Traction sheave iron mold sand-coated casting process optimization method and system

Abstract

The present invention discloses a method and system for optimizing the iron mold sand-coated casting process of a traction wheel, including: constructing a target optimization function according to the hardness difference of the traction wheel groove; obtaining the thickness and material of the casting, using the improved fireworks algorithm to iteratively solve the optimization parameters of the traction wheel iron mold sand-coated casting process, and obtaining target data; the improved fireworks algorithm uses the Dropout algorithm in deep learning to screen the explosion radius of fireworks on the basis of the fireworks algorithm; and performing the traction wheel iron mold sand-coated casting according to the obtained target data. An optimization target function of the iron mold sand-coated casting process is proposed, and then the fireworks optimization algorithm is improved in a targeted manner according to the characteristics of the iron mold sand-coated casting process, and an intelligent iron mold sand-coated casting process optimization method is obtained, which can improve the casting quality of the elevator traction wheel.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent manufacturing, and the present invention relates to a method and system for optimizing a traction wheel iron mold sand-coated casting process. Background Art

[0002] In the traction system of the elevator, the traction sheave is the core component and plays a vital role in the stable and safe operation of the elevator. In order to complete the traction work normally, the wheel groove surface of the elevator traction sheave must meet a certain hardness standard and have good comprehensive mechanical properties. In the production process of the elevator traction sheave, it is usually necessary to go through casting, mechanical cutting and other process flows. In the traditional casting method, the control of process parameters mostly depends on human experience, and it is easy to have problems such as the traction sheave quality not meeting the standards due to the failure to control the forming process parameters during the manufacturing process. At present, the iron mold sand-coated casting process combines the advantages of sand casting and coated sand shell casting process. The modeling is convenient and fast, and the consistency of the sand density and the surface hardness of the casting is good. It can effectively improve the comprehensive strength of the casting, so it is more and more widely used in the casting of elevator traction sheaves. In the sand-coated iron mold casting process, due to the influence of factors such as the forced cooling effect of the iron mold on the molten iron and the shrinkage feeding riser effect of the pouring system, the thickness of the sand coating layer and the pouring speed of the molten iron jointly affect the hardness and other properties of the finished product of the elevator traction sheave. However, there is currently no optimization method for the iron mold sand-coated casting process. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for optimizing the iron mold sand-coated casting process of a traction sheave, propose an optimization objective function for the iron mold sand-coated casting process, and then improve the fireworks optimization algorithm in a targeted manner according to the characteristics of the iron mold sand-coated casting process, thereby obtaining an intelligent iron mold sand-coated casting process optimization method, which can improve the casting quality of the elevator traction sheave.

[0004] The technical solution to achieve the purpose of the present invention is:

[0005] A method for optimizing a traction sheave iron mold sand-coated casting process comprises the following steps:

[0006] S01: construct the target optimization function according to the hardness difference of the traction sheave groove;

[0007] S02: Obtain the thickness and material of the casting, use the improved fireworks algorithm to iteratively solve the optimization parameters of the traction wheel iron mold sand-coated casting process, and obtain the target data;

[0008] The improved fireworks algorithm uses the Dropout algorithm in deep learning to screen the explosion radius of fireworks based on the fireworks algorithm;

[0009] S03: Perform sand-coated casting of the traction sheave iron mold according to the obtained target data.

[0010] In the preferred technical solution, the target optimization function is:

[0011]

[0012] in, The hardness of the traction sheave groove is poor. is the thickness of the sand layer, is the pouring speed of molten iron, , is the control coefficient, is the wall thickness parameter of the casting, is the correlation coefficient of the molten iron pouring speed, is the average wall thickness of the casting, is the number of different structures in the casting cavity, It is the cavity structure complexity parameter.

[0013] In the preferred technical solution, the thickness of the sand covering layer is :

[0014]

[0015] in, is the wall thickness of the casting, , is the control coefficient, is the wall thickness parameter of the casting.

