutions per minute; Q feed is the slurry feed rate value, in cubic meters per hour; C over is the content of the material on the sieve, dimensionless; θ is the mapping function of the Kalman filter algorithm; δ is the noise term.

[0078] The state equation and observation equation of the Kalman filter algorithm are:

[0079] x k =Ax k-1 +Bu k +w k ;

[0080] z k =Hx k +v k ;

[0081] In the formula, x k is the state vector; u k is the control vector; z k is the observation vector; A is the state transfer matrix; B is the control matrix; H is the observation matrix; w k , v k are process noise and observation noise.

[0082] 6. The mathematical expression of the secondary stirring control function is as follows:

[0083] N second =ξ 1 (F s , Q s )+η 1 ;

[0084] T second =ξ 2 (F s , Q s )+η 2 ;

[0085] Where N secondis the speed of the secondary mixer, in revolutions per minute; T second is the secondary stirring time value, in minutes; F s Q is the fluidity value of the slurry after screening, in millimeters; s is the feed rate of the slurry after screening, in cubic meters per hour; 1 ,ξ 2 is the model predictive control algorithm mapping function; η 1 , η 2 is the error correction term.

[0086] The optimization objective function of model predictive control is:

[0087]

[0088] Where N p is the prediction time domain; N c To control the time domain; k+i|k is the output prediction value; r k+i is the reference trajectory; Δu k+i|k is the control increment; Q, R are the weight matrices.

[0089] The principle and significance of constructing these equations are: all functions take into account the physical constraints and process requirements in actual engineering, and improve the robustness and adaptability of the model by introducing multiple correction coefficients and error terms; different optimization algorithms are used to deal with the characteristics of different processes, such as genetic algorithms are suitable for multi-objective optimization problems, support vector machines are suitable for nonlinear classification problems, deep neural networks are suitable for complex mapping relationships, particle swarm algorithms are suitable for continuous optimization problems, Kalman filters are suitable for noisy state estimation, and model predictive control is suitable for multivariable constrained control problems. Each control function includes a feedback correction mechanism that can adjust parameters according to real-time monitoring data to ensure system stability and control accuracy.

[0090] The derivation and establishment process of each equation is explained in detail below.

[0091] 1. Derivation process of feed parameter optimization function:

[0092] 1. Based on the principles of mass conservation and energy conservation, the power balance equation of the mud mixer is first established:

[0093] P rated =k p N init (ρ s Q soil +ρ w Q water )V m ;

[0094] In the formula, kp is the power coefficient, which is related to the shape and installation method of the stirring paddle and is obtained through experimental calibration; ρ w is the density of water.

[0095] 2. Consider the optimal slurry concentration requirements of the material and establish the water-soil ratio constraint equation:

[0096]

[0097] Where γ is the optimal water-soil ratio, ranging from 0.4 to 0.6, which is determined by orthogonal experiments.

[0098] 3. Combine the above equations and consider the equipment capacity constraints to obtain the final feed parameter optimization function. The optimization process uses a genetic algorithm, and the genetic coding includes Q water , Q soil 、N init Three variables, the fitness function is:

[0099]

[0100] In the formula, Q water，max , Q soil，max 、N init，max It is the maximum allowable value of each parameter, which is determined by the equipment specifications.

[0101] 2. Derivation process of mixing condition judgment function:

[0102] 1. Based on the relationship between stirring power and Reynolds number, the initial model of stirring uniformity is established:

[0103]

[0104] 2. Considering the influence of temperature on slurry viscosity, the Arrhenius equation is introduced to correct it:

[0105]

[0106] Where η is the slurry viscosity; η 0 is the base viscosity; E a is the apparent activation energy; R is the gas constant.

[0107] 3. Use support vector machine to establish nonlinear mapping relationship. The training samples are obtained through online monitoring data during the mixing process. The kernel function selects radial basis function:

[0108] K(x i , x j ) = exp(-γ||x i -x j || 2 );

[0109] The optimization goal of the support vector machine is:

[0110]

[0111] s.t.y i (w T x i +b)≥1-ξ i ,ξ i ≥0;

[0112] In the formula, w is the weight vector; b is the bias term; C is the penalty factor; ξ i is the slack variable.

[0113] 3. Speed ​​regulation function derivation process:

[0114] 1. Establish the relationship equation between stirring power and speed:

[0115]

[0116] Where D is the diameter of the stirring paddle; ρ mix is the density of the mixture.

