Austenitic steel for manufacturing furnace roller and process of austenitic steel

Abstract

The invention provides austenitic steel for manufacturing a furnace roller and a process of the austenitic steel, and belongs to the technical field of austenitic steel, the austenitic steel for manufacturing the furnace roller comprises the chemical components of carbon, silicon, manganese, chromium, nickel, molybdenum, niobium, titanium, aluminum, phosphorus, sulfur and the like according to a specific proportion, and the austenitic steel is manufactured through the processes of smelting, casting, hot working, solution treatment, cold working, aging treatment and the like after raw materials are mixed through a stirring mechanism. The stirring mechanism comprises a stirring barrel, a rotating disc and a shaft which are rotationally connected, various stirring blades, a motor and related transmission parts, accurate temperature, time, deformation and the like are controlled in each step in the process, the austenitic steel has good high-temperature performance and comprehensive mechanical performance, the service life of a furnace roller can be prolonged, the manufacturing process meets the high-performance requirement, the cost can be reduced, and the production cost is reduced. The problems that an existing furnace roller material is insufficient in high-temperature strength, prone to deformation and abrasion, limited in manufacturing process and the like are solved, and the material is suitable for manufacturing furnace rollers of equipment such as industrial furnaces and has important practical application value.

Description

Technical Field

[0001] The invention belongs to the technical field of austenitic steel, and in particular relates to austenitic steel for manufacturing furnace rollers and a process thereof. Background Art

[0002] Austenitic steel is steel with austenite structure after normalizing. The alloying elements (Ni, Mn, N, Cr, etc.) added to the steel can make the metal have a stable austenite structure after normalizing. Steels such as 1Cr18Ni9 and 1Cr18Ni9Ti commonly used in water treatment equipment and steam sampling pipes in thermal power plants belong to this category. This type of steel has good oxidation resistance and acid resistance, and can work at 540-875℃ for a long time, but it is easy to produce brittle damage of intergranular corrosion.

[0003] Furnace rollers play an important role in industrial furnaces and other equipment, and they need to operate stably for a long time under high temperature and complex working conditions. Existing furnace roller materials often have some problems during use, such as insufficient high temperature strength, easy deformation and wear, resulting in a short service life of the furnace rollers, requiring frequent replacement, increasing production costs and equipment downtime. At the same time, the existing manufacturing process may not meet certain high performance requirements, or have certain limitations in process complexity and cost control. Therefore, it is of great practical significance to develop an austenitic steel for furnace rollers with good high temperature performance and comprehensive mechanical properties and its optimized manufacturing process. . Summary of the invention

[0004] The object of the present invention is to provide an austenitic steel for manufacturing furnace rollers and a process thereof, aiming to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: an austenitic steel for manufacturing furnace rollers, the chemical composition of which includes, by mass percentage, carbon 0.05%-0.12%, silicon 1.0%-2.0%, manganese 0.5%-2.5%, chromium 18%-22%, nickel 8%-12%, molybdenum 0.5%-1.5%, niobium 0.05%-0.15%, titanium 0.05%-0.15%, aluminum 0.02%-0.08%, phosphorus ≤0.03%, sulfur ≤0.03%, and the balance is iron and unavoidable impurities

[0006] As a preferred solution of the present invention, its chemical composition, measured by mass percentage, includes: a carbon content of 0.08%-0.1%, a silicon content of 1.5%-1.8%, a manganese content of 2.0%-2.2%, a chromium content of 20%-21%, a nickel content of 10%-11%, a molybdenum content of 1.0%-1.2%, a niobium content of 0.1%-0.12%, a titanium content of 0.1%-0.12%, and an aluminum content of 0.05%-0.06%.

[0007] A preparation process of austenitic steel for manufacturing furnace rollers, the specific steps are:

[0008] S1, smelting: the raw materials are first mixed according to the proportion by a stirring mechanism, and then added into an electric furnace for smelting after the mixing is completed. The smelting temperature is controlled at 1550°C-1650°C. After the raw materials are completely melted, they are refined for 30-60 minutes. Argon blowing and stirring are used during the refining process. Then the molten steel is poured into a ladle for secondary refining outside the furnace to make the oxygen content in the molten steel ≤20ppm and the sulfur content ≤0.01%;

[0009] S2. Casting: Pour the refined molten steel into a pre-prepared casting mold, control the casting temperature at 1450°C-1550°C, and use the pouring method or continuous casting method for casting. A proper amount of modifier can be added during the casting process;

[0010] S3, hot working: heating the steel billet at a temperature of 1150°C-1250°C for 2-4 hours, and then hot rolling or forging, with the starting rolling (forging) temperature controlled at 1100°C-1150°C, the final rolling (forging) temperature not less than 900°C, and the hot working deformation controlled at 30%-60%;

[0011] S4, solution treatment: the hot-processed steel is subjected to solution treatment at a solution temperature of 1050°C-1150°C for 1-3 hours, followed by rapid cooling at a cooling rate greater than 50°C / s;

[0012] S5. Cold working: cold working the steel after solution treatment according to actual needs, and the cold working deformation is controlled within 10%-30%;

[0013] S6. Aging treatment: Finally, the cold-processed steel is subjected to aging treatment at a temperature of 550°C-650°C, with a holding time of 4-8 hours, followed by air cooling.

[0014] As a preferred solution of the present invention, the stirring mechanism comprises:

[0015] Mixing tank;

[0016] A lower turntable, rotatably connected to the bottom wall of the mixing barrel;

[0017] A driving shaft, one end of which is fixed to the top of the lower turntable, and the other end of which is rotatably connected to the top wall of the mixing barrel;

[0018] A central stirring blade, which is provided in plurality and is evenly fixed on the circumferential surface of the driving shaft;

[0019] There are two driven shafts, both of which are rotatably connected to the top of the lower turntable and are respectively located on both sides of the driven shaft;

[0020] Edge stirring blades, which are provided in plurality and are evenly fixed to the circumferential surfaces of the two driven shafts, and the plurality of edge stirring blades and the plurality of central stirring blades are staggered;

[0021] A motor is fixed to the top of the mixing barrel, and the output end of the motor movably penetrates into the mixing barrel and is fixed to the top of the driving shaft;

[0022] An annular rack fixed to the upper portion of the circumferential inner wall of the mixing barrel;

[0023] The gears are provided with two gears, which are respectively fixed to the circumferential surfaces of the two driven shafts and mesh with the annular racks.

[0024] As a preferred solution of the present invention, an upper turntable is fixed on the upper part of the circumferential surface of the driving shaft, the upper turntable is sleeved on the circumferential surfaces of the two driven shafts and is rotatably connected to the two driven shafts, a supporting ring plate is fixed on the circumferential inner wall of the mixing barrel, and the upper turntable is rotatably connected to the upper part of the supporting ring plate.

[0025] As a preferred solution of the present invention, a feed hopper is fixedly installed on the upper part of the circumferential surface of the mixing barrel, a discharge pipe is fixedly installed on the bottom of the mixing barrel, and an arc-shaped scraper plate is fixed on the circumferential surface of the lower turntable, and one end of the arc-shaped scraper plate away from the lower turntable matches the circumferential inner wall of the mixing barrel.

[0026] As a preferred solution of the present invention, the secondary refining outside the furnace in the smelting step adopts a ladle refining furnace or a vacuum degassing process, and the modifier added in the casting step is a rare earth element.

[0027] As a preferred solution of the present invention, in the hot rolling or forging in the thermal processing step, when hot rolling is adopted, the surface roughness Ra of the rolling roller is controlled to be 0.8-1.6 μm.

[0028] As a preferred solution of the present invention, in the solution treatment step, rapid cooling is performed by water quenching or oil quenching, and the cold working method in the cold working step is cold rolling or cold drawing.

[0029] As a preferred solution of the present invention, the steel material made of austenitic steel has a yield strength of ≥450MPa, a tensile strength of ≥650MPa, an elongation of ≥30%, and a hardness (HB) of ≤200.

[0030] A preparation process of austenitic steel for furnace roller manufacturing, wherein a neural fuzzy inference system is integrated in the solution treatment system, and the system is used to monitor and dynamically adjust the solution temperature and holding time in real time, and the input of the neural fuzzy inference system includes a solution temperature sensor, a holding time monitoring relay and a cooling wind speed sensor, and its output is optimized by a differential evolution algorithm and transmitted to the solution treatment electric heating controller and the heat treatment digital timing relay, and the cooling control module receives the cooling speed adjustment signal. The specific actuators in the system include an inductive heater, a thermocouple temperature sensor and a digital timing relay. The optimization process uses the Arctic Sea Parrot optimization algorithm to further adjust the solution parameters to ensure the uniformity of the steel microstructure and the stability of the mechanical properties, and is characterized in that the following steps are included:

[0031] Step 1: Data collection and preprocessing

[0032] 1. Data Collection

[0033] The following key parameters need to be collected as input to the model:

[0034] Solution treatment temperature T s :

[0035] Thermocouple temperature sensors are used. This type of sensor is suitable for high temperature measurement, with high accuracy and fast response. It is suitable for real-time monitoring of the high temperature environment of solution treatment. The thermocouple probe is directly in close contact with the workpiece to ensure the accuracy and response speed of temperature measurement and obtain the solution treatment temperature T. s ;

[0036] Insulation time t h :

[0037] Use the heat treatment digital timing relay to accurately record and control the duration of the holding process. It has high accuracy and can be automatically controlled. Integrate the heat treatment digital timing relay into the control system. When the process starts or reaches the set temperature, the timing will automatically start, and it will automatically stop after the set holding time is completed. The holding time t h ;

[0038] Cooling speed v c :

[0039] The cooling air velocity sensor is used to monitor the speed of the cooling air flow to ensure accurate control of the cooling speed, monitor the flow rate of the cooling medium, and combine with the thermocouple temperature sensor to monitor the temperature drop rate of the workpiece surface to comprehensively evaluate the cooling speed v c ;

[0040] Steel microstructure changes μ:

[0041] Take steel samples, perform standardized cutting, polishing and etching, observe the samples microscopically under a scanning electron microscope, obtain images of grain size, phase structure and tissue distribution, analyze the scanning electron microscope images using image processing software, calculate the grain size and phase content, and quantify the microstructural changes μ;

[0042] Mechanical properties of steel P:

[0043] Prepare steel samples that meet the test standards, use a tensile testing machine to apply gradually increasing tensile loads to the samples, record the load and displacement, calculate the tensile strength, yield strength, elongation and other parameters based on the load-displacement curve, and comprehensively obtain the mechanical properties P of the steel;

[0044] These data will be used to train the neuro-fuzzy inference system, but before that, they need to be normalized to ensure that data of different scales will not have an unbalanced impact due to different units during the training process;

[0045] These data will be used to train the neuro-fuzzy inference system, but before that, they need to be normalized to ensure that data of different scales will not have an unbalanced impact due to different units during the training process;

[0046] 2. Data Normalization

[0047] The purpose of normalization is to map data of different units and magnitudes to the same range. The normalization formula is:

[0048]

[0049] X′ is the normalized value, X is the original data, and X min is the minimum value of the data, X max is the maximum value of the data;

[0050] 3. Normalization formula of each parameter

[0051] Normalization of solution treatment temperature:

[0052]

[0053] T s ′ is the normalized solution treatment temperature, T s is the actual solution treatment temperature, T s,min is the minimum solution treatment temperature, T s,max is the maximum value of solution treatment temperature;

[0054] Normalization of holding time:

[0055]

[0056] t′ h is the normalized holding time, t h is the actual holding time, t h,min is the minimum holding time, t h,max is the maximum value of the holding time;

[0057] Normalization of cooling rate:

[0058]

[0059] v′ c is the normalized cooling rate, v c is the actual cooling rate, v c,min is the minimum cooling rate, v c,max is the maximum value of cooling rate;

[0060] Normalization of microstructural changes:

[0061]

[0062] μ′ is the normalized microstructure change, μ is the actual measured microstructure change value, and μ min is the minimum value of microstructure change, μ max is the maximum value of microstructural change;

[0063] Normalization of mechanical properties:

[0064]

[0065] P′ is the normalized mechanical property value, P is the actual measured mechanical property, and P min is the minimum value of mechanical properties, P max is the maximum value of mechanical properties;

[0066] Step 2: ANFIS modeling of neural fuzzy inference system

[0067] 1. Fuzzification of input variables

[0068] Each input variable is divided into several fuzzy sets through fuzzification. i is divided into m fuzzy sets, then the fuzzy membership function is:

[0069]

[0070] is the membership function of the jth fuzzy set of the ith input variable, X i is the i-th input variable, c i,j is the center of the jth fuzzy set, b is the parameter that controls the shape of the membership function, ai,j To control the width of the membership function;

[0071] 2. Fuzzy rule base

[0072] Based on the fuzzy set of input variables, a fuzzy rule base is constructed, and the output variables μ and P can be expressed as linear combinations of the input variables:

[0073] f1(T s ,t h ,v c )=p1T s +q1t h +r1v c +s1

[0074] f2(T s ,t h ,v c )=p2T s +q2t h +r2v c +s2

[0075] f1(T s ,t h ,v c ) is the output function of the fuzzy inference system under a certain fuzzy rule, p1 is the solution treatment temperature T s The linear coefficient, T s is the solution treatment temperature, q1 is the holding time t h The linear coefficient, t h is the holding time, r1 is the cooling rate v c The linear coefficient, v c is the cooling rate, s1 is a constant term, f2(T s ,t h ,v c ) is one of the outputs of the fuzzy rule reasoning system, p2 is the solution treatment temperature T s The linear coefficient, q2 is the holding time t h The linear coefficient, r2 is the cooling rate v c The linear coefficient of , s2 is the constant term;

[0076] 3. Reasoning and output calculation

[0077] Using the fuzzy rule base for reasoning, the weight calculation method for each fuzzy rule is:

[0078]

[0079] w k is the activation weight of the kth fuzzy rule, ∏ is the continuous product symbol, μ is the membership function, A i,j is the i-th input variable Xi The jth fuzzy set, X i is the i-th input variable;

[0080] The final output is calculated by weighted average:

[0081]

[0082] μ is the predicted microstructure change of steel, ∑ is the summation symbol, k is the current rule number, n is the total number of fuzzy rules, and w k is the activation weight of the kth fuzzy rule, f 1,k (T s ,t h ,v c ) is the output function of the fuzzy inference system under a certain fuzzy rule, T s is the solution treatment temperature, t h is the holding time, v c is the cooling rate;

[0083] P is the mechanical property of steel, f 2,k (T s ,t h ,v c ) is the output function of the fuzzy inference system under a certain fuzzy rule;

[0084] 4. Learning ability of neural networks

[0085] The ANFIS system optimizes the parameters in the fuzzy rule base through the learning ability of the neural network. The optimization process is usually carried out through the error back propagation algorithm, and the goal is to minimize the error of the system;

[0086] The error is defined as:

[0087]

[0088] E is the total error function, To sum all data samples from k = 1 to k = N, is the predicted output value of the kth sample, μ k is the true output value of the kth sample, is the predicted output value of the kth sample, P k is the true output value of the kth sample;

[0089] Step 3: Differential evolution algorithm DE optimizes ANFIS model

[0090] 1. Initialize the population

[0091] In the differential evolution algorithm, the population needs to be initialized first. Each individual in the population represents a candidate solution of the ANFIS model, which contains the parameters of the fuzzy rule base and the membership function. The population size is N, and each individual represents the i-th individual in the t-th generation of the ANFIS model, which is defined as:

[0092]

[0093] is, t is the number of generations in the evolutionary process, i is the index number of the individual in the population, is the central parameter of the jth fuzzy rule membership function of the ith individual in the tth generation, is the width parameter of the membership function of the jth fuzzy rule of the i-th individual in the t-th generation, One of the linear regression parameters used in the ANFIS model to predict the microstructural homogeneity output is related to the solution treatment temperature. It is one of the linear regression parameters used in the ANFIS model to predict the uniformity of microstructure and is related to the holding time. It is one of the linear regression parameters used in the ANFIS model to predict the microstructure uniformity and is related to the cooling rate. is the adjustment amount of the bias term used in the ANFIS model to predict the uniformity of microstructure and is independent of the input variables. One of the linear regression parameters used in the ANFIS model to predict the mechanical property output is related to the solution treatment temperature. One of the linear regression parameters used in the ANFIS model to predict mechanical property output is related to the holding time. One of the linear regression parameters used in the ANFIS model to predict the mechanical property output is related to the cooling rate. It is the bias term used in the ANFIS model to predict the output of mechanical properties, and is an adjustment quantity that is independent of the input variables;

[0094] 2. Mutation Operation

[0095] In each generation t, for each individual Generate mutation vector, mutation vector The generation method is as follows:

[0096]

[0097] is the mutation vector of the ith individual in the t+1th generation, is the solution of the r1th individual in the tth generation, F is the scaling factor, is the solution of the r2th individual in the tth generation, is the solution of the r3th individual in the tth generation;

[0098] 3. Crossover Operation

[0099] Through the crossover operation, the mutant individuals With the original individual Combination to generate new candidate solutions The crossover formula is as follows:

[0100]

[0101] is the jth parameter of the i-th individual of the new candidate solution generated in the t+1th generation, is the jth parameter of the ith individual generated by mutation in the t+1th generation, is the jth parameter of the ith individual of the original individual in the tth generation, rand(0,1) is a random number generation function, and CR is the crossover probability;

[0102] 4. Select an action

[0103] The selection operation selects individuals with better performance to enter the next generation through the fitness function. The fitness function evaluates the performance of the ANFIS model based on the microstructural uniformity μ and mechanical properties P of the steel, which is defined as:

[0104]

[0105] is the fitness function, N is the total number of samples, To sum all data samples from k = 1 to k = N, is the predicted output value of the kth sample, μ k is the true output value of the kth sample, is the predicted output value of the kth sample, P k is the true output value of the kth sample;

[0106] The selection steps are as follows:

[0107]

[0108] is the state of the i-th individual in the next generation, is a new individual generated by mutation and crossover operations, is the state of the i-th individual in the current generation, For newly generated individuals The fitness value of For the current individual The fitness value of

[0109] Through this selection process, candidate solutions with lower individual fitness enter the next generation population, thereby continuously optimizing the fuzzy rules and membership function parameters in the ANFIS model;

[0110] Step 4: Arctic Puffin Optimization Algorithm APO Global Optimization

[0111] 1. Initialization

[0112] Based on the optimization of the differential evolution algorithm, the Arctic Sea Parrot optimization algorithm uses this result as a reference for the initial solution space and sets the initial population, which represents the temperature T of the solution treatment. s , insulation time t h and cooling rate v c Initial value of:

[0113]

[0114] X i is the parameter vector of the i-th puffin individual, is the solution treatment temperature of the ith individual, is the holding time of the ith individual, is the cooling rate of the i-th individual;

[0115] 2. Foraging strategy simulation

[0116] The foraging strategy of the Arctic Puffin simulates its process of finding food in the environment. The algorithm randomly selects different combinations of process parameters for exploration and evaluates the effects of these combinations on the microstructural uniformity and mechanical properties of the steel.