[0016] In the preferred technical solution, the molten iron pouring speed is :

[0017]

[0018] in, is the correlation coefficient of the molten iron pouring speed, is the average wall thickness of the casting, is the total mass of the casting.

[0019] In the preferred technical solution, the hardness of the traction sheave groove is poor :

[0020]

[0021] in, is the natural logarithm constant.

[0022] In the preferred technical solution, in step S02, the formula for screening the explosion radius of fireworks using the Dropout algorithm in deep learning is:

[0023]

[0024] in, is the explosion radius of the fireworks, is the minimum value of the fireworks explosion radius. is the probability of discarding an individual, is the discard probability threshold set, that is, when the explosion radius of a firework is smaller than When The probability of size is retained, and its radius is set to .

[0025] In the preferred technical solution, the improved fireworks algorithm in step S02 also includes introducing adaptive weights on the basis of Gaussian variant sparks to accelerate the convergence of the algorithm during the update process of individual fireworks, which is expressed as:

[0026]

[0027] in, Represents a firework No. Dimensional parameters, represents a Gaussian distribution with a mean of 1 and a variance of 1. represents the Gaussian mutation spark obtained after the mutation operation, is the adaptive weight parameter, is the random disturbance factor.

[0028] In the preferred technical solution, the adaptive weight parameter for:

[0029]

[0030] in, is the average explosion radius of the current group, is the number of iterations.

[0031] The present invention also discloses a traction sheave iron mold sand-coated casting process optimization system, comprising:

[0032] A target optimization function building module is used to build a target optimization function according to the hardness difference of the traction sheave groove;

[0033] Iterative optimization module: obtains the thickness and material of the casting, uses the improved fireworks algorithm to iteratively solve the optimization parameters of the traction wheel iron mold sand-coated casting process, and obtains the target data;

[0034] The improved fireworks algorithm uses the Dropout algorithm in deep learning to screen the explosion radius of fireworks based on the fireworks algorithm;

[0035] The traction sheave iron mold sand-coated casting module performs traction sheave iron mold sand-coated casting according to the obtained target data.

[0036] The present invention further discloses a computer storage medium on which a computer program is stored. When the computer program is executed, the above-mentioned method for optimizing the traction wheel iron mold sand-coated casting process is implemented.

[0037] Compared with the prior art, the present invention has the following significant advantages:

[0038] The present invention proposes an optimization objective function for the iron mold sand-coated casting process, and then improves the fireworks optimization algorithm in a targeted manner according to the characteristics of the iron mold sand-coated casting process, thereby obtaining an intelligent iron mold sand-coated casting process optimization method, which can improve the casting quality of the elevator traction wheel and has a broad application market space and economic value. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of the method for optimizing the traction sheave iron mold sand-coated casting process of this embodiment;

[0040] Figure 2 This is a comparison chart of the process structures of ordinary sand casting and iron mold sand casting;

[0041] Figure 3 It is a schematic diagram of the numerical relationship between the thickness of the sand coating and the wall thickness of the casting;

[0042] Figure 4 Execute the flow chart for the fireworks algorithm;

[0043] Figure 5 This is a comparison chart between the algorithm of the present invention and other optimization algorithms. DETAILED DESCRIPTION

[0044] The principle of the present invention is: an optimization objective function of the traction sheave iron mold sand-coated casting process is proposed, and then according to the characteristics of the iron mold sand-coated casting process, the fireworks optimization algorithm is improved in a targeted manner, and an intelligent iron mold sand-coated casting process optimization method is obtained, which can improve the casting quality of the elevator traction sheave.

