A method and system for constructing a drive controller module based on an expert system

Abstract

The application discloses a kind of based on expert system's drive controller module construction method and system, belong to rudder machine design field, method mainly includes: S1.Establish controller module database;S2.Adopt improved Apriori algorithm and dig the association rule between controller module, and construct rule base;S3.Based on the rule base, construct controller module reasoning system, for automatically recommending controller module according to input information;S4.Construction evaluation function carries out comprehensive evaluation to the reasoning result of controller module reasoning system, optimizes the selection of controller module, obtains drive controller design scheme;S5.Based on the drive controller design scheme obtained in step S4 updates controller module database.The application realizes the construction of controller module application system by the rule that is dug, promotes the application of modularization, realizes controller autonomous design, and then improves the research and development efficiency of electric rudder, shortens development cycle, satisfies market demand.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of steering engine design, and in particular to a driving controller module construction method and system based on an expert system. BACKGROUND

[0002] As a power actuator on an aircraft, an electric steering engine has been widely used in many types of aircraft systems. With the increasing number of aircraft types, how to achieve rapid design of servo mechanisms to adapt to rich application scenarios and meet the growing social needs is the focus of the industry.

[0003] Currently, the modularization scheme of the electric mechanism controller has shortened the development cycle of the electric steering engine to a certain extent, but the new problem that follows is that there is no complete system for the management and application of the modules, the selection of the modularization is highly dependent on the designers, and the quality of the products is closely related to the experience and level of the designers, resulting in poor popularization of the modularization application.

[0004] In summary, the main problems existing in the modular design scheme are as follows:

[0005] ① The electric steering engine controller modules are scattered, and there is a lack of a complete management system;

[0006] ② The association rules between the electric steering engine controller modules are not clear, and the knowledge reuse is not complete;

[0007] ③ The application mode of the electric steering engine controller module is not clear, the automation degree is low, and the selection of the module is not convenient. SUMMARY

[0008] The present application relates to the field of steering engine design, and in particular to a driving controller module construction method and system based on an expert system.

[0009] The purpose of the present application is achieved by the following technical solutions:

[0010] In a first aspect, a driving controller module construction method based on an expert system is provided, comprising:

[0011] S1. Establishing a controller module database;

[0012] S2. Mining the association rules between the controller modules by using an improved Apriori algorithm, and constructing a rule base;

[0013] S3. Constructing a controller module reasoning system based on the rule base, for automatically recommending controller modules according to input information;

[0014] S4. Constructing an evaluation function to comprehensively evaluate the inference results of the controller module inference system, optimizing the selection of the controller module, and obtaining a driving controller design scheme;

[0015] S5. Updating the controller module database based on the driving controller design scheme obtained in step S4.

[0016] In some embodiments, the establishment of the controller module database comprises the following steps:

[0017] S11. Establishing an electric rudder case library for storing design cases of electric rudder controller modules;

[0018] S12. Constructing a correlation degree function between microcircuit components;

[0019] S13. Determining the optimal cluster number of the controller modules based on the contour coefficient method and error sum of squares;

[0020] S14. Labeling each controller module by using a fuzzy C-means clustering algorithm, dividing each controller module into different intervals, generating first-level information and second-level information of each controller module, and calculating the membership degree of each controller module belonging to the corresponding interval;

[0021] S15. Constructing a controller module database based on the classification information and membership degree of each controller module.

[0022] In some embodiments, the correlation degree function is

[0023]

[0024] wherein, represents the correlation degree function between microcircuit components, represents the multi-dimensional factors affecting the correlation degree, including fault dependence, physical connection, fault propagation, and environmental influence, represents the weight of each factor.

[0025] In some embodiments, the improved Apriori algorithm is used to mine the association rules between the controller modules, comprising:

[0026] The membership degree is brought into the Apriori algorithm to calculate the confidence of the association rules.

[0027] In some embodiments, the controller module inference system is constructed based on the rule base, comprising:

[0028] The known information delivered by the servo mechanism task is processed into the representation form of each rule in the association rules, and the processed information is used as the input of the knowledge inference system to complete the knowledge inference. The rules and confidence obtained by the inference are recorded.