[0117] Each sea parrot individual X i The objective function f(X i ) is defined as the evaluation of microstructural homogeneity and mechanical properties based on a combination of process parameters:

[0118]

[0119] f(X i ) is, w1 is the weight coefficient, N is the total number of samples, To sum all data samples from k = 1 to k = N, is the predicted output value of the kth sample, μ k is the true output value of the kth sample, w2 is the weight coefficient, is the predicted output value of the kth sample, P k is the true output value of the kth sample;

[0120] 3. Global Search and Local Optimization

[0121] APO simulates global search and local exploitation through the foraging behavior of sea parrots, where each individual moves randomly in the sea to find the optimal combination of process parameters;

[0122] Global search: In the early stages, the puffins will conduct a large-scale exploration to find new parameter combinations. At this stage, the solution update formula is:

[0123]

[0124] is the solution vector of the ith individual in the t+1th generation, is the solution vector of the ith individual in the tth generation, r is a random number between the interval [0,1], is the solution vector of the individual with the best fitness in the tth generation, F is the scaling factor, is the solution vector of the r1th individual randomly selected from the tth generation population, is the solution vector of the r2th individual randomly selected from the tth generation population;

[0125] Local development: When the Puffin is close to the optimal solution, it enters the local optimization stage. Local optimization aims to fine-tune the process parameters and further improve the results. The formula for local search is:

[0126]

[0127] is the solution vector of the ith individual in the t+1th generation, is the solution vector of the ith individual in the tth generation, α is the shrinkage factor, is the local optimal solution in the tth generation;

[0128] 4. Convergence judgment

[0129] The convergence of the APO algorithm is judged by evaluating whether the fitness function value of the population converges to a certain threshold. If the fitness change of the optimal individual in multiple consecutive generations is less than a preset threshold ∈, the algorithm stops. The convergence judgment condition is:

[0130]

[0131] is the fitness value of the best individual in the t+1 generation population, is the best individual in the t+1 generation population, is the fitness value of the best individual in the t-th generation population, is the solution vector of the individual with the best fitness in the tth generation;

[0132] Once the convergence condition is met, the algorithm stops and outputs the final optimal parameter combination X best, that is, the optimal solution treatment temperature, holding time and cooling rate;

[0133] Step 5: Real-time feedback and optimization

[0134] 1. Real-time data collection and monitoring

[0135] During the solution treatment process, sensors monitor the following key process parameters in real time:

[0136] Solution treatment temperature T s (t);

[0137] Insulation time t h (t);

[0138] Cooling speed v c (t);

[0139] t is the real-time data at the current time point. The sensor collects real-time data and sends it as input to the ANFIS model optimized by the differential evolution algorithm for processing;

[0140] 2. Real-time feedback input to ANFIS model

[0141] The data collected by the sensor in real time s (t),t h (t), and v c (t) is used as input to the previously DE-optimized ANFIS model, which outputs two main results:

[0142] Prediction of microstructure homogeneity

[0143] Mechanical properties prediction

[0144] The input-output relationship of the ANFIS model can be expressed as:

[0145]

[0146] is the predicted uniformity of the steel microstructure at time t, is the predicted mechanical properties of steel at time t, ANFIS is an adaptive neural fuzzy inference system, T s (t) is the solution treatment temperature at time t, t h (t) is the holding time at time t, v c (t) is the cooling rate at time t;

[0147] 3. Real-time feedback and adjustment mechanism

[0148] When the predicted value output by the ANFIS model is and When the expected target value is deviated, the system will automatically adjust according to the feedback;

[0149] The target microstructure uniformity is μ target and the target mechanical properties are P target , the error is expressed as:

[0150]

[0151] E μ (t) is the microstructure uniformity error at time t, μ target To achieve microstructural uniformity, is the predicted uniformity of the steel microstructure at time t;

[0152] E P (t) is the mechanical property error at time t, P target For the target mechanical properties, is the predicted mechanical properties of steel at time t;

[0153] When the error E μ (t) and E P (t) When it exceeds the set tolerance range, the system automatically adjusts the current process parameter T in combination with the Arctic Sea Parrot optimization algorithm. s (t),t h (t),v c (t), in order to reduce the error;

[0154] 4. APO is used for process parameter adjustment

[0155] Combined with the APO algorithm, the current process parameters are dynamically adjusted to ensure the stable performance of the steel. The adjusted process parameters can be expressed by the following formula:

[0156]

[0157] is the solution treatment temperature after dynamic adjustment at time t+1, T s (t) is the solution treatment temperature at the current time t, α1 is the error E of temperature to microstructure uniformity μ (t) Response level, E μ (t) is the microstructure uniformity error, α2 is the temperature error on mechanical properties E P (t) Response level, E P (t) is the mechanical property error;

[0158] is the holding time after dynamic adjustment at time t+1, t h(t) is the holding time at the current time t, β1 is the error E of the holding time on the microstructure uniformity μ (t) Response level, E μ (t) is the microstructure uniformity error, β2 is the mechanical property error of the holding time E P (t) Response level, E P (t) is the mechanical property error;

[0159] is the cooling rate after dynamic adjustment at time t+1, v c (t) is the cooling rate at the current time t, γ1 is the error E of the cooling rate to the microstructure uniformity μ (t) Response level, E μ (t) is the microstructure uniformity error, γ2 is the cooling rate error on mechanical properties E P (t) Response level, E P (t) is the mechanical property error;

[0160] 5. Continuous optimization and data update

[0161] As time goes by, the system will continue to obtain real-time data from sensors and dynamically update process parameters. Due to the adaptive nature of the APO algorithm, the system will continuously perform global and local optimization to keep the process parameters moving in the optimal direction.

[0162] At each time step t, the new process parameters The data are input into the system and fed back into the ANFIS model for re-prediction. This closed-loop control system ensures that the process is always optimized.

[0163] 6. Convergence conditions

[0164] The real-time optimization system will continue to adjust the process parameters until the error E μ (t) and E P (t) converges to the preset threshold ∈ μ and ∈ P ,Right now:

[0165] |E μ (t)|<∈ μ

[0166] |E P (t)|<∈ P

[0167] E μ (t) is the microstructure uniformity error, ε μ is the error convergence threshold of microstructure uniformity, E P (t) is the mechanical property error, ∈P is the error convergence threshold of mechanical properties;

[0168] Once the error meets the convergence conditions, the system enters a stable state and the process parameters also stabilize, ensuring that the microstructure and mechanical properties of the steel are optimal.

[0169] Compared with the prior art, the present invention has the following beneficial effects:

[0170] 1. In the present invention, through the optimized chemical composition design, the alloy elements in the austenitic steel cooperate with each other, which effectively improves the high-temperature strength of the furnace roller. The chromium element can form a stable oxide film to prevent oxygen from further corroding the steel matrix, thereby improving the steel's resistance to high-temperature oxidation and, at the same time, enhancing the steel's high-temperature strength to a certain extent; the molybdenum element can solid-solution strengthen the matrix and increase the recrystallization temperature of the steel, thereby enhancing the high-temperature strength; the carbonitrides formed by niobium and titanium elements can refine the grains, and the refined grains can improve the strength of the steel, especially the high-temperature strength, because the fine grain structure can hinder the movement of dislocations at high temperatures, so that the furnace roller can better withstand various loads in a high-temperature environment, reduce the risk of deformation and damage, and improve the stability and service life of the furnace roller under high-temperature conditions.

[0171] 2. In the present invention, the steel has good oxidation resistance and corrosion resistance through reasonable chemical composition design. The silicon element helps to improve the oxidation resistance of the steel and forms a silicon oxide film on the steel surface. The silicon oxide film and the chromium oxide film formed by the chromium element work together to enhance the oxidation resistance of the steel and reduce the damage of high-temperature oxidation to the furnace roller. The content of harmful impurities such as phosphorus and sulfur is strictly controlled, and elements such as molybdenum are added to improve the corrosion resistance of the steel, so that the furnace roller has better resistance in a complex corrosive environment, prolongs the service life of the furnace roller, and reduces the maintenance cost.

[0172] 3. The present invention significantly improves the comprehensive mechanical properties of steel by combining optimized chemical composition with manufacturing process. Nickel expands the austenite phase region and improves the toughness of steel, making the furnace roller less likely to break when subjected to impact and other loads. Manganese improves the hardenability and strength of steel, and is beneficial to the formation of austenite, providing a good organizational basis for subsequent heat treatment and other processes. Through process steps such as hot working, solution treatment, cold working and aging treatment, the grains are further refined, and the comprehensive properties of steel such as strength, toughness and plasticity are improved. The deformation amount and temperature are controlled during the hot working process to refine and homogenize the grains. The solution treatment obtains a single-phase austenite structure, which improves the toughness and corrosion resistance of the steel. The cold working improves the strength and dimensional accuracy of the steel. The aging treatment makes the precipitation phase in the steel precipitate evenly, further improving the strength and hardness of the steel, while maintaining good toughness. These improvements in comprehensive mechanical properties enable the furnace roller to better adapt to complex working conditions and improve the operating reliability and efficiency of the equipment.

[0173] 4. This patent realizes the intelligence and optimization of the solution treatment process by introducing intelligent algorithms. First of all, the neural fuzzy reasoning system combines fuzzy logic with the learning ability of neural networks, and can accurately predict the effects of solution treatment temperature, holding time and cooling rate on the microstructure uniformity and mechanical properties of steel. The differential evolution algorithm optimizes the fuzzy rule base and membership function parameters during the training process, improving the prediction accuracy of the model. Further, the global and local process parameter optimization is achieved through the Arctic Puffin optimization algorithm to ensure the stability and consistency of steel performance. In practical applications, sensors monitor process parameters in real time, and dynamically adjust process parameters in combination with real-time feedback mechanisms to ensure that product performance is continuously in the optimal state. The entire system not only improves the automation and intelligence level of the process, but also greatly enhances the stability, efficiency and quality consistency of steel production. This optimization method based on intelligent algorithms can effectively adapt to different production conditions and raw material changes, and has significant practicality and creativity.

[0174] The present invention significantly improves the high-temperature strength, oxidation resistance and corrosion resistance of the furnace rollers by optimizing the chemical composition and intelligent process control. The synergistic effect of elements such as chromium, molybdenum, niobium and titanium enhances the stability and durability of steel in high-temperature environments and reduces the risk of deformation and damage; the oxide films of silicon and chromium further improve the oxidation resistance, and molybdenum enhances the corrosion resistance, which extends the service life of the furnace rollers and reduces maintenance costs. Combined with process steps such as hot working, solution treatment, cold working and aging treatment, the grains are further refined, and the strength, toughness and plasticity of the steel are improved. The introduction of intelligent algorithms realizes the automatic optimization of process parameters, ensures the stability and consistency of product performance, adapts to changing production conditions, and greatly improves the automation level and production efficiency of the process. BRIEF DESCRIPTION OF THE DRAWINGS

[0175] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0176] Figure 1 This is a flow chart of a process for preparing austenitic steel for manufacturing furnace rollers according to the present invention;

[0177] Figure 2 It is a three-dimensional diagram of the stirring mechanism in the present invention;

[0178] Figure 3 It is a first cross-sectional stereoscopic view of the stirring mechanism in the present invention;

[0179] Figure 4 It is a second cross-sectional stereoscopic view of the stirring mechanism in the present invention;

[0180] Figure 5It is a third cross-sectional stereoscopic view of the stirring mechanism in the present invention;

[0181] Figure 6 Iterative convergence curve diagram for process parameter optimization;

[0182] Figure 7 Optimization flow chart for solution treatment process based on intelligent algorithm.