[0117] 2. Consider the influence of slurry viscosity change and introduce correction items:

[0118]

[0119] In the formula, k η is the viscosity correction factor.

[0120] 3. A three-layer feedforward neural network is used to establish the mapping relationship, and the hidden layer uses the hyperbolic tangent activation function:

[0121]

[0122] The network training uses the back propagation algorithm, and the loss function is the mean square error:

[0123]

[0124] In the formula, m is the number of samples; y is i is the actual value; is the predicted value.

[0125] 4. Derivation process of conveying control function:

[0126] 1. Based on pipeline transportation dynamics, establish the pressure loss equation:

[0127]

[0128] Where f is the pipeline friction coefficient; L is the pipeline length; D is the pipeline diameter; ρ is the slurry density; v is the flow velocity; and K is the local loss coefficient.

[0129] 2. Considering the non-Newtonian fluid characteristics of the curing agent slurry, the correction term is introduced:

[0130]

[0131] Where τ is the shear stress; τ 0 is the yield stress; K is the consistency coefficient; is the shear rate; n is the rheological index.

[0132] 3. The particle swarm optimization algorithm is used to solve the optimal transportation parameters. The objective function is:

[0133]

[0134] Where P max is the maximum pressure of the mud pump; Q max It is the maximum flow rate of mud pump.

[0135] 5. Derivation process of screening control function:

[0136] 1. Establish screening efficiency model:

[0137]

[0138] In the formula, E is the screening efficiency; k is the screening coefficient; A is the effective area of ​​the screen.

[0139] 2. Consider the influence of screen blockage and introduce time-varying coefficient:

[0140] k=k 0 (1-βt);

[0141] In the formula, k 0 is the initial screening coefficient; β is the blocking coefficient; t is the running time.

[0142] 3. Use Kalman filter algorithm to realize state estimation, and the state space model is:

[0143]

[0144] z k =[1 0]x k +v k ;

[0145] Prediction equation:

[0146]

[0147] P k|k-1 =APk-1|k-1 A T +Q;

[0148] Update equation:

[0149] K k =P k|k-i H T (HP k|k-1 H T +R) -1 ;

[0150]

[0151] P k|k =(IK k H)P k|k-1 ;

[0152] Where P is the estimation error covariance matrix; K is the Kalman gain; Q is the process noise covariance; and R is the observation noise covariance.

[0153] 6. Derivation process of secondary stirring control function:

[0154] 1. Establish a slurry fluidity prediction model:

[0155] F k+1 =aF k +bN second +cQ s +d;

[0156] Where a, b, c, and d are model parameters, which are identified by the least squares method.

[0157] 2. Consider the effect of temperature on fluidity and introduce a temperature correction term:

[0158] F T =F k+1 [1+α(TT 0 )];

[0159] Where α is the temperature correction coefficient; T 0 is the base temperature.

[0160] 3. Using the model predictive control algorithm, the prediction model is:

[0161]

[0162] Constraints:

[0163] u min ≤u k ≤u max ;

[0164] Δu min ≤Δuk ≤Δu max ;

[0165] y min ≤y k ≤y max ;

[0166] Solve the quadratic programming problem:

[0167]

[0168] The effects of these equations are reflected in: the feed parameter optimization function achieves a balance between production efficiency and slurry performance; the mixing condition determination function can accurately evaluate the degree of mixing uniformity; the speed adjustment function ensures the dynamic optimization of the mixing process; the conveying control function ensures the accurate addition of the curing agent; the screening control function improves the screening efficiency and classification accuracy; the secondary mixing control function achieves accurate control of the final product performance. All parameters can be obtained through online detection or experimental calibration, and a closed-loop control system is formed between the equations to ensure the stability of the production process and the consistency of product quality. The entire control system is adaptive and robust, and can cope with the impact of fluctuations in raw material properties and changes in process conditions.

[0169] Specifically, the principle of the present invention is: the technical principle of the present invention is a method based on the combination of physical process modeling and intelligent algorithm optimization. First, by analyzing the relationship between mass conservation, momentum conservation and energy conservation in the production process of fluidized solidified soil, the basic physical model of each process is established. For example, the relationship between power and speed in the mixing process, the pressure loss equation in the transportation process, the efficiency model in the screening process, etc. These models reflect the inherent connection between process parameters.

