Intelligent manufacturing control method for adaptive interval type-2 fuzzy logic enhancement

Abstract

The invention provides a self-adaptive interval type-2 fuzzy logic enhanced intelligent manufacturing control method, which comprises the steps that a fuzzy set influencing a target system is predicted through a particle swarm optimization algorithm, the fuzzy set comprises one or more dimensions, and each dimension determines a numerical range capable of generating corresponding influence through an upper membership function and a lower membership function; activating a corresponding fuzzy reasoning rule in response to the obtained fuzzy set output by the particle swarm optimization algorithm, performing fuzzy reasoning based on upper and lower membership functions in the fuzzy set, and obtaining a reasoning set; and performing defuzzification on the reasoning set to obtain a control instruction, and controlling the target system to perform intelligent manufacturing based on the control instruction. According to the method, the capability of representing and processing uncertainty in the intelligent manufacturing control process can be improved, so that a more accurate decision instruction is provided to improve the product quality and the production performance.

Description

Technical Field

[0001] The present invention relates to the field of intelligent manufacturing, and in particular to an intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic. Background Art

[0002] With the intensification of global competition and the diversification of customer needs, intelligent manufacturing, as a key technology to improve product quality, has received widespread attention. In the existing intelligent manufacturing system, although various advanced automation technologies and information technologies have been adopted, there are still limitations in dealing with the uncertainty and dynamic changes in the production process.

[0003] Traditional manufacturing systems often rely on deterministic models and static parameter settings, which are difficult to adapt to various uncertainties in the production process. These limitations lead to the inability to continuously optimize production efficiency and product quality, while also increasing production costs and resource waste.

[0004] In recent years, fuzzy logic control has been introduced into the field of intelligent manufacturing due to its advantages in dealing with uncertainty and ambiguity. However, most existing fuzzy logic control systems adopt single-point fuzzy logic based on traditional fuzzy sets, which have limited capabilities in representing and dealing with uncertainty. Summary of the invention

[0005] In order to improve the ability to represent and process uncertainty in the intelligent manufacturing control process, the present invention proposes an intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic, the method comprising: predicting a fuzzy set that affects a target system by a particle swarm optimization algorithm, the fuzzy set comprising one or more dimensions, each dimension having a numerical range that can produce a corresponding impact determined by an upper membership function and a lower membership function; activating corresponding fuzzy inference rules in response to obtaining a fuzzy set output by the particle swarm optimization algorithm, performing fuzzy inference based on the upper and lower membership functions in the fuzzy set, and obtaining an inference set; defuzzifying the inference set to obtain a control instruction, and controlling the target system to perform intelligent manufacturing based on the control instruction.

[0006] In one or more embodiments, a fuzzy set affecting a target system is predicted by a particle swarm optimization algorithm, including: constructing a multidimensional search space according to controllable parameters of the target system; constructing an upper membership function and a lower membership function for determining a particle position range for each dimension in the multidimensional search space, and initializing a particle cluster based on the upper membership function and the lower membership function; evaluating the influence of the membership function in each particle on the system according to a preset fitness function, and determining the individual optimal fitness and the global optimal fitness respectively; updating the position and velocity of each particle according to the individual optimal fitness and the global optimal fitness to adjust the membership function in each particle, and outputting a fuzzy set formed by the membership function in each particle.

[0007] In one or more embodiments, the particle cluster is initialized based on the upper membership function and the lower membership function, including: constructing multidimensional particles according to the center position and width in the upper membership function and the lower membership function corresponding to each dimension respectively; randomly assigning the center position of each particle in each dimension of the search space to randomly assign the initial position of the particle.

[0008] In one or more embodiments, evaluating the influence of the membership function in each particle on the system according to a preset fitness function includes: The preset fitness function is determined as: , in, represents the control error of the corresponding controllable parameter; The influence of the membership function in each particle on the system is evaluated according to the above fitness function.

[0009] In one or more embodiments, the method further includes: adjusting the fitness function according to an actual feedback result of the system to the control instruction.

[0010] In one or more embodiments, the method further includes determining the individual optimal fitness and the global optimal fitness of the particle by the following steps: if the current fitness of the particle is better than its individual historical optimal fitness, updating the individual optimal fitness according to the current fitness of the particle; if the individual optimal fitness of the particle is better than the global optimal fitness, updating the global optimal fitness of the particle according to the individual optimal fitness of the particle.