[0183] In the figure: 1, mixing barrel; 101, feed hopper; 102, discharge pipe; 2, driving shaft; 201, central stirring blade; 3, driven shaft; 301, edge stirring blade; 4, lower turntable; 5, upper turntable; 6, support ring plate; 7, motor; 8, ring rack; 9, gear; 10, arc scraper plate. DETAILED DESCRIPTION

[0184] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0185] See also Figure 1-Figure 4 , the present invention provides the following technical solutions:

[0186] The invention discloses an austenitic steel for manufacturing furnace rollers, wherein the chemical composition thereof comprises, by mass percentage, 0.05%-0.12% carbon, 1.0%-2.0% silicon, 0.5%-2.5% manganese, 18%-22% chromium, 8%-12% nickel, 0.5%-1.5% molybdenum, 0.05%-0.15% niobium, 0.05%-0.15% titanium, 0.02%-0.08% aluminum, ≤0.03% phosphorus, ≤0.03% sulfur, and the balance is iron and unavoidable impurities.

[0187] In a specific embodiment of the present invention, in the present invention, the content of carbon is controlled within the above range, which can not only ensure the strength of the steel, but also avoid the formation of excessive carbides at high temperatures to reduce the toughness and corrosion resistance of the steel. The silicon element helps to improve the oxidation resistance and strength of the steel, but too high a silicon content will affect the toughness and welding performance of the steel, so it is controlled at 1.0%-2.0%. The manganese element can improve the hardenability and strength of the steel, while reducing the critical cooling rate of the steel, which is conducive to the formation of austenite, and the content is set to 1.5%-2.5%. The chromium element can significantly improve the oxidation resistance and corrosion resistance of the steel, and is between 18%-22%. A stable oxide film can be formed at a content of 8%-12%. Nickel can expand the austenite phase area, improve the toughness and low-temperature performance of steel, and the content is 8%-12%. Molybdenum can improve the high-temperature strength and corrosion resistance of steel, and the content of 0.5%-1.5% can effectively play its role. Niobium and titanium can form carbonitrides, which play the role of grain refinement and precipitation strengthening, and the content is controlled at 0.05%-0.15% and 0.05%-0.15% respectively. Aluminum can deoxidize and refine grains, and the content is controlled at 0.02%-0.08%. The content of phosphorus and sulfur is strictly controlled to ensure the purity and good comprehensive performance of steel.

[0188] Specifically, a preparation process of austenitic steel for manufacturing furnace rollers is described.

[0189] S1, smelting: weigh 0.08% carbon, 1.5% silicon, 2.0% manganese, 20% chromium, 10% nickel, 1.0% molybdenum, 0.1% niobium, 0.1% titanium, 0.05% aluminum, 0.02% phosphorus, 0.01% sulfur, and the balance is iron (Fe) and inevitable impurities. Mix all the raw materials according to the proportion through a stirring mechanism, add them into an electric furnace for smelting after the mixing is completed, and control the smelting temperature at 1550°C-1650°C. After the raw materials are completely melted, perform refining treatment for 30-60 minutes. During the refining process, argon blowing and stirring are adopted, and then the molten steel is poured into a ladle for secondary refining outside the furnace to make the oxygen content in the molten steel ≤20ppm and the sulfur content ≤0.01%;

[0190] Adding the raw materials into the electric furnace for smelting according to the proportion and controlling the temperature at 1550°C-1650°C can fully melt the raw materials and provide uniform molten steel for subsequent refining. (In Example 1, the raw materials are completely melted by smelting in this temperature range, which lays the foundation for subsequent processes.) This temperature range can not only ensure the full melting of the raw materials, but also avoid excessive temperature causing energy waste and excessive wear on the furnace body;

[0191] Refining for 30-60 minutes and blowing argon for stirring can make the composition and temperature of the molten steel uniform, remove inclusions, and improve the quality of the molten steel. (In Example 1, the refining was performed for 45 minutes, and the argon blowing and stirring made the composition of the molten steel more uniform, reduced inclusions, and improved the purity of the steel.) The control of the refining time is to ensure that the inclusions have enough time to float up and be removed. The argon blowing and stirring can promote the uniform mixing of the molten steel and the floating of the inclusions, thereby improving the quality of the steel and reducing the defects caused by inclusions during the use of the furnace roller;

[0192] Secondary refining outside the furnace (such as LF or VD process) makes the oxygen content in the molten steel ≤20ppm and the sulfur content ≤0.01%, further adjusts the composition and purity of the molten steel, and ensures the performance of the steel. (Example 1 adopts secondary refining outside the furnace to meet the requirements of low oxygen and low sulfur, and improves the quality of the steel.) Low oxygen content can reduce oxide inclusions and improve the toughness and fatigue performance of the steel; low sulfur content can enhance the corrosion resistance of the steel, which is crucial for the long-term stable operation of the furnace roller under complex working conditions;

[0193] S2. Casting: Pour the refined molten steel into a pre-prepared casting mold, control the casting temperature at 1450°C-1550°C, and use the pouring method or continuous casting method for casting. A proper amount of modifier can be added during the casting process;

[0194] The casting temperature is controlled at 1450℃-1550℃, and the pouring method or continuous casting method is used for casting, which can smoothly shape the molten steel into a billet. (In Example 1, the casting temperature is 1500℃, and the pouring method is used for casting to obtain a billet of good quality.) This temperature range is suitable for the solidification of the molten steel. The pouring method or continuous casting method can be selected according to the production scale and product requirements, which can ensure the quality and production efficiency of the billet;

[0195] Adding an appropriate amount of modifier (such as rare earth elements) during the casting process can refine the grains and improve the structure and performance of the steel; (the rare earth modifier is added in Example 1 to refine the grains and improve the performance of the steel;) Rare earth elements can be adsorbed at positions such as grain boundaries to hinder the growth of grains, thereby refining the grains, making the structure of the steel more uniform, improving the strength, toughness and other properties, and extending the service life of the furnace roller;

[0196] S3, hot working: heating the steel billet at a temperature of 1150°C-1250°C for 2-4 hours, and then hot rolling or forging, with the starting rolling (forging) temperature controlled at 1100°C-1150°C, the final rolling (forging) temperature not less than 900°C, and the hot working deformation controlled at 30%-60%;

[0197] The heating temperature is 1150°C-1250°C, and the holding time is 2-4 hours, so that the steel billet is fully and evenly austenitized; (In Example 1, the steel billet is heated to 1200°C and held for 3 hours to achieve austenitization of the steel billet, providing a good organizational foundation for subsequent processing;) Within this temperature and time range, the organization in the steel billet can be fully transformed into austenite, which is a kind of organization with good plasticity and toughness, which is conducive to subsequent hot rolling or forging processing and improves the comprehensive performance of the steel;

[0198] The starting rolling (forging) temperature is controlled at 1100°C-1150°C, and the final rolling (forging) temperature is not less than 900°C. By controlling the hot working deformation at 30%-60%, the grains of the steel can be further refined and the comprehensive mechanical properties of the steel can be improved; (In Example 1, the starting rolling temperature is 1120°C, the final rolling temperature is 950°C, and the hot working deformation is 45%, which effectively refines the grains and improves the mechanical properties;) Appropriate starting and final rolling temperatures can ensure the smooth progress of the processing process. At the same time, controlling the hot working deformation can cause the grains to break and rearrange during the deformation process, thereby achieving the purpose of refining the grains, thereby improving the strength, toughness and other mechanical properties of the steel and meeting the use requirements of the furnace roller;

[0199] When hot rolling is used, the surface roughness Ra of the roller is controlled at 0.8-1.6μm, which can improve the surface quality of the steel and reduce surface defects; (In the thermal processing step, if hot rolling is used, the control of the surface roughness of the roller can improve the contact state between the steel surface and the roller, reduce the occurrence of surface scratches, cracks and other defects, improve the surface quality of the steel, and thus improve the quality and service life of the furnace roller;)

[0200] S4, solution treatment: the hot-processed steel is subjected to solution treatment at a solution temperature of 1050°C-1150°C for 1-3 hours, followed by rapid cooling at a cooling rate greater than 50°C / s;

[0201] The solution treatment temperature is 1050°C-1150°C, the holding time is 1-3 hours, and then the steel is rapidly cooled (the cooling rate is greater than 50°C / s) to obtain a single-phase austenite structure, thereby improving the corrosion resistance and toughness of the steel; (In Example 1, the solution treatment temperature is 1100°C, and the steel is rapidly cooled after holding for 2 hours to obtain a single-phase austenite structure, thereby improving the performance of the steel;) During the solution treatment process, the alloy elements are fully dissolved in the austenite by high-temperature heating, the holding time ensures that the elements are evenly distributed, and the rapid cooling can inhibit the precipitation of the alloy elements and maintain the single-phase state of the austenite, thereby improving the corrosion resistance and toughness of the steel, and making the furnace roller have better corrosion resistance and deformation resistance in a high temperature environment;

[0202] Rapid cooling uses water quenching or oil quenching. The appropriate cooling method can be selected according to the specific requirements of the steel and production conditions to achieve the best organization and performance; (Water quenching has a fast cooling speed and can obtain finer grains, but may produce greater internal stress; Oil quenching has a relatively slow cooling speed and less internal stress, but the grains may be slightly coarser; According to the composition and performance requirements of the steel, choosing the appropriate cooling method can ensure performance while reducing defects caused by improper cooling;)