[0029] In some embodiments, the evaluation function is

[0030]

[0031] wherein, represents the evaluation function, and is a weight coefficient, represents the association function between each component and the microcircuit component, is a function representing the functional attribute of each component.

[0032] In some embodiments, the evaluation function is The calculation formula is as follows:

[0033]

[0034] wherein, represents the association function between other types of components and the microcircuit in different dimensional factors and the inferred controller module, represents the weight.

[0035] In some embodiments, the evaluation function is According to the component type, it is respectively constructed based on the polynomial attenuation, the empirical model, the exponential attenuation and the lookup table and interpolation model function.

[0036] In some embodiments, the selection of the optimized controller module comprises:

[0037] For the controller module with low evaluation function value, if it is accepted, it is used, and if it is not accepted, it is replaced.

[0038] In a second aspect, a driving controller module construction system based on an expert system is provided, comprising:

[0039] A controller module database construction module is used to establish a controller module database.

[0040] A rule base construction module is used to mine the association rules between controller modules by using the improved Apriori algorithm, and construct a rule base.

[0041] An inference system construction module is used to construct a controller module inference system based on the rule base, and automatically recommend a controller module according to input information.

[0042] An inference evaluation module is used to construct an evaluation function to comprehensively evaluate the inference results of the controller module inference system, optimize the selection of the controller module, and obtain a driving controller design scheme.

[0043] An iterative updating module is configured to update the controller module database based on the driving controller design scheme obtained in the inference evaluation module.

[0044] It should be further explained that the technical features of the above embodiments can be combined or replaced with each other to form new technical solutions without conflict.

[0045] The application first establishes a controller module database, and then performs association rule mining between modules based on the association rule algorithm on the basis of the controller module database. The inference system of the module application is built through the mined rules, the modular application is promoted, the controller is designed autonomously, and the research and development efficiency of the electric rudder is improved, the development cycle is shortened, and the market demand is met. Compared with the prior art, the application has the following advantages:

[0046] 1. The application establishes an electric rudder case library, and realizes unified management of electric rudder controller modules;

[0047] 2. The application realizes the mining of electric rudder controller design knowledge, completes the reuse of design knowledge, and improves the inheritance of design knowledge;

[0048] 3. The application establishes an expert system of electric rudder controller modules, and constructs an evaluation function to comprehensively evaluate the selected modules, so as to facilitate the selection and optimization of the modules and reduce the dependence on the design personnel;

[0049] 4. It is universal. The inference system established in the application is suitable for the design of electric rudder controllers of various types of aircraft, and can quickly complete the design of electric rudder controllers;

[0050] 5. It is generalizable. The application proposes a driving controller module construction method based on an expert system, which facilitates the promotion of electric rudder controller modular application, realizes the rapid design of electric rudder controllers, meets the rich market demand, and has promotional significance. BRIEF DESCRIPTION OF DRAWINGS

[0051] Figure 1 The flowchart of the driving controller module construction method based on the expert system of the application;

[0052] Figure 2 The flowchart of the fuzzy C-means clustering algorithm for module labeling processing of the application;

[0053] Figure 3 The flowchart of the improved Apriori algorithm for obtaining the association rules between the controller modules of the application;

[0054] Figure 4 The application flowchart of the inference system. DETAILED DESCRIPTION

[0055] The technical solutions of the present application will be described clearly and completely below in conjunction with the drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings can be arranged and designed in various different configurations. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0056] It should be noted that the defects of the above prior art solutions are the results obtained by the inventors after practice and careful study. Therefore, the discovery process of the above problems and the solutions proposed by the embodiments of the present application to the above problems should be the contributions made by the inventors to the present application in the process of invention and creation, and should not be understood as technical contents known to those skilled in the art.

[0057] In an exemplary embodiment, referring to Figure 1 , a method for constructing a drive controller module based on an expert system is provided, comprising:

[0058] S1. establishing a controller module database;

[0059] S2. using an improved Apriori algorithm to mine the association rules between controller modules, and constructing a rule base;

[0060] S3. constructing a controller module reasoning system based on the rule base, for automatically recommending a controller module according to input information;

[0061] S4. constructing an evaluation function to comprehensively evaluate the reasoning results of the controller module reasoning system, optimizing the selection of the controller module, and obtaining a drive controller design scheme;

[0062] S5. updating the controller module database based on the drive controller design scheme obtained in step S4.