[0011] In one or more embodiments, updating the position and velocity of each particle according to the individual optimal fitness and the global optimal fitness to adjust the membership function in each particle includes: Determine the particle position update formula as ; The particle velocity update formula is determined as ; in, represents the position of particle i in the last t-th iteration, represents the speed of particle i in the current t+1th iteration process, is the inertia weight, c1 and c2 are learning factors, is the individual optimal fitness of particle i, is the global optimal fitness.

[0012] In one or more embodiments, in response to obtaining a fuzzy set output by the particle swarm optimization algorithm, the corresponding fuzzy inference rules are activated and fuzzy inference is performed based on the upper and lower membership functions in the fuzzy set to obtain an inference set, including: activating the corresponding fuzzy inference rules according to the influence of each particle on the system; and selecting or performing logical operations on the membership functions in the particles according to the fuzzy inference rules using a preset rule algorithm to obtain inference results to form an inference set.

[0013] In one or more embodiments, defuzzifying the inference set to obtain the control instruction includes: using a weighted average method to convert the aggregated inference results into a clear control instruction.

[0014] The beneficial effects of the present invention include: the present invention proposes to define the membership function of the fuzzy set with interval type-2 fuzzy values ​​to more effectively express and process uncertainty and ambiguity, and uses particle swarm optimization (PSO) to solve the uncertainty problem, thereby improving the ability to represent and process uncertainty in the intelligent manufacturing control process, thereby providing more accurate decision-making instructions to improve product quality and production performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other embodiments can be obtained based on these drawings without paying creative work.

[0016] Figure 1 This is a workflow diagram of an intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic according to an embodiment of the present invention; Figure 2 A flowchart of data collection and preprocessing according to an embodiment of the present invention; Figure 3 The following is a flowchart of the fuzzy processing process according to an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the embodiments of the present invention are further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0018] It should be noted that all expressions using "first" and "second" in the embodiments of the present invention are for distinguishing two non-identical entities with the same name or non-identical parameters. It can be seen that "first" and "second" are only for the convenience of expression and should not be understood as limitations on the embodiments of the present invention. The subsequent embodiments will not explain this one by one.

[0019] In order to improve the ability to represent and process uncertainty in the intelligent manufacturing control process, thereby providing more accurate decision-making instructions to improve product quality and production performance, the present invention proposes to define the membership function of the fuzzy set with interval type-2 fuzzy values ​​to more effectively express and process uncertainty and fuzziness, and use particle swarm optimization (PSO) to calculate and solve the uncertainty problem. Specifically, the present invention proposes an adaptive interval type-2 fuzzy logic enhanced intelligent manufacturing control method, characterized in that the method includes: step S1, predicting the fuzzy set that affects the target system through a particle swarm optimization algorithm, the fuzzy set contains one or more dimensions, and each dimension is determined by an upper membership function and a lower membership function to determine the numerical range that can produce a corresponding impact; step S2, in response to obtaining the fuzzy set output by the particle swarm optimization algorithm, activating the corresponding fuzzy inference rule and performing fuzzy inference based on the upper and lower membership functions in the fuzzy set and obtaining an inference set; step S3, defuzzifying the inference set to obtain a control instruction, and controlling the target system for intelligent manufacturing based on the control instruction.

[0020] In one embodiment, a fuzzy set affecting a target system is predicted by a particle swarm optimization algorithm, including: constructing a multidimensional search space according to controllable parameters of the target system; constructing an upper membership function and a lower membership function for determining a particle position range for each dimension in the multidimensional search space, and initializing a particle cluster based on the upper membership function and the lower membership function; evaluating the influence of the membership function in each particle on the system according to a preset fitness function, and determining the individual optimal fitness and the global optimal fitness respectively; updating the position and speed of each particle according to the individual optimal fitness and the global optimal fitness to adjust the membership function in each particle, and outputting a fuzzy set formed by the membership function in each particle.

[0021] Specifically, data collection and preprocessing are key components of the intelligent manufacturing process optimization system, which are responsible for real-time collection of key parameters on the production line and necessary preprocessing to facilitate subsequent fuzzy logic reasoning and optimization algorithm processing; therefore, before executing step S1 and constructing a multidimensional search space based on the controllable parameters of the target system, the following data collection and preprocessing work needs to be performed, including: (1) Main components Sensor integration unit: Integrates various types of sensors, including but not limited to temperature sensors, pressure sensors, humidity sensors, speed sensors, etc., for real-time monitoring of the status of the production line.