[0203] S5. Cold working: cold working the steel after solution treatment according to actual needs, and the cold working deformation is controlled within 10%-30%;

[0204] According to actual needs, the steel after solution treatment is cold processed (such as cold rolling, cold drawing), and the cold processing deformation is controlled at 10%-30%, which can improve the strength and dimensional accuracy of the steel; (cold rolling is performed in Example 1, and the cold processing deformation is 20%, which improves the strength and dimensional accuracy of the steel and meets the manufacturing requirements of the furnace roller;) Cold processing increases the strength of the steel by causing plastic deformation of the steel and increasing the dislocation density; at the same time, cold processing can further improve the surface quality and dimensional accuracy of the steel, making the manufacturing of the furnace roller more accurate, and improving its matching accuracy and performance with other components;

[0205] S6, aging treatment: finally, the cold-worked steel is subjected to aging treatment at a temperature of 550°C-650°C, a holding time of 4-8 hours, and then air-cooled; after testing, the austenitic steel prepared in this embodiment has a yield strength of 480MPa, a tensile strength of 680MPa, an elongation of 35%, and a hardness (HB) of 180, having good comprehensive properties;

[0206] Finally, the cold-worked steel is subjected to aging treatment, the aging temperature is 550°C-650°C, the holding time is 4-8 hours, and then air-cooling is performed; the aging treatment can make the precipitation phase in the steel precipitate uniformly, further improve the strength and hardness of the steel, and maintain good toughness; (in Example 1, the aging treatment temperature is 600°C, and air-cooling is performed after holding for 6 hours, so that the strength and hardness of the steel are improved, the toughness is maintained well, and the comprehensive performance of the furnace roller steel is improved;) During the aging treatment process, with the control of temperature and time, the alloy elements in the steel will precipitate in the form of precipitation phases, and these precipitation phases can hinder the dislocation movement, thereby improving the strength and hardness of the steel; at the same time, a reasonable aging treatment process will not excessively damage the toughness of the steel, so that the steel can maintain a certain toughness while having higher strength and hardness, meeting the use requirements of the furnace roller under complex working conditions.

[0207] In this embodiment: the carbon content is 0.08%, which ensures the strength while cooperating with other elements to make the steel have better comprehensive properties. If the carbon content is too low, the strength of the steel may be insufficient; if the carbon content is too high, in a high temperature environment, excessive carbide precipitation will weaken the toughness and corrosion resistance of the steel, affecting the long-term stable operation of the furnace roller under high temperature conditions; the silicon content is 1.5%, which plays a certain role in improving oxidation resistance and strength, while avoiding excessive adverse effects on toughness and welding performance. Silicon can form a silicon oxide film in steel, improve the oxidation resistance of steel, and increase strength at the same time. However, too high a silicon content will reduce the toughness of the steel and easily produce cracks during welding, etc. Defects affect the manufacturing process and performance of the furnace roller; the manganese content is 2.0%, which promotes the formation of austenite and improves the hardenability and strength of the steel. Manganese can expand the austenite phase area, making it easier for the steel to form austenite structure during the cooling process. Improving hardenability helps the steel to obtain uniform structure and performance during heat treatment, improves the overall strength, and provides better mechanical performance support for the furnace roller under complex working conditions; the chromium content is 20%, which effectively improves the oxidation resistance and corrosion resistance of the steel, making it suitable for the furnace roller to work in harsh environments such as high temperature. Chromium can form a dense chromium oxide film on the steel surface to prevent oxygen and other media from further corroding the steel matrix and improve the corrosion resistance of the steel. At the same time, this oxygen The chemical film can also maintain a certain stability at high temperatures, ensuring the oxidation resistance of the furnace roller in high temperature environments and extending its service life; the nickel content is 10%, which enhances the toughness and low-temperature performance of the steel, so that the furnace roller can maintain good performance under different temperature conditions. Nickel is an austenite-forming element, which can stabilize the austenite structure and improve the toughness of the steel, making it less likely to break when subjected to impact and other loads. In low-temperature environments, nickel can also prevent the brittle transition of steel, ensuring the safe use of the furnace roller under low-temperature conditions; the molybdenum content is 1.0%, which improves the high-temperature strength and corrosion resistance of the steel, meeting the use requirements of the furnace roller in high-temperature environments. Molybdenum can solid-solution strengthen the matrix in steel and improve the recrystallization of steel. The crystallization temperature is increased, thereby enhancing the high-temperature strength. At the same time, molybdenum can also improve the steel's resistance to pitting and intergranular corrosion, and work synergistically with elements such as chromium to further improve the steel's corrosion resistance, ensuring the stable performance of the furnace roller in a high-temperature corrosion environment; the niobium content is 0.1%, and the titanium content is 0.1%. By forming carbonitrides, the grains are refined, and the comprehensive properties of the steel are improved. Niobium and titanium combine with carbon and nitrogen in the steel to form fine carbonitride particles. These particles can hinder the growth of grains during the solidification and heat treatment of the steel, thereby refining the grains. The refined grains can improve the comprehensive properties of the steel, such as strength, toughness and plasticity, and make the furnace roller have better mechanical properties and fatigue resistance; the aluminum content is 0.05%, which plays a role in deoxidation and grain refinement, and improves the purity and organizational properties of steel. Aluminum can combine with oxygen in the steelmaking process to form inclusions such as alumina. These inclusions float up and are discharged, thereby achieving the purpose of deoxidation. At the same time, aluminum can also promote grain refinement, improve the strength and toughness of steel, and improve the quality of steel for furnace rollers; the phosphorus content is 0.02%, and the sulfur content is 0.01%. The lower phosphorus and sulfur content helps to improve the quality and performance of steel. Phosphorus and sulfur are generally regarded as harmful impurities in steel. Phosphorus can reduce the toughness and welding performance of steel, especially at low temperatures, which can easily cause brittle fracture. Sulfur can form sulfide inclusions, which reduce the hot working performance and fatigue performance of steel. Therefore, strictly controlling its content can improve the purity and comprehensive performance of steel and ensure the reliability and service life of furnace rollers. .

[0208] Specifically, the stirring mechanism includes:

[0209] Mixing tank 1;

[0210] A lower turntable 4 is rotatably connected to the bottom wall of the mixing barrel 1;

[0211] A driving shaft 2, one end of which is fixed to the top of the lower turntable 4, and the other end of which is rotatably connected to the top wall of the mixing barrel 1;

[0212] A central stirring blade 201, which is provided in plurality and is evenly fixed on the circumferential surface of the driving shaft 2;

[0213] There are two driven shafts 3, both of which are rotatably connected to the top of the lower turntable 4 and are respectively located on both sides of the driven shaft 3;

[0214] The edge stirring blades 301 are provided in plurality and are evenly fixed on the circumferential surfaces of the two driven shafts 3, and the plurality of the edge stirring blades 301 are staggered with the plurality of the central stirring blades 201;

[0215] The motor 7 is fixed to the top of the mixing barrel 1, and the output end of the motor 7 is movable through the mixing barrel 1 and fixed to the top of the driving shaft 2;

[0216] An annular rack 8 is fixed to the upper portion of the circumferential inner wall of the mixing barrel 1;

[0217] There are two gears 9 which are respectively fixed to the circumferential surfaces of the two driven shafts 3 and evenly mesh with the annular racks 8 .

[0218] Specifically, an upper turntable 5 is fixed to the upper part of the circumferential surface of the driving shaft 2, the upper turntable 5 is sleeved on the circumferential surfaces of the two driven shafts 3 and is rotatably connected to the two driven shafts 3, a supporting ring plate 6 is fixed to the circumferential inner wall of the mixing barrel 1, and the upper turntable 5 is rotatably connected to the upper part of the supporting ring plate 6.

[0219] In this embodiment: Stirring mechanism composition and function

[0220] Mixing barrel 1: provides space for accommodating raw materials for mixing. Its structural design must ensure that it can withstand the pressure and impact of materials during the mixing process. During the entire mixing process, the mixing barrel, as a container, provides a stable environment for the mixing of raw materials. Its material and structural strength must meet the requirements of long-term use and mixing force. First, multiple raw materials that need to be refined are mixed to enable better refining; the motor 7 drives the driving shaft 2 to rotate, so that the central stirring blade 201 stirs the raw materials, which plays a role in preliminary mixing and dispersing the raw materials. The motor 7 drives the driving shaft 2 to rotate, and the central stirring blade 201 rotates accordingly, stirring the raw materials in the central area of ​​the barrel, so that the raw materials have a certain movement in the vertical direction and mixing; the gear 9 on the driven shaft 3 is meshed with the annular rack 8 on the upper part of the inner wall of the mixing barrel. When the driving shaft 2 drives the lower turntable to rotate, the driven shaft 3 rotates due to the meshing action with the annular rack 8, and the edge stirring blades 301 stir the raw materials close to the barrel wall, and cooperate with the central stirring blades to achieve comprehensive and uniform stirring of the raw materials. During the stirring process, the edge stirring blades 301 can fully stir the raw materials close to the barrel wall to avoid accumulation of raw materials near the barrel wall or uneven mixing, and work together with the central stirring blades 201 to fully mix the raw materials in the entire mixing barrel 1, ensuring the uniformity of the raw materials before entering the electric furnace for smelting, thereby improving the uniformity and stability of the steel composition. ; The motor 7 provides power for the stirring mechanism, and its power and speed should be reasonably selected according to the size of the stirring barrel and the characteristics of the raw materials to ensure the stirring effect. The output end of the motor 7 is connected to the driving shaft 2, providing power for the rotation of the driving shaft 2, driving the entire stirring mechanism to operate, and is the power source for achieving raw material mixing; the annular rack 8 is fixed to the upper part of the circumferential inner wall of the stirring barrel 1, and the gear 9 is fixed to the circumferential surface of the driven shaft 3 and meshes with the annular rack 8. The self-rotation of the driven shaft is achieved through this structural design. During the stirring process, the cooperation of the annular rack 8 and the gear 9 enables the driven shaft 3 to rotate while revolving with the lower turntable 4, which increases the complexity and uniformity of stirring and improves the stirring efficiency and effect; the main An upper turntable 5 is fixed to the upper part of the circumferential surface of the driving shaft 2, and the upper turntable 5 is sleeved on the circumferential surfaces of the two driven shafts 3 and rotatably connected thereto. A supporting ring plate 6 is fixed to the circumferential inner wall of the mixing barrel 1, and the upper turntable 5 is rotatably connected to the upper part of the supporting ring plate 6. The upper turntable 5 and the supporting ring plate 5 play a role in supporting and stabilizing the driven shaft 3, so that the driven shaft 3 is more stable during rotation, reducing shaking and friction, and improving the operating stability and reliability of the stirring mechanism. When the stirring mechanism is running, the upper turntable 5 rotates on the supporting ring plate 6, providing an upper support point for the driven shaft 3, and working together with the lower turntable 4 to ensure the stable operation of the driven shaft 3, thereby ensuring that the edge stirring blades 301 can stably and effectively stir the raw materials.