[0045] Embodiment 1:

[0046] like Figure 1 As shown, a method for optimizing the traction sheave iron mold sand-coated casting process comprises the following steps:

[0047] S01: construct the target optimization function according to the hardness difference of the traction sheave groove;

[0048] S02: Obtain the thickness and material of the casting, use the improved fireworks algorithm to iteratively solve the optimization parameters of the traction wheel iron mold sand-coated casting process, and obtain the target data;

[0049] The improved fireworks algorithm uses the Dropout algorithm in deep learning to screen the explosion radius of fireworks based on the fireworks algorithm;

[0050] S03: Perform sand-coated casting of the traction sheave iron mold according to the obtained target data.

[0051] In a preferred embodiment, the objective optimization function is:

[0052]

[0053] in, The hardness of the traction sheave groove is poor. is the thickness of the sand layer, is the pouring speed of molten iron, , is the control coefficient, is the wall thickness parameter of the casting, is the correlation coefficient of the molten iron pouring speed, is the average wall thickness of the casting, is the number of different structures in the casting cavity, It is the cavity structure complexity parameter.

[0054] In a preferred embodiment, the thickness of the sand covering layer is :

[0055]

[0056] in, is the wall thickness of the casting, , is the control coefficient, is the wall thickness parameter of the casting.

[0057] In the preferred technical solution, the molten iron pouring speed is :

[0058]

[0059] in, is the correlation coefficient of the molten iron pouring speed, is the average wall thickness of the casting, is the total mass of the casting.

[0060] In a preferred embodiment, the traction sheave groove has a hardness difference :

[0061]

[0062] in, is the natural logarithm constant.

[0063] In a preferred embodiment, in step S02, the formula for screening the explosion radius of fireworks using the Dropout algorithm in deep learning is:

[0064]

[0065] in, is the explosion radius of the fireworks, is the minimum value of the fireworks explosion radius. is the probability of discarding an individual, is the discard probability threshold set, that is, when the explosion radius of a firework is smaller than When The probability of size is retained, and its radius is set to .

[0066] In a preferred embodiment, the improved fireworks algorithm in step S02 further includes introducing adaptive weights based on Gaussian variation sparks to accelerate the convergence of the algorithm during the update of individual fireworks, which is expressed as:

[0067]

[0068] in, Represents a firework No. Dimensional parameters, represents a Gaussian distribution with a mean of 1 and a variance of 1. represents the Gaussian mutation spark obtained after the mutation operation, is the adaptive weight parameter, is the random disturbance factor.

[0069] In a preferred embodiment, the adaptive weight parameter for:

[0070]

[0071] in, is the average explosion radius of the current group, is the number of iterations.

[0072] In another embodiment, a computer storage medium stores a computer program, which, when executed, implements the above-mentioned method for optimizing the traction sheave iron mold sand-coated casting process.

[0073] The specific optimization method adopts the above-mentioned traction sheave iron mold sand-coated casting process optimization method, which will not be repeated here.

[0074] In another embodiment, a traction sheave iron mold sand-coated casting process optimization system includes:

[0075] A target optimization function building module is used to build a target optimization function according to the hardness difference of the traction sheave groove;

[0076] Iterative optimization module: obtains the thickness and material of the casting, uses the improved fireworks algorithm to iteratively solve the optimization parameters of the traction wheel iron mold sand-coated casting process, and obtains the target data;

[0077] The improved fireworks algorithm uses the Dropout algorithm in deep learning to screen the explosion radius of fireworks based on the fireworks algorithm;

[0078] The traction sheave iron mold sand-coated casting module performs traction sheave iron mold sand-coated casting according to the obtained target data.

[0079] Specifically, the workflow of the traction sheave iron mold sand-coated casting process optimization system is described below by taking a preferred embodiment as an example:

[0080] The specific steps include:

[0081] Step 1: Establish the optimization function of the elevator traction wheel iron mold sand-coated casting process objectives.