[0022] Data acquisition interface: A variety of interfaces (such as RS-232, RS-485, Ethernet, wireless communication, etc.) are designed to meet the data output requirements of different sensors.

[0023] Data synchronization and time stamping: ensure the time consistency of data collection and provide an accurate time benchmark for data analysis.

[0024] Data caching and management unit: Temporarily stores collected data to prevent data loss and supports further processing and management of data.

[0025] (2) Workflow (e.g. Figure 2 shown) Real-time data acquisition: The sensor integrated unit monitors various parameters on the production line in real time and converts the detected signals into electrical or digital signals; Signal conditioning: amplify, filter, isolate and process the collected original signal to eliminate noise and interference and improve signal quality; Analog-to-digital conversion: converting analog signals into digital signals for easy processing and analysis by digital systems; Data standardization: Use the following formula to standardize the collected data:

[0026] Where x is the original data, xmin and xmax are the minimum and maximum values ​​of the predefined data range respectively; Outlier detection and elimination: Detect outliers in the data and eliminate or correct them to ensure data accuracy; Data fuzzification: convert the standardized data into fuzzy values ​​and determine their membership in each fuzzy set; Data output: Output the preprocessed data to the fuzzy logic reasoning mechanism to provide input for subsequent decision-making.

[0027] Among them, the collected data is used to train particle clusters and inference rules, so as to realize the training of particle clusters and fuzzy inference rules with real production data; secondly, it is used to input the collected data into the fuzzy set model and the inference model in real time after the fuzzy set model and the inference model are trained, so as to generate control instructions in real time to adjust the production process according to various production conditions, requirements or environmental changes.

[0028] In one embodiment, initializing a particle cluster based on an upper membership function and a lower membership function includes: constructing multidimensional particles according to the center position and width in the upper membership function and the lower membership function corresponding to each dimension respectively; and randomly assigning the center position of each particle in each dimension of the search space to randomly assign the initial position of the particle.

[0029] Specifically, each particle can be represented as a multidimensional vector, where each dimension corresponds to a key parameter in the upper membership function (UMF) and lower membership function (LMF) of an input or output variable in the collected data. The particle can be represented as a parameter vector: ; in, c i_u and c i_l are the center points of the upper and lower membership functions of the i-th input variable, respectively. w i is the width; wherein the value range of the center point and the width of the membership function can be determined according to the numerical range of the corresponding data collected, or determined according to experience.

[0030] In one embodiment, the influence of the membership function in each particle on the system is evaluated according to a preset fitness function, including: The preset fitness function is determined as: , Among them, represents the control error of the corresponding controllable parameter; The influence of the membership function in each particle on the system is evaluated according to the above fitness function.

[0031] Specifically, the fitness function evaluates the impact of the membership function parameters of the particle representatives on the system performance. In IT2FL systems, the fitness is usually related to the control error.

[0032] In one embodiment, the method of the present invention further includes: adjusting the fitness function according to the actual feedback result of the system to the control instruction.

[0033] In one embodiment, the method of the present invention further comprises determining the individual optimal fitness and the global optimal fitness of the particle by the following steps: if the current fitness of the particle is better than its individual historical optimal fitness, updating the individual optimal fitness according to the current fitness of the particle; if the individual optimal fitness of the particle is better than the global optimal fitness, updating the global optimal fitness of the particle according to the individual optimal fitness of the particle.

[0034] In one embodiment, the position and velocity of each particle are updated according to the individual optimal fitness and the global optimal fitness to adjust the membership function in each particle, including: Determine the particle position update formula as ; The particle velocity update formula is determined as ; in, represents the position of particle i in the last t-th iteration, represents the speed of particle i in the current t+1th iteration process, is the inertia weight, c1 and c2 are learning factors, is the individual optimal fitness of particle i, is the global optimal fitness.

[0035] Specifically, according to the position of the optimal particle, the parameters of the membership function are adjusted to optimize the performance of the fuzzy controller. This may include adjusting the center point and width of the UMF and LMF to better adapt to the uncertainty in the production process.

[0036] In one embodiment, in response to obtaining a fuzzy set output by a particle swarm optimization algorithm, the corresponding fuzzy inference rules are activated and fuzzy inference is performed based on the upper and lower membership functions in the fuzzy set to obtain an inference set, including: activating the corresponding fuzzy inference rules according to the influence of each particle on the system; selecting or performing logical operations on the membership functions in the particles according to the fuzzy inference rules using a preset rule algorithm to obtain inference results to form an inference set. Wherein, defuzzifying the inference set to obtain a control instruction includes: using a weighted average method to convert the aggregated inference results into clear control instructions.