[0221] Specifically, a feed hopper 101 is fixedly installed on the upper part of the circumferential surface of the mixing barrel 1, a discharge pipe 102 is fixedly installed on the bottom of the mixing barrel 1, and an arc-shaped scraper plate 10 is fixed on the circumferential surface of the lower turntable 4, and one end of the arc-shaped scraper plate 10 away from the lower turntable 4 matches the circumferential inner wall of the mixing barrel 1.

[0222] In the present embodiment: the feed hopper 101 is used for feeding, 102 is used for discharging after mixing, a valve is arranged on 102, and an arc-shaped scraper plate 10 is fixed on the circumferential surface of the lower turntable 4, and its end away from the lower turntable 4 matches the circumferential inner wall of the mixing barrel 1. During the mixing process, as the lower turntable 4 rotates, when the material is discharged through the discharge pipe 102, the arc-shaped scraper plate 10 can scrape off the raw materials attached to the bottom wall of the mixing barrel 1, so as to avoid residual and waste of raw materials, ensure the full utilization of raw materials, and also help to keep the inside of the mixing barrel clean and reduce the impact on subsequent mixing. During the mixing process, the arc-shaped scraper plate 10 continuously scrapes the raw materials on the bottom wall of the mixing barrel 1 to the discharge pipe 102, so as to ensure that the raw materials can be discharged from the discharge pipe 102 to avoid waste.

[0223] Specifically, the secondary refining outside the furnace in the smelting step adopts a ladle refining furnace or a vacuum degassing process, and the modifier added in the casting step is a rare earth element. The rare earth element is added as a modifier during the casting process. The rare earth element has a strong deoxidation and desulfurization ability, which can further reduce the oxygen and sulfur content in the steel and reduce the formation of inclusions. At the same time, the rare earth element can refine the grains and improve the cast structure of the steel. The refined grains can improve the comprehensive properties of the steel such as strength, toughness and plasticity, so that the furnace roller has better mechanical properties when bearing various loads. In addition, the rare earth element can also improve the high-temperature oxidation resistance and corrosion resistance of the steel, which is undoubtedly very beneficial for the furnace roller working in high temperature and complex environment, and can significantly improve the service life and reliability of the furnace roller.

[0224] Specifically, in the hot rolling or forging in the hot working step, when hot rolling is adopted, the surface roughness Ra of the rolling roller is controlled to be 0.8-1.6 μm.

[0225] Specifically, in the solution treatment step, rapid cooling is performed by water quenching or oil quenching, and the cold working method in the cold working step is cold rolling or cold drawing.

[0226] Specifically, the steel material made of austenitic steel has a yield strength of ≥450 MPa, a tensile strength of ≥650 MPa, an elongation of ≥30%, and a hardness (HB) of ≤200.

[0227] In this embodiment, the service life of the furnace roller made of the austenitic steel in a high temperature environment is increased by 2-3 times compared with the existing furnace roller material.

[0228] This embodiment also provides a preparation process of austenitic steel for manufacturing furnace rollers, which is different from that of embodiment 1 only in the percentage of raw materials, and the preparation method is the same, as follows:

[0229] S1, smelting: weigh 0.08% carbon, 1.5% silicon, 2.0% manganese, 20% chromium, 10% nickel, 1.0% molybdenum, 0.1% niobium, 0.1% titanium, 0.05% aluminum, 0.02% phosphorus, 0.01% sulfur, and the balance is iron (Fe) and inevitable impurities. Mix all the raw materials according to the proportion through a stirring mechanism, add them into an electric furnace for smelting after the mixing is completed, and control the smelting temperature at 1550°C-1650°C. After the raw materials are completely melted, perform refining treatment for 30-60 minutes. During the refining process, argon blowing and stirring are adopted, and then the molten steel is poured into a ladle for secondary refining outside the furnace to make the oxygen content in the molten steel ≤20ppm and the sulfur content ≤0.01%;

[0230] S2. Casting: Pour the refined molten steel into a pre-prepared casting mold, control the casting temperature at 1450°C-1550°C, and use the pouring method or continuous casting method for casting. A proper amount of modifier can be added during the casting process;

[0231] S3, hot working: heating the steel billet at a temperature of 1150°C-1250°C for 2-4 hours, and then hot rolling or forging, with the starting rolling (forging) temperature controlled at 1100°C-1150°C, the final rolling (forging) temperature not less than 900°C, and the hot working deformation controlled at 30%-60%;

[0232] S4, solution treatment: the hot-processed steel is subjected to solution treatment at a solution temperature of 1050°C-1150°C for 1-3 hours, followed by rapid cooling at a cooling rate greater than 50°C / s;

[0233] S5. Cold working: cold working the steel after solution treatment according to actual needs, and the cold working deformation is controlled within 10%-30%;

[0234] S6. Aging treatment: Finally, the cold-worked steel is subjected to aging treatment at a temperature of 550°C-650°C for 4-8 hours, followed by air cooling. According to testing, the austenitic steel prepared in this embodiment has a yield strength of 480 MPa, a tensile strength of 680 MPa, an elongation of 35%, and a hardness (HB) of 180, and has good comprehensive properties.

[0235] Example 3

[0236] Neuro-fuzzy inference system input:

[0237] Solution treatment temperature T s , the range is [1050℃,1150℃];

[0238] Insulation time t h , the range is [1,3] hours;

[0239] Cooling speed v c , range is [50,300]℃ / s;

[0240] Target output: microstructure homogeneity μ and mechanical properties P;

[0241] Output:

[0242] Optimized solution treatment temperature

[0243] Optimized holding time

[0244] Optimized cooling speed

[0245] The corresponding microstructure uniformity μ opt And mechanical properties P opt ;

[0246] ANFIS modeling: First, based on a large amount of experimental data, the ANFIS model is established, with the input being the process parameter T s ,t h ,v c , the output is the predicted microstructural uniformity and mechanical properties

[0247] The formula is as follows:

[0248]

[0249] is the predicted microstructure homogeneity value, is the predicted mechanical property value, ANFIS is the neuro-fuzzy inference system, T s is the solution treatment temperature, t h is the holding time, v c is the cooling rate;

[0250] The ANFIS model gradually optimizes the membership function by learning the complex nonlinear relationship between process parameters and output performance, so that the error between the predicted value and the actual value is minimized;

[0251] Differential Evolution (DE): To further improve the accuracy of the ANFIS model, the differential evolution algorithm is used to optimize the membership function parameters in ANFIS. Each generation generates a new combination of membership function parameters, with the goal of minimizing the error function:

[0252]

[0253] f(x) is the objective function or fitness function, N is the total number of training samples, To accumulate all samples from k = 1 to k = N, μ target To achieve microstructural uniformity, is the predicted microstructure uniformity value of the kth sample, P target For the target mechanical properties, is the predicted mechanical property value of the kth sample;

[0254] Arctic Puffin Optimization (APO): APO algorithm is further used for global optimization to solve the optimal combination of solution treatment parameters in actual production. APO finds the optimal T based on differential evolution optimization. s ,t h ,v c The parameter combination aims to make the error E μ (t) and E P (t) Convergence:

[0255]

[0256] E μ (t) is the microstructure uniformity error at time t, μ target To achieve microstructural uniformity, is the predicted uniformity of the steel microstructure at time t;

[0257] When|E μ (t)|<∈ μ and |E P (t)|<∈ P When , the system stops and outputs the optimal process parameters, such as Figure 6 As shown, the lines from top to bottom are the solution treatment temperature T s , cooling rate v c , insulation time t h , process parameters T s ,t h ,v c As the optimization algorithm iterates, it gradually converges and eventually reaches the optimal value;

[0258] By applying the DE and APO algorithms, the optimal process parameters finally determined by the system are:

[0259] Solution treatment temperature

[0260] Holding time:

[0261] Cooling rate:

[0262] The output microstructure uniformity is μ opt =0.91, mechanical properties are P opt =505MPa, both reaching or exceeding the target value. This shows that the accuracy of process parameters and the consistency of steel performance can be significantly improved through the optimization of intelligent algorithms.

[0263] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An austenitic steel for manufacturing furnace rollers, characterized in that: Its chemical composition, by mass percentage, includes: carbon 0.05%-0.12%, silicon 1.0%-2.0%, manganese 0.5%-2.5%, chromium 18%-22%, nickel 8%-12%, molybdenum 0.5%-1.5%, niobium 0.05%-0.15%, titanium 0.05%-0.15%, aluminum 0.02%-0.08%, phosphorus ≤0.03%, sulfur ≤0.03%, and the balance is iron and unavoidable impurities.

2. The austenitic steel for manufacturing furnace rollers according to claim 1, characterized in that: Its chemical composition by mass percentage includes: carbon content of 0.08%-0.1%, silicon content of 1.5%-1.8%, manganese content of 2.0%-2.2%, chromium content of 20%-21%, nickel content of 10%-11%, molybdenum content of 1.0%-1.2%, niobium content of 0.1%-0.12%, titanium content of 0.1%-0.12%, and aluminum content of 0.05%-0.06%.