[0082] The traction sheave is a key casting on the elevator traction machine. It must have high strength, good toughness, wear resistance, impact resistance and other properties. Its quality directly affects the safety performance and service life of the traction machine. The iron mold sand-coated casting process combines the advantages of sand casting and coated sand shell casting processes. It is convenient and fast to shape, and the consistency of sand density and casting surface hardness is good. It can effectively improve the comprehensive strength of the casting. Therefore, it is more and more widely used in elevator traction sheave casting. Figure 2 In the sand-coated iron mold casting process, due to the forced cooling effect of the iron mold on the molten iron and the shrinkage-feeding riser effect of the pouring system, the thickness of the sand layer and the pouring speed of the molten iron jointly affect the hardness and other properties of the finished elevator traction sheave casting.

[0083] In order to realize intelligent optimization of elevator traction sheave sand-coated iron mold casting, the present invention starts with the traction sheave groove hardness difference index, establishes a target optimization function, and guides the traction sheave sand-coated iron mold casting process.

[0084] First, based on the long-term experimental results, it is concluded that the thickness of the sand coating directly affects the cooling rate of the casting, and thus affects the surface strength of the casting. Assume that the thickness of the sand coating is , and its function graph with the casting wall thickness is as follows Figure 3 As shown. Figure 3 It can be found that the numerical relationship between the two presents an inverted U shape, which can be fitted using formula (10), where is the natural logarithm constant, is the wall thickness of the casting, , is the control coefficient, is the wall thickness parameter of the casting.

[0085]

[0086] The pouring speed of molten iron is closely related to the wall thickness, structure and hardness of the finished product of the casting. According to the quality standard requirements of the elevator traction sheave, the pouring speed of molten iron is The relationship between the wall thickness and the quality of the traction sheave casting is shown in formula (11), where is the correlation coefficient of the molten iron pouring speed, is the average wall thickness of the casting, is the total mass of the casting.

[0087]

[0088] According to national standards, the hardness difference of the traction wheel rope groove surface should be less than 15HB. Suppose the hardness difference of the traction wheel rope groove surface is , which is related to the thickness of the sand layer and molten iron pouring speed The relationship between can be described by the following formula, where is the natural logarithm constant.

[0089]

[0090] The present invention achieves intelligent optimization of the traction wheel sand-coated iron mold casting process by minimizing the hardness difference of the traction wheel rope surface. Therefore, the target optimization function is defined as shown in formula (13), where is the thickness of the sand layer, is the pouring speed of molten iron, , is the control coefficient, is the wall thickness parameter of the casting, is the correlation coefficient of the molten iron pouring speed, is the average wall thickness of the casting, is the number of different structures in the casting cavity, It is the cavity structure complexity parameter.

[0091]

[0092] It can be seen from formula (13) that the optimization of the traction wheel sand-coated iron mold casting process is a multi-parameter optimization problem, and the parameters are continuous parameters. There is currently no optimal solution under polynomial time complexity for this type of problem, and a relatively optimal solution is often solved by intelligent algorithms. The present invention uses an improved fireworks algorithm to complete the optimization solution of the traction wheel sand-coated iron mold casting process.

[0093] Step 2: Improve the fireworks algorithm.

[0094] The fireworks algorithm has a different search mechanism compared to the genetic algorithm and particle swarm optimization algorithm. It is mainly used to solve optimization problems in continuous space. During the iteration process, each fireworks individual in the population will produce multiple individuals in one iteration, while the particle swarm optimization algorithm usually only produces one individual. The explosion mechanism of the fireworks algorithm makes its search for the area near the fireworks more thorough, so it is more suitable for scenarios such as process optimization that solve problems in continuous space.

[0095] The fireworks algorithm simulates fireworks exploding in the air to form explosion sparks randomly scattered around, and regards this process as the optimization process of particles in the solution space. The algorithm is mainly composed of explosion intensity, mapping rules, selection strategy, mutation operator, and displacement operation. Its execution process is as follows Figure 4 shown.