[0037] Specifically, interval type-2 fuzzy logic reasoning is the core part of the intelligent manufacturing process optimization system. It uses interval type-2 fuzzy sets (IT2FS) (i.e., fuzzy sets formed by the value range determined by the upper and lower membership functions in this application) to represent and process the uncertainty and ambiguity in the production process, thereby improving the decision-making quality and adaptability of the system. The preliminary preparations required for fuzzy reasoning include: Fuzzy set definition: A series of fuzzy sets are defined to represent key parameters in the production process, such as temperature, pressure, speed, etc. Membership function: A membership function is defined for each fuzzy set to measure the degree of membership of the data in the fuzzy set; in an optional implementation, the membership function of the fuzzy set is defined as:

[0038] Where X is a certain interval, u(x) and v(x) are the upper membership function and the lower membership function, respectively, which define the possibility that the data x belongs to the fuzzy set A; Fuzzy rule base: A fuzzy rule base containing multiple production scenarios is constructed to guide the decision-making of the production process; in an optional implementation, the fuzzy rules include: High temperature, high pressure and low humidity can improve product quality, but they consume more energy and reduce efficiency; High temperature, high pressure and high humidity can improve product quality, but they consume more energy and increase production efficiency; Low temperature, low pressure and low humidity will reduce product quality, increase energy consumption and reduce production efficiency.

[0039] Fuzzy inference engine: implements an inference algorithm based on fuzzy inference rules, used to make decisions based on input data and fuzzy rules. In an optional embodiment, the fuzzy inference rules include: Fuzzy logic AND operation:

[0040] For two fuzzy sets A and B, the membership of their intersection A∩B is the minimum of the membership of x and y.

[0041] Fuzzy logic OR operation:

[0042] For two fuzzy sets A and B, the membership of their union A∪B is the maximum of the membership of x and y.

[0043] In an optional implementation, the fuzzy reasoning process includes: ① Membership function update: According to the actual collected data of the system, for UMF and LMF, the center point and width are adjusted according to the particle parameters to update the membership.

[0044] ②Fuzzy reasoning aggregation Aggregate the reasoning results of all activated fuzzy inference rules to obtain the final fuzzy set

[0045] Where B(y) is the aggregation matrix of parameter y. The aggregation matrix is ​​formed by performing AND operations on the parameters Aij internally and then performing OR operations between the parameters.

[0046] ③ Defuzzification Use the weighted average method to convert the aggregated fuzzy sets into clear control instructions:

[0047] ④ Adaptive control command generation Based on the defuzzification results, control instructions are generated and applied to the production process.

[0048] ⑤Performance evaluation Evaluate system performance and adjust the fuzzy membership function based on the evaluation results to optimize system performance.

[0049] The complete workflow of the present invention is as follows Figure 3 As shown, firstly, particle encoding is performed, that is, the particles are initialized and the positions of the particles are randomly assigned in the search space; the fitness value is calculated, that is, the collected parameters are substituted into the fitness function obtained based on the upper and lower membership functions of each particle to determine the influence of the parameter on the system according to the fitness; the optimal solution is updated, that is, the individual optimal fitness update and the global optimal fitness update are performed according to the fitness; the speed and position of the ion are updated, that is, if the individual optimal fitness update or the global optimal fitness update occurs, the speed and position of the ion need to be updated based on this; the fuzzy controller parameters are updated, that is, the fuzzy inference rules constructed based on the upper and lower membership functions are updated, that is, the fuzzy inference model is updated; the performance evaluation, that is, whether the system output results based on the particle swarm model and the fuzzy inference model meet the expectations, if not, the particle swarm model and the fuzzy inference model need to be adjusted accordingly, and the fitness value calculation step is returned; if it meets, the optimal solution is output and the process ends.

[0050] Through the above process, the intelligent manufacturing process optimization system can realize adaptive control and optimization of the production process. The system can handle the uncertainty and dynamic changes in the production process and improve production efficiency and product quality.

[0051] The above are exemplary embodiments disclosed in the present invention, but it should be noted that various changes and modifications may be made without departing from the scope of the embodiments disclosed in the claims. The functions, steps and / or actions of the method claims according to the disclosed embodiments described herein do not need to be performed in any particular order.