[0170] Secondly, considering the nonlinear characteristics of the process and the coupling relationship between parameters, different types of intelligent algorithms are selected to construct the control function. Genetic algorithm is suitable for feed parameter optimization because it can find the global optimal solution in a complex parameter space; support vector machine is suitable for mixing condition determination because it has good classification ability and generalization performance; deep neural network is used for speed regulation and can learn complex nonlinear mapping relationships; particle swarm algorithm is suitable for conveying control and can quickly converge to the optimal solution; Kalman filter is suitable for screening control and can effectively handle measurement noise; model predictive control is suitable for secondary mixing and can consider multivariable constraints.

[0171] Finally, by establishing an information interaction mechanism between the control functions, a multi-level closed-loop control system is formed. The control results of the upstream process are used as the input parameters of the downstream process, and the operating status of the downstream process is fed back to the upstream process for parameter adjustment. This system structure ensures the coordinated optimization of process parameters and can adapt to changes in raw material properties and process conditions.

[0172] A specific embodiment 1 of the present invention is provided below, and the specific implementation method of each step in this embodiment 1 is described in detail as follows.

[0173] The specific implementation method of step S1 is: use a power detection module to obtain the rated power value of the mud mixer, which is generally between 15 kilowatts and 45 kilowatts. The power detection module uses a temperature-compensated Hall sensor with a measurement accuracy of 0.1% of the full scale. Use a volume measurement device constructed with a laser rangefinder to obtain the volume value of the mud mixer. The conventional volume value is 2 cubic meters to 6 cubic meters, and the measurement accuracy is ±0.05 cubic meters. The density value of the original soil is measured by a densimeter based on the principle of ray attenuation, which is usually between 1.6 grams per cubic centimeter and 2.2 grams per cubic centimeter, and the measurement accuracy is ±0.01 grams per cubic centimeter. Based on the principles of conservation of mass and conservation of energy, a feed parameter optimization function is constructed, and the expression is: P rated =k p N init (ρ s Q soil +ρ w Q water )V m , where k p is the power coefficient, which is related to the shape and installation method of the stirring paddle and is obtained through experimental calibration. Its value range is 0.8 to 1.2; ρ w is the density of water. Considering the optimal slurry concentration requirement of the material, the water-soil ratio constraint equation is established: In the formula, γ is the optimal water-soil ratio, ranging from 0.4 to 0.6. Genetic algorithm is used for parameter optimization, and the chromosome encoding includes Q water , Q soil 、N init Three variables, population size is 100, number of iterations is 500, crossover probability is 0.8, mutation probability is 0.1, and the fitness function is: In the formula, Q water，max , Q soil，max 、N init，max is the maximum allowable value of each parameter. This step determines the optimal feed parameter combination by combining theoretical modeling and optimization algorithm, providing basic parameter support for the subsequent mixing process.

[0174] The specific implementation method of step S2 is: according to the water consumption per unit time obtained in step S1, a variable frequency water pump is used to deliver clean water to the mixing cylinder of the mud mixer. The working frequency of the water pump is automatically adjusted according to the required flow rate, and the frequency adjustment range is 20 Hz to 50 Hz. A piezoelectric pressure sensor is used to measure the outlet pressure of the water pump, and an alarm signal is triggered when the pressure exceeds 2.0 MPa. The rated head of the water pump is not less than 20 meters of water column to ensure that the water delivery is stable and controllable. A capacitive liquid level sensor is set in the mixing cylinder with a measurement accuracy of ±1 mm. When the water level reaches the set height, the water supply is automatically stopped. When the mixing system is started, a soft start mode is adopted, and the start time is 10 seconds to 30 seconds. A paddle is installed on the mixing shaft, and the paddle inclination angle is 30 degrees to 45 degrees. This step realizes the precise control of the water volume and the smooth start of the mixing system.

[0175] The specific implementation of step S3 is: an excavator using a strain gauge weighing sensor with a weighing accuracy of ±0.5 kg is used to deliver raw soil to the mud mixer according to the soil feed rate per unit time determined in step S1. The feeding process uses a fuzzy control algorithm to achieve accurate adjustment of the feeding rate, and establishes a fuzzy set of two input variables, soil feed rate deviation E and deviation change rate EC, with domains of [-6, 6] and [-3, 3] respectively. The output variable is the excavator action adjustment amount U, with a domain of [-2, 2]. The fuzzy rule adopts a double-input single-output form: If E is A i and EC is B j then U is C k , where A i , B j , C k They are the fuzzy subsets on the corresponding domain. The centroid method is used for defuzzification: Where μ(x) is the membership function. The feed amount deviation threshold is set to ±5%, and an alarm signal is triggered when it exceeds this range. The dynamic response time of the feeding process does not exceed 3 seconds, and the adjustment accuracy is better than 1%. This step realizes the quantitative addition of raw soil and the intelligent control of the feeding process.