[0096] In the fireworks algorithm, the explosion operator, mutation operator and selection strategy are the three most important components, which directly determine the performance of the fireworks algorithm. In the fireworks algorithm, the explosion radius and the number of sparks produced by each firework are closely related to its fitness value relative to other fireworks in the population. The calculation formula for the explosion radius and the number of explosion sparks of a firework is shown in the following formula, where For the index of fireworks, , They represent the explosion radius and number of explosion sparks of the fireworks respectively. is the fireworks explosion radius adjustment factor, is the adjustment factor for the number of fireworks explosion sparks, , are constants, It is the penalty coefficient, which is used to avoid division by zero errors during the algorithm calculation process. represents the target optimization function, represents a firework in the population, , They are respectively the minimum and maximum values ​​of fitness in the current fireworks population.

[0097]

[0098] In order to increase the diversity of the explosion spark population, the fireworks algorithm introduces a mutation operator to generate mutation sparks, namely Gaussian mutation sparks. The process of generating Gaussian mutation sparks is shown in the following formula, where Represents a firework No. Dimensional parameters, represents a Gaussian distribution with a mean of 1 and a variance of 1. Represents the Gaussian mutation spark obtained after the mutation operation.

[0099]

[0100] In order to pass the better information in the fireworks population to the next generation population, after the explosion sparks and Gaussian mutation sparks are generated, the algorithm will select a certain number of individuals as the next generation of fireworks. Assume that the candidate set is , fireworks individual Probability of being selected The calculation formula is as follows, where , All are indexes of fireworks.

[0101]

[0102] The above formula shows that in the candidate set, if the regional population density is high, that is, there are many other candidate individuals around an individual, the probability of the individual being selected will decrease.

[0103] In order to make the fireworks algorithm more suitable for the optimization of the traction wheel sand-coated casting process, the present invention makes the following improvements to the fireworks algorithm.

[0104] First, according to formula (14), if the explosion radius of fireworks is very small, then these fireworks contribute very little to the search process of the solution space, thus resulting in a waste of computing resources, and directly discarding the fireworks may result in an incomplete search of the solution space. To overcome this problem, the present invention uses the Dropout algorithm in deep learning to screen the explosion radius of fireworks, as shown in formula (17), where is the explosion radius of the fireworks, is the minimum value of the fireworks explosion radius. is the probability of discarding an individual, is the set discard probability threshold. That is, when the explosion radius of a firework is smaller than When The probability of size is retained, and its radius is set to .

[0105]

[0106] At the same time, in order to accelerate the convergence of the algorithm, the present invention introduces adaptive weights based on formula (15) during the update of individual fireworks to better control the calculation process of the algorithm, as shown in formula (18). is the adaptive weight parameter, is a random perturbation factor used to increase randomness during the iteration process.

[0107]

[0108] is a quantity related to the number of iterations and the explosion radius of the fireworks, and its calculation is shown in the following formula, where is the average explosion radius of the current group, is the number of iterations.

[0109]

[0110] The process of optimizing the elevator traction wheel sand coating casting process by using the improved fireworks algorithm in the present invention is as follows: first, the thickness, material and other data of the casting are obtained; then, the fireworks algorithm is used to iteratively solve various optimization parameters of the process; finally, target data such as the thickness of the sand coating layer and the pouring speed of molten iron are calculated according to the obtained parameters; finally, the casting of the elevator traction wheel is completed according to these data.

[0111] The improved fireworks algorithm was experimented and compared with the particle swarm algorithm and the original fireworks algorithm. The results are as follows: Figure 5 As shown, it can be seen that the improved fireworks algorithm proposed in the present invention has a faster convergence speed and can effectively prevent the algorithm from falling into a local optimal solution, thereby obtaining a better result, and thus can effectively improve the quality of the finished product of the elevator traction wheel casting.

[0112] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications made without departing from the spirit and principles of the present invention shall be equivalent replacement modes and shall be included in the protection scope of the present invention.