3. A process for preparing austenitic steel for manufacturing furnace rollers, based on the austenitic steel for manufacturing furnace rollers as claimed in claims 1-2, characterized in that: The following steps are involved: S1, smelting: the raw materials are first mixed according to the proportion by a stirring mechanism, and then added into an electric furnace for smelting after the mixing is completed. The smelting temperature is controlled at 1550°C-1650°C. After the raw materials are completely melted, they are refined for 30-60 minutes. Argon blowing and stirring are used during the refining process. Then the molten steel is poured into a ladle for secondary refining outside the furnace to make the oxygen content in the molten steel ≤20ppm and the sulfur content ≤0.01%; S2. Casting: Pour the refined molten steel into a pre-prepared casting mold, control the casting temperature at 1450°C-1550°C, and use the pouring method or continuous casting method for casting. A proper amount of modifier can be added during the casting process; S3, hot working: heating the steel billet at a temperature of 1150°C-1250°C for 2-4 hours, and then hot rolling or forging, with the starting rolling (forging) temperature controlled at 1100°C-1150°C, the final rolling (forging) temperature not less than 900°C, and the hot working deformation controlled at 30%-60%; S4, solution treatment: the hot-processed steel is subjected to solution treatment at a solution temperature of 1050°C-1150°C for 1-3 hours, followed by rapid cooling at a cooling rate greater than 50°C / s; S5. Cold working: cold working the steel after solution treatment according to actual needs, and the cold working deformation is controlled within 10%-30%; S6. Aging treatment: Finally, the cold-processed steel is subjected to aging treatment at a temperature of 550°C-650°C, with a holding time of 4-8 hours, followed by air cooling.

4. The process for preparing austenitic steel for manufacturing furnace rollers according to claim 3, characterized in that: The stirring mechanism comprises: Mixing bucket (1); A lower turntable (4) rotatably connected to the bottom wall of the mixing barrel (1); A driving shaft (2), one end of which is fixed to the top of the lower turntable (4), and the other end of which is rotatably connected to the top wall of the mixing barrel (1); A plurality of central stirring blades (201) are provided and are evenly fixed on the circumferential surface of the driving shaft (2); Two driven shafts (3) are provided, both of which are rotatably connected to the top of the lower turntable (4) and are respectively located on both sides of the driven shaft (3); A plurality of edge stirring blades (301) are provided and are evenly fixed on the circumferential surfaces of the two driven shafts (3), and the plurality of edge stirring blades (301) and the plurality of central stirring blades (201) are staggeredly distributed; A motor (7) is fixed to the top of the mixing barrel (1), and an output end of the motor (7) is movably inserted into the mixing barrel (1) and is fixed to the top of the driving shaft (2); An annular rack (8) is fixed to the upper portion of the circumferential inner wall of the mixing barrel (1); The gears (9) are provided in two pieces and are respectively fixed to the circumferential surfaces of the two driven shafts (3) and are evenly meshed with the annular racks (8).

5. The process for preparing austenitic steel for manufacturing furnace rollers according to claim 4, characterized in that: An upper turntable (5) is fixed to the upper portion of the circumferential surface of the driving shaft (2), the upper turntable (5) is sleeved on the circumferential surfaces of the two driven shafts (3) and is rotatably connected to the two driven shafts (3), a supporting ring plate (6) is fixed to the circumferential inner wall of the mixing barrel (1), and the upper turntable (5) is rotatably connected to the upper portion of the supporting ring plate (6).

6. The process for preparing austenitic steel for manufacturing furnace rollers according to claim 5, characterized in that: A feed hopper (101) is fixedly mounted on the upper portion of the circumferential surface of the mixing barrel (1), a discharge pipe (102) is fixedly mounted on the bottom of the mixing barrel (1), and an arc-shaped scraper plate (10) is fixedly mounted on the circumferential surface of the lower turntable (4), wherein one end of the arc-shaped scraper plate (10) away from the lower turntable (4) matches the circumferential inner wall of the mixing barrel (1).

7. The process for preparing austenitic steel for manufacturing furnace rollers according to claim 6, characterized in that: The secondary refining outside the furnace in the smelting step adopts a ladle refining furnace or a vacuum degassing process, and the modifier added in the casting step is a rare earth element.

8. The process for preparing austenitic steel for manufacturing furnace rollers according to claim 7, characterized in that: In the hot rolling or forging in the hot working step, when hot rolling is adopted, the surface roughness Ra of the rolling roller is controlled at 0.8-1.6 μm, in the solution treatment step, rapid cooling is carried out by water quenching or oil quenching, and the cold working method in the cold working step is cold rolling or cold drawing.

9. The process for preparing austenitic steel for manufacturing furnace rollers according to claim 8, characterized in that: The steel material made of the austenitic steel has a yield strength of ≥450MPa, a tensile strength of ≥650MPa, an elongation of ≥30%, and a hardness (HB) of ≤200.