[0176] The specific implementation method of step S4 is: during the mud mixing process, a Hall effect power detection module is used to collect the power value of the mixer in real time, with a sampling frequency of 10 Hz and a measurement accuracy of ±0.1% of the full scale. A platinum thermal resistor temperature sensor is used to collect the temperature value of the mixer, with a temperature measurement range of 0 degrees Celsius to 100 degrees Celsius, an accuracy of ±0.5 degrees Celsius, and a response time of less than 1 second. The torque value of the mixer is collected by a magnetoelectric torque sensor, with a range of 0 to 1000 Nm, an accuracy of ±0.5% of the full scale, and the sampling frequency is synchronized with the power detection. The mixing condition judgment function is implemented based on the support vector machine algorithm. First, an initial model of mixing uniformity is established: Considering the influence of temperature on the viscosity of the slurry, the Arrhenius equation is introduced: In the formula, η is the viscosity of the slurry, with the unit of Pa·s; η 0 is the reference viscosity, with the unit of Pa·s; E a is the apparent activation energy, with the unit of J / mol; R is the gas constant, with a value of 8.314 J / (mol·K); T real is the Kelvin temperature. The radial basis kernel function is used to construct the mapping relationship of the support vector machine: K(x i , x j ) = exp(-γ||x i - x j || 2 ), where γ is the kernel function parameter, with a value of 0.1. The optimization objective of the support vector machine is established: Constraint condition: y i (w T x i + b) ≥ 1 - ξ i , ξ i ≥0, where w is the weight vector; b is the bias term; C is the penalty factor, with a value of 100; ξ i is the slack variable. The value range of the stirring uniformity is from 0 to 1, and when it is greater than 0.85, it is considered to reach the uniform stirring state; when the change rate value of the slurry viscosity is less than 0.05 per second, it indicates that the properties of the slurry tend to be stable. This step realizes the real-time monitoring and evaluation of the stirring process through the collaborative work of multiple sensors and the application of the support vector machine algorithm, providing a basis for subsequent parameter adjustment.

[0177] The specific implementation method of step S5 is: based on the stirring uniformity value and the change rate value of the slurry viscosity obtained in step S4, establish the relationship equation between the stirring power and the rotation speed: In the formula, D is the diameter of the stirring paddle, with the unit of m; ρmi x is the density of the mixture, with the unit of kg / m³. Considering the influence of the change in the slurry viscosity, establish the rotation speed correction equation: In the formula, k η is the viscosity correction coefficient, with a value range of 0.2 to 0.8, obtained through experimental calibration. A three-layer feedforward neural network is used to establish the mapping relationship of the rotation speed adjustment function. The number of neurons in the input layer is 2, the number of neurons in the hidden layer is 20, and the number of neurons in the output layer is 2. The hyperbolic tangent activation function is used in both the hidden layer and the output layer: The network training uses the backpropagation algorithm, with a learning rate of 0.01 and a momentum factor of 0.9. The mean square error is used as the loss function: The training data includes historical operation data and laboratory calibration data, with a data volume of no less than 1,000 groups. The adjustment range of the real-time speed value of the mud mixer is ±30% of the initial speed, and the speed change rate does not exceed 10 revolutions per minute per second; the mixing time value is between 10 minutes and 30 minutes. This step realizes the dynamic optimization of mixing parameters through the neural network algorithm, improving the mixing efficiency and slurry quality.