Claims

1. A method for optimizing the traction sheave iron mold sand-coated casting process, characterized in that: The following steps are involved: S01: Construct a target optimization function based on the hardness difference of the traction sheave groove; the target optimization function is: ； in, The hardness of the traction sheave groove is poor. is the thickness of the sand layer, is the pouring speed of molten iron, , is the control coefficient, is the wall thickness parameter of the casting, is the correlation coefficient of the molten iron pouring speed, is the average wall thickness of the casting, is the number of different structures in the casting cavity, is the cavity structure complexity parameter; S02: Obtain the thickness and material of the casting, use the improved fireworks algorithm to iteratively solve the optimization parameters of the traction wheel iron mold sand-coated casting process, and obtain the target data; The improved fireworks algorithm uses the Dropout algorithm in deep learning to screen the explosion radius of fireworks based on the fireworks algorithm; S03: Perform sand-coated casting of the traction sheave iron mold according to the obtained target data.

2. The method for optimizing the traction sheave iron mold sand-coated casting process according to claim 1, characterized in that: Thickness of sand layer : ； in, is the wall thickness of the casting, , is the control coefficient, is the wall thickness parameter of the casting.

3. The traction sheave iron mold sand-coated casting process optimization method according to claim 1 is characterized in that: Liquid iron pouring speed : ； in, is the correlation coefficient of the molten iron pouring speed, is the average wall thickness of the casting, is the total mass of the casting.

4. The method for optimizing the traction sheave iron mold sand-coated casting process according to claim 1, characterized in that: The hardness of the traction sheave groove is poor : ； in, is the natural logarithm constant.

5. The method for optimizing the traction sheave iron mold sand-coated casting process according to claim 1, characterized in that: In step S02, the formula for screening the explosion radius of fireworks using the Dropout algorithm in deep learning is: ； in, is the explosion radius of the fireworks, is the minimum value of the fireworks explosion radius. is the probability of discarding an individual, is the discard probability threshold set, that is, when the explosion radius of a firework is smaller than When The probability of size is retained, and its radius is set to .

6. The method for optimizing the traction sheave iron mold sand-coated casting process according to claim 1, characterized in that: The improved fireworks algorithm in step S02 also includes introducing adaptive weights on the basis of Gaussian variation sparks to accelerate the convergence of the algorithm during the update process of individual fireworks, which is expressed as: ； in, Represents a firework No. Dimensional parameters, represents a Gaussian distribution with a mean of 1 and a variance of 1. represents the Gaussian mutation spark obtained after the mutation operation, is the adaptive weight parameter, is the random disturbance factor.

7. The method for optimizing the traction sheave iron mold sand-coated casting process according to claim 6, characterized in that: Adaptive weight parameters for: ； in, is the average explosion radius of the current group, is the number of iterations.

8. A traction sheave iron mold sand-coated casting process optimization system, characterized in that: include: A target optimization function building module is used to build a target optimization function according to the hardness difference of the traction sheave groove; The objective optimization function is: ； in, The hardness of the traction sheave groove is poor. is the thickness of the sand layer, is the pouring speed of molten iron, , is the control coefficient, is the wall thickness parameter of the casting, is the correlation coefficient of the molten iron pouring speed, is the average wall thickness of the casting, is the number of different structures in the casting cavity, is the cavity structure complexity parameter; Iterative optimization module: obtains the thickness and material of the casting, uses the improved fireworks algorithm to iteratively solve the optimization parameters of the traction wheel iron mold sand-coated casting process, and obtains the target data; The improved fireworks algorithm uses the Dropout algorithm in deep learning to screen the explosion radius of fireworks based on the fireworks algorithm; The traction sheave iron mold sand-coated casting module performs traction sheave iron mold sand-coated casting according to the obtained target data.

9. A computer storage medium having a computer program stored thereon, characterized in that: When the computer program is executed, the method for optimizing the traction sheave iron mold sand-coated casting process as described in any one of claims 1 to 7 is implemented.

Images