[0052] The serial numbers of the embodiments disclosed in the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0053] A person skilled in the art should understand that the discussion of any of the above embodiments is only exemplary and is not intended to imply that the scope of the disclosure of the embodiments of the present invention (including the claims) is limited to these examples; under the concept of the embodiments of the present invention, the technical features in the above embodiments or different embodiments can also be combined, and there are many other changes in different aspects of the above embodiments of the present invention, which are not provided in detail for the sake of simplicity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present invention should be included in the protection scope of the embodiments of the present invention.

Claims

1. An adaptive interval type-2 fuzzy logic enhanced intelligent manufacturing control method, characterized in that: The method comprises: Predicting a fuzzy set that affects a target system by a particle swarm optimization algorithm, wherein the fuzzy set includes one or more dimensions, and each dimension has a numerical range that can produce a corresponding impact determined by an upper membership function and a lower membership function; In response to obtaining a fuzzy set output by the particle swarm optimization algorithm, a corresponding fuzzy inference rule is activated and fuzzy inference is performed based on upper and lower membership functions in the fuzzy set to obtain an inference set; The inference set is defuzzified to obtain control instructions, and the target system is controlled to perform intelligent manufacturing based on the control instructions.

2. The intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic according to claim 1 is characterized in that: The particle swarm optimization algorithm is used to predict the fuzzy sets that affect the target system, including: Construct a multidimensional search space based on the controllable parameters of the target system; For each dimension in the multidimensional search space, an upper membership function and a lower membership function for determining a particle position range are constructed respectively, and a particle cluster is initialized based on the upper membership function and the lower membership function; According to the preset fitness function, the influence of the membership function in each particle on the system is evaluated, and the individual optimal fitness and the global optimal fitness are determined respectively; The position and speed of each particle are updated according to the individual optimal fitness and the global optimal fitness to adjust the membership function in each particle, and a fuzzy set formed by the membership function in each particle is output.

3. The intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic according to claim 2 is characterized in that: Initializing a particle cluster based on the upper membership function and the lower membership function includes: Constructing multidimensional particles according to the center position and width in the upper membership function and the lower membership function corresponding to each dimension respectively; The center position of each particle is randomly assigned in each dimension of the search space to randomly assign the initial position of the particle.

4. The intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic according to claim 2 is characterized in that: The influence of the membership function in each particle on the system is evaluated according to the preset fitness function, including: The preset fitness function is determined as: ， in, represents the control error of the corresponding controllable parameter; The influence of the membership function in each particle on the system is evaluated according to the above fitness function.

5. The intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic according to claim 4 is characterized in that: The method further comprises: The fitness function is adjusted according to the actual feedback result of the system to the control instruction.

6. The intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic according to claim 2 is characterized in that: The method further comprises determining the individual optimal fitness of the particles and the global optimal fitness by the following steps: If the current fitness of a particle is better than its individual historical optimal fitness, the individual optimal fitness is updated according to the current fitness of the particle; If the individual optimal fitness of a particle is better than the global optimal fitness, the global optimal fitness of the particle is updated according to the individual optimal fitness of the particle.

7. The intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic according to claim 2 is characterized in that: The position and velocity of each particle are updated according to the individual optimal fitness and the global optimal fitness to adjust the membership function in each particle, including: Determine the particle position update formula as ； Among them, represents the position of particle i in the previous t-th iteration, and represents the speed of particle i in the current t+1-th iteration.

8. The intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic according to claim 7 is characterized in that: The method further comprises: The particle velocity update formula is determined as ； in, represents the position of particle i in the last t-th iteration, represents the speed of particle i in the current t+1th iteration process, is the inertia weight, c1 and c2 are learning factors, is the individual optimal fitness of particle i, is the global optimal fitness.

9. The intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic according to claim 4 is characterized in that: In response to obtaining a fuzzy set output by the particle swarm optimization algorithm, a corresponding fuzzy inference rule is activated and fuzzy inference is performed based on upper and lower membership functions in the fuzzy set to obtain an inference set, including: Activate the corresponding fuzzy inference rules according to the influence of each particle on the system; According to the fuzzy inference rules, a preset rule algorithm is used to select or perform logical operations on the membership functions in the particles to obtain inference results to form an inference set.

10. The intelligent manufacturing control method enhanced by adaptive interval type-2 fuzzy logic according to claim 9 is characterized in that: Defuzzifying the inference set to obtain a control instruction includes: The weighted average method is used to convert the aggregated reasoning results into clear control instructions.

Images