[0178] The specific implementation of step S6 is: using a Hall-type speed sensor to collect the speed value of the curing agent pulping machine, the measurement range is 0 to 300 revolutions per minute, and the accuracy is ±1 revolution per minute. The output value of the curing agent pulping machine is measured by an electromagnetic flowmeter, the range is 0 to 10 cubic meters per hour, and the measurement accuracy is ±0.5% of the full scale. Based on the principle of pipeline transportation dynamics, the pressure loss equation is established: Where f is the pipeline friction coefficient; L is the pipeline length, in meters; D is the pipeline diameter, in meters; ρ is the slurry density, in kilograms per cubic meter; v is the flow rate, in meters per second; K is the local loss coefficient. Considering the non-Newtonian fluid characteristics of the curing agent slurry, a correction term is introduced: Where, τ is the shear stress, in Pa; τ 0 is the yield stress, in Pa; K is the consistency coefficient; is the shear rate in seconds; n is the rheological index. The particle swarm optimization algorithm is used to solve the optimal conveying parameters, and the objective function is: The algorithm's velocity and position update equations are: In the formula, ω is the inertia weight, which is 0.8; c 1 , c 2 is the acceleration constant, all of which are taken as 2; r 1 , r 2 is a random number between 0 and 1. This step optimizes the delivery parameters through the particle swarm optimization algorithm to ensure the stable delivery of the curing agent slurry.

[0179] The specific implementation method of step S7 is: use a variable frequency speed-regulated screw pump to deliver the curing agent slurry to the mud mixer, the maximum delivery pressure of the pump is 2 MPa, and the flow rate range is 0.5 cubic meters per hour to 5 cubic meters per hour. A piezoelectric pressure sensor and an electromagnetic flow sensor are installed in the delivery pipeline, the sampling frequency is 5 Hz, and a proportional integral control variable frequency speed-regulated screw pump is used to achieve closed-loop control of the pressure and flow in the pipeline. The control equation is: In the formula, K p is the proportionality coefficient, K iis the integral coefficient, and the parameter value is determined by the Ziegler-Nichols setting method. When the pressure exceeds 120% of the set value, the pressure relief circuit is automatically opened through the solenoid valve; when the flow deviation exceeds ±10% of the set value, the sound and light alarm signal is triggered. This step achieves accurate addition and uniform mixing of the curing agent.

[0180] The specific implementation of step S8 is: using a Doppler flow meter to measure the feed rate of the slurry after mixing and stirring, the measurement range is 0 to 10 cubic meters per hour, and the measurement accuracy is ±1% of the full scale. The content of the sieve material is determined by sampling, and the detection accuracy is ±0.5%. Establishing a screening efficiency model: In the formula, E is the screening efficiency; k is the screening coefficient; A is the effective area of ​​the screen, in square meters. Considering the influence of screen blockage, the time-varying coefficient is introduced: k = k 0 (1-βt), where k 0 is the initial screening coefficient; β is the blocking coefficient, ranging from 0.001 to 0.005 per minute; t is the running time, in minutes. The Kalman filter algorithm is used for state estimation, and the state space model includes the prediction equation: P k|k-1 =AP k-1|k-1 A T +Q and update equation: K k =P k|k-1 H T (HP k|k-1 H T +R) -1 ; P k|k =(IK k H)P k|k-1 The rotating speed of the drum screen ranges from 20 to 60 rpm, and the speed adjustment accuracy is ±1 rpm. This step achieves optimal control of the screening process through the Kalman filter algorithm, thereby improving the screening efficiency and classification accuracy.

[0181] The specific implementation method of step S9 is: according to the drum rotary screen speed value determined in step S8, a permanent magnet synchronous motor is used to drive the drum rotary screen, the motor power is 5 kW to 15 kW, and the start-up adopts a variable frequency soft start mode, and the start-up time is 5 seconds to 15 seconds. The screen adopts a stainless steel polyurethane composite screen with a mesh size of 2 mm to 5 mm and a tension of 2 kN to 4 kN. During the screening process, an electromagnetic exciter is used to generate high-frequency vibration, with a vibration frequency of 30 Hz to 50 Hz and an amplitude of 1 mm to 3 mm. The vibration parameters are optimized by the vibration equation: In the formula, F 0 is the vibration force amplitude, ω is the angular frequency, is the phase angle. An ultrasonic cleaning device is set up. When the screen blockage rate exceeds 30%, the cleaning procedure is started, and the cleaning efficiency is not less than 95%. The screening efficiency is controlled to be above 90%, and the content of particles with a particle size larger than the size of the screen hole in the screened material does not exceed 1%. This step achieves efficient screening of the slurry through the synergistic effect of multiple mechanical vibrations.