[0063] In some embodiments, the establishment of the controller module database comprises the following steps:

[0064] S11. establishing an electric rudder case library for storing design cases of electric rudder controller modules as a basis for association rule mining;

[0065] S12. constructing an association degree function between microcircuit components;

[0066] S13. determining the optimal cluster number of the controller module based on the contour coefficient method and error sum of squares;

[0067] S14. The fuzzy C-means clustering algorithm is used to label each controller module, divide each controller module into different intervals, generate the first-level information and second-level information of each controller module, and calculate the membership of each controller module belonging to the corresponding interval;

[0068] S15. A controller module database is constructed based on the classification information and membership of each controller module.

[0069] In step S12, it is known that the electromechanical servo system driving controller is generally composed of the following electronic components: microcircuit components, semiconductor discrete components, resistors, capacitors, and oscillator components. In this embodiment, the microcircuit components are the core components, the correlation function between the microcircuit components is constructed according to multiple dimensions such as fault dependence, physical connection, fault propagation, and environmental influence, the correlation between the microcircuits is quantified, and the obtained quantified continuity value is used to determine the number of controller module category clusters. Exemplarily, the correlation function is

[0070]

[0071] wherein, represents the correlation function between the microcircuit components, represents the multiple-dimensional factors affecting the correlation, and , including fault dependence, physical connection, fault propagation, and environmental influence, represents the weight of each dimension factor, and .

[0072] In step S13, the correlation degree between the quantified microcircuit components can be obtained according to the correlation function, and the evaluation function value is a continuous value. Based on the cohesion and separation degrees, the contour coefficient method is used to determine the number of clusters of a series of continuous values.

[0073] The central idea of the contour coefficient method is to calculate the cohesion and separation degrees between the data to ensure that the original data used for clustering analysis is unchanged, and to complete the influence evaluation of different clustering algorithms or the same clustering algorithm with different clustering cluster numbers on the data clustering processing to obtain the final result. The calculation formula of the contour coefficient S for the entire data set is:

[0074]

[0075] wherein, N is the total number of samples; is the contour coefficient of each sample; , wherein represents the average distance from the sample point to other points in the same cluster, and the average distance from the sample point to other clusters is also calculated, and the smaller distance is .

[0076] After the application of clustering algorithm to complete the data of no label classification, the silhouette coefficient of the data contained in the K data clusters after clustering is calculated respectively, and the silhouette coefficient Must change in [-1, 1] interval, and the silhouette coefficient is closer to 1, the better the clustering result is, and the average value of the silhouette coefficient of all data for clustering analysis is the silhouette coefficient when the number of clustering clusters is K.

[0077] After determination by using the silhouette coefficient method, further verification is carried out by using SSE (sum of the squared errors), and the calculation formula of SSE is as follows:

[0078]

[0079] Wherein, K is the number of clustering clusters; Is the kth cluster; Is the centroid of the kth cluster; Is a data point in the cluster.

[0080] According to the silhouette coefficient method and SSE, the number of clusters of the drive controller is K=4, that is, the controller of the electromechanical servo system is divided into four categories, which are control module, communication module, drive module and other functional modules.

[0081] In step S14, as shown in Figure 2 According to the K value calculated in the last step, the fuzzy C mean clustering algorithm is used for module labeling processing, that is, each module is divided into different intervals, and through continuous iteration algorithm processing, the classification shown in table 1 is finally generated, and the membership of the module belonging to the interval The core calculation steps of fuzzy C mean clustering algorithm are as follows:

[0082]

[0083]

[0084] The module meaning after the tag processing is shown in Table 1 below. In the present application, there are three or four types of controllers, which are: control module (code: A), communication module (code: B), drive module (code: C) and other functional modules (D). The above module information is taken as first-level information, wherein the control module, the communication module, the drive module and the other functional modules can be further decomposed into second-level information, for example, the master control module (code: A) can be divided into FPGA module (code: A1m, m≥1, taking integer) and single-chip microcomputer module (code: A2m, m≥1, taking integer) as second-level information, and the code in the second-level information is not limited, i.e. Anm (n≥1, m≥1, n and m taking integer).