10. The process for preparing austenitic steel for manufacturing furnace rollers according to claim 3, characterized in that: The solution treatment step is integrated with a neural fuzzy inference system, which is used to monitor and dynamically adjust the solution temperature and holding time in real time. The input of the neural fuzzy inference system includes the solution treatment temperature T s , insulation time t h and cooling rate v c Its output is optimized by differential evolution algorithm and transmitted to the solution treatment electric heating controller and heat treatment digital timing relay. At the same time, the cooling control module receives the cooling speed adjustment signal. The optimization process uses the Arctic Sea Parrot optimization algorithm to further adjust the solution parameters to ensure the uniformity of the steel microstructure and the stability of the mechanical properties. The following steps are involved: Step 1: Data collection and preprocessing Data Collection: The following key parameters need to be collected as input to the model: Solution treatment temperature T s : Use a thermocouple temperature sensor and place the thermocouple probe in direct close contact with the workpiece to ensure the accuracy and response speed of temperature measurement and obtain the solution treatment temperature T s ; Insulation time t h : Use the heat treatment digital timing relay to record and control the duration of the holding process. Integrate the heat treatment digital timing relay into the control system. When the process starts or reaches the set temperature, the timing will automatically start and stop after the set holding time is completed. The holding time t h ; Cooling speed v c : The temperature reduction of the object per unit time is used in combination with a thermocouple temperature sensor to monitor the temperature drop rate of the workpiece surface and evaluate the cooling rate v c ; Steel microstructure changes μ: Take steel samples, perform standardized cutting, polishing and etching, observe the samples microscopically under a scanning electron microscope, obtain images of grain size, phase structure and tissue distribution, analyze the scanning electron microscope images using image processing software, calculate the grain size and phase content, and quantify the microstructural changes μ; Mechanical properties of steel P: Prepare steel samples that meet the test standards, use a tensile testing machine to apply gradually increasing tensile loads to the samples, record the load and displacement, calculate the tensile strength, yield strength and elongation parameters based on the load-displacement curve, and comprehensively obtain the mechanical properties P of the steel; These data will be used to train the neuro-fuzzy inference system, but before that, they need to be normalized to ensure that data of different scales will not have an unbalanced impact due to different units during the training process; Data normalization: The purpose of normalization is to map data of different units and magnitudes to the same range. The normalization formula is: X′ is the normalized value, X is the original data, and X min is the minimum value of the data, X max is the maximum value of the data; Normalization formula for each parameter: Normalization of solution treatment temperature: T s ′ is the normalized solution treatment temperature, T s is the actual solution treatment temperature, T s,min is the minimum solution treatment temperature, T s,max is the maximum value of solution treatment temperature; Normalization of holding time: t′ h is the normalized holding time, t h is the actual holding time, t h,min is the minimum holding time, t h,max is the maximum value of the holding time; Normalization of cooling rate: v′ c is the normalized cooling rate, v c is the actual cooling rate, v c,min is the minimum cooling rate, v c,max is the maximum value of cooling rate; Normalization of microstructural changes: μ′ is the normalized microstructure change, μ is the actual measured microstructure change value, and μ min is the minimum value of microstructure change, μ max is the maximum value of microstructural change; Normalization of mechanical properties: P′ is the normalized mechanical property value, P is the actual measured mechanical property, and P min is the minimum value of mechanical properties, P max is the maximum value of mechanical properties; Step 2: ANFIS modeling of neural fuzzy inference system Fuzzification of input variables: Each input variable is divided into several fuzzy sets through fuzzification processing. Each input variable Xi is divided into m fuzzy sets, and the fuzzy membership function is: is the membership function of the jth fuzzy set of the ith input variable, X i is the i-th input variable, c i,j is the center of the jth fuzzy set, b is the parameter that controls the shape of the membership function, a i,j To control the width of the membership function; Fuzzy rule base: Based on the fuzzy set of input variables, a fuzzy rule base is constructed, and the output variables μ and P can be expressed as linear combinations of the input variables: f1(T s ,t h ,v c )=p1T s +q1t h +r1v c +s1 f2(T s ,t h ,v c )=p2T s +q2t h +r2v c +s2 f1(T s ,t h ,v c ) is the output function of the fuzzy inference system under a certain fuzzy rule, p1 is the solution treatment temperature T s The linear coefficient, T s is the solution treatment temperature, q1 is the holding time t h The linear coefficient, t h is the holding time, r1 is the cooling rate v c The linear coefficient, v c is the cooling rate, s1 is a constant term, f2(T s ,t h ,v c ) is one of the outputs of the fuzzy rule reasoning system, p2 is the solution treatment temperature T s The linear coefficient, q2 is the holding time t h The linear coefficient, r2 is the cooling rate v c The linear coefficient of , s2 is the constant term; Reasoning and output calculation: Using the fuzzy rule base for reasoning, the weight calculation method for each fuzzy rule is: w k is the activation weight of the kth fuzzy rule, ∏ is the continuous product symbol, μ is the membership function, A i,j is the i-th input variable X i The jth fuzzy set, X i is the i-th input variable; The final output is calculated by weighted average: μ is the predicted microstructure change of steel, ∑ is the summation symbol, k is the current rule number, n is the total number of fuzzy rules, and w k is the activation weight of the kth fuzzy rule, f 1,k (T s ,t h ,v c ) is the output function of the fuzzy inference system under a certain fuzzy rule, T s is the solution treatment temperature, t h is the holding time, v c is the cooling rate; P is the mechanical property of steel, f 2,k (T s ,t h ,v c ) is the output function of the fuzzy inference system under a certain fuzzy rule; Learning capabilities of neural networks: The ANFIS system optimizes the parameters in the fuzzy rule base through the learning ability of the neural network. The optimization process is usually carried out through the error back propagation algorithm, and the goal is to minimize the error of the system; The error is defined as: E is the total error function, To sum all data samples from k = 1 to k = N, is the predicted output value of the kth sample, μ k is the true output value of the kth sample, is the predicted output value of the kth sample, P k is the true output value of the kth sample; Step 3: Differential evolution algorithm DE optimizes ANFIS model Initialize the population: In the differential evolution algorithm, the population needs to be initialized first. Each individual in the population represents a candidate solution of the ANFIS model, which contains the parameters of the fuzzy rule base and the membership function. The population size is N, and each individual represents the i-th individual in the t-th generation of the ANFIS model, which is defined as: is, t is the number of generations in the evolutionary process, i is the index number of the individual in the population, is the central parameter of the jth fuzzy rule membership function of the ith individual in the tth generation, is the width parameter of the membership function of the jth fuzzy rule of the i-th individual in the t-th generation, One of the linear regression parameters used in the ANFIS model to predict the microstructural homogeneity output is related to the solution treatment temperature. It is one of the linear regression parameters used in the ANFIS model to predict the uniformity of microstructure and is related to the holding time. It is one of the linear regression parameters used in the ANFIS model to predict the microstructure uniformity and is related to the cooling rate. is the adjustment amount of the bias term used in the ANFIS model to predict the uniformity of microstructure and is independent of the input variables. One of the linear regression parameters used in the ANFIS model to predict the mechanical property output is related to the solution treatment temperature. One of the linear regression parameters used in the ANFIS model to predict mechanical property output is related to the holding time. One of the linear regression parameters used in the ANFIS model to predict the mechanical property output is related to the cooling rate. It is the bias term used in the ANFIS model to predict the output of mechanical properties, and is an adjustment quantity that is independent of the input variables; Mutation operation: In each generation t, for each individual Generate mutation vector, mutation vector The generation method is as follows: is the mutation vector of the ith individual in the t+1th generation, is the solution of the r1th individual in the tth generation, F is the scaling factor, is the solution of the r2th individual in the tth generation, is the solution of the r3th individual in the tth generation; Crossover operation: Through the crossover operation, the mutant individuals With the original individual Combination to generate new candidate solutions The crossover formula is as follows: is the jth parameter of the i-th individual of the new candidate solution generated in the t+1th generation, is the jth parameter of the ith individual generated by mutation in the t+1th generation, is the jth parameter of the ith individual of the original individual in the tth generation, rand(0,1) is a random number generation function, and CR is the crossover probability; Select an action: The selection operation selects individuals with better performance to enter the next generation through the fitness function. The fitness function evaluates the performance of the ANFIS model based on the microstructural uniformity μ and mechanical properties P of the steel, which is defined as: is the fitness function, N is the total number of samples, To sum all data samples from k = 1 to k = N, is the predicted output value of the kth sample, μ k is the true output value of the kth sample, is the predicted output value of the kth sample, P k is the true output value of the kth sample; The selection steps are as follows: is the state of the i-th individual in the next generation, is a new individual generated by mutation and crossover operations, is the state of the i-th individual in the current generation, For newly generated individuals The fitness value of For the current individual The fitness value of Through this selection process, candidate solutions with lower individual fitness enter the next generation population, thereby continuously optimizing the fuzzy rules and membership function parameters in the ANFIS model; Step 4: Arctic Puffin Optimization Algorithm APO Global Optimization initialization: Based on the optimization of the differential evolution algorithm, the Arctic Sea Parrot optimization algorithm uses this result as a reference for the initial solution space and sets the initial population, which represents the temperature T of the solution treatment. s , insulation time t h and cooling rate v c Initial value of: X i is the parameter vector of the i-th puffin individual, is the solution treatment temperature of the ith individual, is the holding time of the ith individual, is the cooling rate of the i-th individual; Foraging strategy simulation: The foraging strategy of the Arctic Puffin simulates its process of finding food in the environment. The algorithm randomly selects different combinations of process parameters for exploration and evaluates the effects of these combinations on the microstructural uniformity and mechanical properties of the steel. Each sea parrot individual X i The objective function f(X i ) is defined as the evaluation of microstructural homogeneity and mechanical properties based on a combination of process parameters: f(X i ) is, w1 is the weight coefficient, N is the total number of samples, To sum all data samples from k = 1 to k = N, is the predicted output value of the kth sample, μ k is the true output value of the kth sample, w2 is the weight coefficient, is the predicted output value of the kth sample, P k is the true output value of the kth sample; Global search and local optimization: APO simulates global search and local exploitation through the foraging behavior of sea parrots, where each individual moves randomly in the sea to find the optimal combination of process parameters; Global search: In the early stages, the puffins will conduct a large-scale exploration to find new parameter combinations. At this stage, the solution update formula is: is the solution vector of the ith individual in the t+1th generation, is the solution vector of the ith individual in the tth generation, r is a random number between the interval [0,1], is the solution vector of the individual with the best fitness in the tth generation, F is the scaling factor, is the solution vector of the r1th individual randomly selected from the tth generation population, is the solution vector of the r2th individual randomly selected from the tth generation population; Local development: When the Puffin is close to the optimal solution, it enters the local optimization stage. Local optimization aims to fine-tune the process parameters and further improve the results. The formula for local search is: is the solution vector of the ith individual in the t+1th generation, is the solution vector of the ith individual in the tth generation, α is the shrinkage factor, is the local optimal solution in the tth generation; Convergence judgment: The convergence of the APO algorithm is judged by evaluating whether the fitness function value of the population converges to a certain threshold. If the fitness change of the optimal individual in multiple consecutive generations is less than a preset threshold ∈, the algorithm stops. The convergence judgment condition is: is the fitness value of the best individual in the t+1 generation population, is the best individual in the t+1 generation population, is the fitness value of the best individual in the t-th generation population, is the solution vector of the individual with the best fitness in the tth generation; Once the convergence condition is met, the algorithm stops and outputs the final optimal parameter combination X best , that is, the optimal solution treatment temperature, holding time and cooling rate; Step 5: Real-time feedback and optimization Real-time data collection and monitoring: During the solution treatment process, sensors monitor the following key process parameters in real time: Solution treatment temperature T s (t); Insulation time t h (t); Cooling speed v c (t); t is the real-time data at the current time point. The sensor collects real-time data and sends it as input to the ANFIS model optimized by the differential evolution algorithm for processing; Real-time feedback input to the ANFIS model: The data collected by the sensor in real time s (t),t h (t), and v c (t) is used as input to the previously DE-optimized ANFIS model, which outputs two main results: Prediction of microstructure homogeneity Mechanical properties prediction The input-output relationship of the ANFIS model can be expressed as: is the predicted uniformity of the steel microstructure at time t, is the predicted mechanical properties of steel at time t, ANFIS is an adaptive neural fuzzy inference system, T s (t) is the solution treatment temperature at time t, t h (t) is the holding time at time t, v c (t) is the cooling rate at time t; Real-time feedback and adjustment mechanism: When the predicted value output by the ANFIS model is and When the expected target value is deviated, the system will automatically adjust according to the feedback; The target microstructure uniformity is μ target and the target mechanical properties are P target , the error is expressed as: E μ (t) is the microstructure uniformity error at time t, μ target To achieve microstructural uniformity, is the predicted uniformity of the steel microstructure at time t; E P (t) is the mechanical property error at time t, P target For the target mechanical properties, is the predicted mechanical properties of steel at time t; When the error E μ (t) and E P (t) When it exceeds the set tolerance range, the system automatically adjusts the current process parameter T in combination with the Arctic Sea Parrot optimization algorithm. s (t),t h (t),v c (t), in order to reduce the error; APO is used for process parameter adjustment: Combined with the APO algorithm, the current process parameters are dynamically adjusted to ensure the stable performance of the steel. The adjusted process parameters can be expressed by the following formula: is the solution treatment temperature after dynamic adjustment at time t+1, T s (t) is the solution treatment temperature at the current time t, α1 is the error E of temperature to microstructure uniformity μ (t) Response level, E μ (t) is the microstructure uniformity error, α2 is the temperature error on mechanical properties E P (t) Response level, E P (t) is the mechanical property error; is the holding time after dynamic adjustment at time t+1, t h (t) is the holding time at the current time t, β1 is the error E of the holding time on the microstructure uniformity μ (t) Response level, E μ (t) is the microstructure uniformity error, β2 is the mechanical property error of the holding time E P (t) Response level, E P (t) is the mechanical property error; is the cooling rate after dynamic adjustment at time t+1, v c (t) is the cooling rate at the current time t, γ1 is the error E of the cooling rate to the microstructure uniformity μ (t) Response level, E μ (t) is the microstructure uniformity error, γ2 is the cooling rate error on mechanical properties E P (t) Response level, E P (t) is the mechanical property error; Continuous optimization and data update: As time goes by, the system will continue to obtain real-time data from sensors and dynamically update process parameters. Due to the adaptive nature of the APO algorithm, the system will continuously perform global and local optimization to keep the process parameters moving in the optimal direction. At each time step t, the new process parameters The data are input into the system and fed back into the ANFIS model for re-prediction. This closed-loop control system ensures that the process is always optimized. Convergence conditions: The real-time optimization system will continue to adjust the process parameters until the error E μ (t) and E P (t) converges to the preset threshold ∈ μ and ∈ P ,Right now: |E μ (t)|<∈ μ |E P (t)|<∈ P E μ (t) is the microstructure uniformity error, ∈ μ is the error convergence threshold of microstructure uniformity, E P (t) is the mechanical property error, ∈ P is the error convergence threshold of mechanical properties; Once the error meets the convergence conditions, the system enters a stable state and the process parameters also stabilize, ensuring that the microstructure and mechanical properties of the steel are optimal.

Images