[0182] The specific implementation of step S10 is: use a standard fluidity tester to measure the fluidity value of the slurry after screening, with a measurement range of 160 mm to 260 mm and a measurement accuracy of ±1 mm. Use a Coriolis mass flowmeter to measure the feed rate value of the slurry after screening, with an accuracy of ±0.2%. Establish a slurry fluidity prediction model: F k+1 =aF k +bN second +cQ s +d, where a, b, c, d are model parameters, which are identified online by recursive least squares method: Where λ is the forgetting factor, which ranges from 0.95 to 0.99. Considering the effect of temperature on fluidity: F T =F k+1 [1+α(TT 0 )], where α is the temperature correction coefficient, ranging from 0.001 to 0.003 per degree Celsius. The secondary stirring control function is established based on the model predictive control algorithm, and the prediction model is: The constraints include: min ≤u k ≤u max ; Δu min ≤Δu k ≤Δu max ;y min ≤y k ≤y max . Solve the quadratic programming problem: Where N p is the prediction time domain, with a value of 10, N c To control the time domain, the value is 3. This step achieves precise control of slurry properties through model predictive control.

[0183] The specific implementation method of step S11 is: use a vertical secondary mixer with variable frequency speed regulation to perform final mixing, and establish a fluidized solidified soil performance prediction model: S k =f(N second , T second , F T , Q s ), where S kis the performance index. Adaptive fuzzy neural network is used for online prediction. The network structure is: 4 neurons in the input layer, Gaussian membership function in the hidden layer, and 1 neuron in the output layer. The network learning adopts a hybrid learning algorithm: forward propagation calculates the network output; back propagation optimizes the conclusion layer parameters; and the least squares method optimizes the antecedent layer parameters. The learning rate dynamic adjustment equation is: η k =η 0 exp(-βk), where η 0 is the initial learning rate, which takes a value of 0.1, and β is the attenuation coefficient, which takes a value of 0.01. When the stirring time reaches the set value and the slurry performance index meets the requirements, the stirring is automatically stopped and qualified products are output. This step realizes the quality control of fluidized solidified soil through the application of intelligent algorithms.

[0184] The control algorithms and mathematical models used in all the above steps take into account the physical nature and engineering practice of the process. Through multi-level closed-loop control and online optimization, the precise regulation of the fluidized solidified soil production process is achieved. The selection of various parameters and variables is based on a large amount of experimental data and field operation experience, and has strong practicality and reliability. The entire control system is adaptive and robust, and can effectively cope with the impact of fluctuations in raw material properties and changes in process conditions, ensuring the stability and consistency of product quality.

[0185] In order to better understand and implement the present invention, the following is an example 2 of a specific application scenario of the present invention: A construction team implements a nomadic fluidized solidified soil production process. The original soil at the construction site is mainly beach aquaculture sludge, with a density of 1.85 grams per cubic centimeter, a water content of 25%, and a plasticity index of 18. The construction team uses a mud mixer with a rated power of 30 kilowatts and a volume of 4 cubic meters. The curing agent uses 42.5 grade ordinary Portland cement.

[0186] First, the rated power of the mud mixer is 30 kilowatts, the volume is 4 cubic meters, and the original soil density is 1.85 grams per cubic centimeter. The optimization function is based on a genetic algorithm. After 500 iterations, the optimal parameter combination is obtained, which is 1.6 cubic meters per hour of water per unit time, 2.4 tons of soil feed per unit time per hour, and 120 revolutions per minute of the initial speed of the mud mixer. The water-soil ratio corresponding to this set of parameters is 0.52, which is within the optimal water-soil ratio range of 0.4 to 0.6.

[0187] The production system is started according to the optimized parameters, and water is delivered to the mixing tank through a variable frequency water pump, and the operating frequency of the water pump is set to 35 Hz. After the mixer is started, an excavator with a weighing sensor is used to feed the material. The weighing sensor accuracy is ±0.5 kg, and the excavator bucket capacity is 0.8 cubic meters. The fuzzy control system uses the feed amount deviation and the deviation change rate as input variables, and adjusts the excavator action in real time to keep the feed rate stable at 2.4 tons per hour.

[0188] During the mixing process, the power detection module collects the mixer power value in real time, the temperature sensor collects the temperature value, and the torque sensor collects the torque value. The sampling frequency is 10 Hz. The data records are shown in Table 1 below:

[0189] Table 1 Data record table

[0190] Run time (minutes) Power value (kW) Temperature value (Celsius) Torque value (Nm) 0 25.5 22.5 680 5 27.8 24.3 720 10 28.2 25.8 735 15 28.0 26.2 730 20 27.9 26.5 725

[0191] The monitoring data is input into the mixing condition judgment function, which is based on the support vector machine algorithm and uses the radial basis kernel function. The calculated mixing uniformity value is 0.88, which exceeds the set threshold of 0.85; the slurry viscosity change rate is 0.03 per second, which is less than the stable threshold of 0.05 per second, indicating that the slurry properties tend to be stable.