[0085] Table 1: Storage information representation of controller module database

[0086]

[0087] In step S15, the components of each case stored in the electric rudder case library are second-level or third-level information, and the specific storage content is shown in Table 2 below.

[0088] Table 2: Electric rudder case library

[0089]

[0090] In step S2, the electric rudder controller design knowledge mining is completed and a rule library is formed, as shown in Table 2. Figure 3 Based on the electric rudder case library, the improved Apriori algorithm is used to obtain the association rules between the controller modules, and the membership degrees obtained by the fuzzy C-means clustering algorithm are brought into the Apriori algorithm to calculate the confidence of the rules. For example, the calculation formula of a certain set is as follows:

[0091]

[0092] In order to obtain high-quality controller module design knowledge, the support threshold and the confidence threshold are set, wherein the support threshold is 0.1 and the confidence threshold is 0.6. The obtained knowledge expression form is shown in Table 3 below, and the first rule is taken as an example for explanation: A11→B21, wherein A11 is known information and B21 is inference information, and 0.785 is the confidence of the association rule.

[0093] Table 3: Specific form of the mined association rules

[0094]

[0095] The above-mentioned mined rules are constructed into a rule library in the format in Table 3.

[0096] In step S3, the establishment of the electric rudder controller module reasoning system is completed based on the principle of evidence theory reasoning, as shown in FIG. 4. First, the known information of the servo mechanism task is processed into the representation form of each rule in the association rule, and the processed information is taken as the input of the knowledge reasoning system to complete the knowledge reasoning. The rules and confidence obtained by reasoning are recorded. The whole process does not depend on the designer and has a high degree of automation.

[0097] Exemplarily, it is known that a rudder system with a torque of X and a speed of Y is needed to be designed. First, the selection and design of the motor are completed based on the torque and speed parameters of the rudder system. The driving module is selected according to the selected motor type. The driving module Cnm(n≥1, integer) is finally determined as the known information as the input of the reasoning system according to the conversion of the design knowledge completed above. After confirming the known information, the process in FIG. 4 is entered. The obtained reasoning result is used to realize the reuse of the design knowledge of the electric rudder.

[0098] In step S4, an evaluation function is set in the reasoning system The evaluation function is used to evaluate the results of each component in the reasoning module, determine the mapping relationship between other types of components and microcircuit components, and construct the evaluation function with the functional attributes as auxiliary measurement tools to quantify the correlation degree between each component and the microcircuit.

[0099] Exemplarily, the evaluation function is

[0100]

[0101] wherein, represents the evaluation function, and is a weight coefficient, represents the correlation degree function of each component and the microcircuit component, is a functional attribute function of each component.

[0102] The calculation formula of the correlation degree function of each component and the microcircuit component is as follows:

[0103]

[0104] wherein, represents the correlation degree function of other types of components and the microcircuit on the controller module in different dimension factors, and if there is no microcircuit component on the circuit, 0, the closer to the value 1, the higher the correlation degree between the component and the microcircuit component, represents the weight, and​ .

[0105] The According to the component category, the functions are respectively constructed based on polynomial attenuation, empirical construction, exponential attenuation and lookup table and interpolation model.

[0106] Further, the controller module with low evaluation function value is analyzed, if acceptable, it is used, if not acceptable, the electronic component with low evaluation value is replaced, and the optimization of the controller module is completed.

[0107] In step S5, the driving controller design scheme obtained based on the above method is stored in the electric rudder case library as a new case, and iteration of the expert system is realized.

[0108] In another exemplary embodiment, a driving controller module construction system based on an expert system is provided, comprising:

[0109] The controller module database construction module is used to establish the controller module database;

[0110] The rule base construction module is used to mine the association rules among the controller modules by using the improved Apriori algorithm, and construct the rule base;

[0111] The reasoning system construction module is used to construct the controller module reasoning system based on the rule base, and automatically recommend the controller module according to the input information;

[0112] The reasoning evaluation module is used to construct the evaluation function to comprehensively evaluate the reasoning result of the controller module reasoning system, optimize the selection of the controller module, and obtain the driving controller design scheme;

[0113] The iterative update module is used to update the controller module database based on the driving controller design scheme obtained by the reasoning evaluation module.