[0192] The speed adjustment function uses a deep neural network to calculate the real-time speed of the mud mixer based on the mixing parameters and the mixing time should be adjusted to 110 rpm and 25 minutes. The mixing system runs according to the new parameters, and real-time monitoring data shows that the fluctuation range of the mixing power is reduced to within ±2%.

[0193] When the curing agent slurry system is running, the speed sensor measures the speed of the slurry machine to be 250 rpm, and the flow meter measures the output to be 3.5 cubic meters per hour. The transport control function is based on the particle swarm optimization algorithm, and the optimal mud pump pressure value is calculated to be 1.2 MPa and the flow value is 2.8 cubic meters per hour. The mud pump transports the curing agent slurry to the mixer for mixing according to the optimized parameters.

[0194] After mixing, the Doppler flow meter measured the slurry feed rate at 6.5 cubic meters per hour, and the online detection device measured the screen content at 8%. These data were input into the screening control function, and the Kalman filter algorithm was used to calculate the optimal drum screen speed of 45 revolutions per minute. The screening system runs at this speed, and the actual screening efficiency reaches 92%.

[0195] After screening, the slurry fluidity value measured by the standard fluidity meter was 220 mm, and the feed rate value measured by the Coriolis mass flowmeter was 5.8 cubic meters per hour. These parameters were input into the secondary mixing control function, and the secondary mixer speed value was calculated to be 80 rpm and the mixing time value was 12 minutes based on the model predictive control algorithm. The unconfined compressive strength of the fluidized solidified soil obtained in the end was 1.2 MPa at 7 days and 2.8 MPa at 28 days, which met the design requirements.

[0196] During the entire production process, the system operates stably and the fluctuation range of various parameters is controlled within the set range. Compared with the traditional empirical control method, this production process has significant advantages. The traditional method mainly relies on the operator's experience to set the process parameters and uses a single-loop PID controller for simple adjustment. There are the following problems: the setting of feed parameters lacks a theoretical basis, and it often takes multiple tests to determine the appropriate ratio; the mixing process lacks effective monitoring means, and abnormal conditions cannot be discovered and handled in time; the amount of curing agent added and the conveying parameters are difficult to accurately control, resulting in large fluctuations in product performance; the screening and secondary mixing process parameter adjustment lags behind, affecting the stability of product quality.

[0197] The intelligent control method adopted by the present invention overcomes the above shortcomings: the theoretical optimization of the feed parameters is achieved through the genetic algorithm, reducing the number of experiments; the support vector machine algorithm provides real-time monitoring capabilities of the mixing process, and abnormal conditions are handled in a timely manner; the deep neural network ensures the dynamic optimization of the mixing parameters; the particle swarm algorithm realizes the precise control of the curing agent delivery; the Kalman filter and model predictive control ensure the stable operation of the subsequent processes. The entire production process forms a complete closed-loop control system, which realizes the real-time and precise regulation of the process parameters, significantly reduces the fluctuation of product quality, and improves the production efficiency by about 30%. These advances fully demonstrate the innovation and practicality of the present invention in solving the real-time and precise regulation of the process parameters in the production process of nomadic fluidized solidified soil.

[0198] It should be noted that the variables involved in the present invention are explained in detail as shown in Table 2 below.

[0199] Table 2 Variable explanation table

[0200]

[0201] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A nomadic fluidized solidified soil production process, characterized in that: The following steps are involved: Collect the rated power value and volume value of the mud mixer, measure the original soil density value, input the feed parameter optimization function, and obtain the water consumption per unit time, soil feed per unit time, and initial speed value of the mud mixer; Pumping water into the mixing cylinder of the mud mixer according to the water consumption per unit time, starting the mud mixer according to the initial speed value of the mud mixer; conveying raw soil into the mud mixer through the excavator according to the soil material feed amount per unit time; Collect the real-time power value, temperature value and torque value of the mud mixer, input the mixing condition judgment function, and obtain the mixing uniformity value and slurry viscosity change rate value; According to the stirring uniformity value and the slurry viscosity change rate value, input the speed adjustment function to adjust the mud mixer operation parameters; collect the speed value of the curing agent pulping machine and the output value of the curing agent pulping machine, input the conveying control function, and obtain the mud pump pressure value and the mud pump flow value; according to the mud pump pressure value and the mud pump flow value, pump the curing agent slurry into the mud mixer for mixing and stirring; collect the slurry feed rate value and the screen content value after mixing and stirring, input the screening control function, and obtain the drum rotating screen speed value for screening; The fluidity value of the slurry after screening is measured, the feed rate value of the slurry after screening is collected, and the secondary mixing control function is input to obtain the secondary mixer speed value and the secondary mixing time value; secondary mixing is performed according to the secondary mixer speed value and the secondary mixing time value to obtain fluidized solidified soil.

2. The nomadic fluidized solidified soil production process according to claim 1 is characterized in that: The feeding parameter optimization function is based on the principle of maximizing production efficiency, taking into account equipment capacity limitations and material flow characteristics. The input includes the rated power value of the mud mixer, the volume value of the mud mixer, and the original soil density value. The output includes the water consumption per unit time, the soil feed amount per unit time, and the initial speed value of the mud mixer. The feeding parameter optimization function is implemented based on a genetic algorithm, with a population size of 100, the number of iterations of 500, the crossover probability of 0.8, and the mutation probability of 0.

1.

3. The nomadic fluidized solidified soil production process according to claim 1 is characterized in that: The stirring condition determination function is based on the slurry uniformity determination principle, taking into account the stirring energy consumption and material state. The input includes the real-time power value of the mud mixer, the real-time temperature value of the mud mixer, and the real-time torque value of the mud mixer. The output includes the stirring uniformity value and the slurry viscosity change rate value. The stirring condition determination function is implemented based on the support vector machine algorithm, using the radial basis kernel function, the penalty factor is 100, and the kernel function parameter is 0.

1.

4. The nomadic fluidized solidified soil production process according to claim 3 is characterized in that: The transportation control function is based on the principle of pipeline transportation dynamics and takes into account the flow characteristics of the curing agent slurry. The input includes the speed value of the curing agent pulping machine and the output value of the curing agent pulping machine, and the output includes the mud pump pressure value and the mud pump flow value. The transportation control function solves the optimal transportation parameters based on the particle swarm optimization algorithm, with a population size of 50, a maximum number of iterations of 200, and a learning factor of 2.

5. The nomadic fluidized solidified soil production process according to claim 4 is characterized in that: The screening control function is based on the particle classification principle, taking into account the material dispersion and equipment screening efficiency. The input includes the slurry feed rate value and the screen content value, and the output includes the drum screen speed value. The screening control function is implemented based on the Kalman filter algorithm, and the system characteristics are determined through the state estimation matrix and the observation matrix.

6. The nomadic fluidized solidified soil production process according to claim 5 is characterized in that: The secondary mixing control function is based on the performance optimization principle of fluidized solidified soil, taking into account the final fluidity requirements of the slurry. The input includes the fluidity value of the slurry after screening and the feed rate value of the slurry after screening. The output includes the secondary mixer speed value and the secondary mixing time value. The secondary mixing control function is implemented based on the model predictive control algorithm, with a prediction time domain of 10 and a control time domain of 3.

7. The nomadic fluidized solidified soil production process according to claim 6, characterized in that: The stirring uniformity value ranges from 0 to 1. When it is greater than 0.85, it is considered that a uniform stirring state is achieved; when the slurry viscosity change rate value is less than 0.05 per second, it indicates that the slurry properties tend to be stable.

8. The nomadic fluidized solidified soil production process according to claim 7, characterized in that: The adjustment range of the real-time speed value of the mud mixer is ±30% of the initial speed, and the speed change rate does not exceed 10 revolutions per minute per second; the stirring time value of the mud mixer is between 10 minutes and 30 minutes.

9. The nomadic fluidized solidified soil production process according to claim 8, characterized in that: The speed range of the drum screen is 20 to 60 rpm, and the speed adjustment accuracy is ±1 rpm; the speed range of the secondary mixer is 40 to 120 rpm, and the secondary mixing time is between 5 minutes and 15 minutes.

10. The nomadic fluidized solidified soil production process according to claim 9, characterized in that: The pressure value of the mud pump is between 0.5 MPa and 2 MPa; the flow value of the mud pump ranges from 0.5 cubic meters per hour to 5 cubic meters per hour; and the mesh size of the screen is from 2 mm to 5 mm.

Images