[0114] The above specific embodiments are detailed descriptions of the present application, which cannot be recognized as the specific embodiments of the present application only limited to these descriptions. For ordinary skilled in the art to which the present application belongs, without departing from the concept of the present application, a number of simple deductions and substitutions can be made, which should be regarded as belonging to the protection scope of the present application.

Claims

1. A method of constructing an expert system based drive controller module, characterized by, The method comprises the following steps: S1. establishing a controller module database; The establishment of the controller module database comprises the following steps: S11. establishing an electric rudder case library for storing design cases of electric rudder controller modules; S12. constructing a correlation degree function between microcircuit components; S13. determining the optimal clustering cluster number of the controller modules based on a profile coefficient method and error sum of squares; S14. performing labelization processing on each controller module by using a fuzzy C-means clustering algorithm, dividing each controller module into different intervals, generating first-level information and second-level information of each controller module, and calculating the membership of each controller module belonging to the corresponding interval; S15. constructing the controller module database based on the classification information and the membership of each controller module; S2. mining association rules between the controller modules by using an improved Apriori algorithm, and constructing a rule base; S3. constructing a controller module reasoning system based on the rule base, for automatically recommending controller modules according to input information; S4. constructing an evaluation function to comprehensively evaluate the reasoning result of the controller module reasoning system, optimizing the selection of the controller modules, and obtaining a drive controller design scheme; S5. updating the controller module database based on the drive controller design scheme obtained in step S4.

2. The method of claim 1, wherein the method further comprises: The correlation degree function is wherein, represents a correlation function between microcircuit components, represents multi-dimensional factors affecting the correlation, including fault dependency, physical connection, fault propagation, and environmental impact, represents the weight of each factor.

3. The method of claim 1, wherein the method further comprises: The mining of the association rules between the controller modules by using the improved Apriori algorithm comprises: The membership is brought into the Apriori algorithm to calculate the confidence of the association rules.

4. The method of claim 3, wherein the expert system-based drive controller module is constructed by, The construction of the controller module reasoning system based on the rule base comprises: Known information delivered by a servo mechanism task is processed into a representation form of each rule in the association rules, and the processed information is taken as input of the knowledge reasoning system to complete knowledge reasoning, and the rules and the confidence obtained by reasoning are recorded.

5. The method of claim 1, wherein the method further comprises: The evaluation function is wherein, represents an evaluation function, and is a weight coefficient, represents a correlation function of each component and microcircuit component, is a function representing the functional attribute of each component.

6. The method of claim 5, wherein the method further comprises: The The calculation formula is as follows: wherein, represents a correlation function between other class components and the inferred controller module on microcircuit in different dimensional factors, represents a weight.

7. The method of claim 5, wherein the method further comprises: The Depending on the component type, the models are constructed based on polynomial decay, empirical, exponential decay, and look-up table and interpolation model functions, respectively.

8. The method of claim 5, wherein the expert system-based drive controller module is constructed by, The optimization of the selection of the controller modules comprises: Controller modules with low evaluation function values are analyzed, and if accepted, the controller modules are used, and if not accepted, the controller modules are replaced.

9. An expert system based drive controller module building system, characterized by, The method comprises the following steps: A controller module database construction module is configured to establish a controller module database; The establishment of the controller module database comprises the following steps: S11. establishing an electric rudder case library for storing design cases of electric rudder controller modules; S12. constructing a correlation degree function between microcircuit components; S13. determining the optimal clustering cluster number of the controller modules based on a profile coefficient method and error sum of squares; S14. performing labelization processing on each controller module by using a fuzzy C-means clustering algorithm, dividing each controller module into different intervals, generating first-level information and second-level information of each controller module, and calculating the membership of each controller module belonging to the corresponding interval; S15. constructing the controller module database based on the classification information and the membership of each controller module; A rule base construction module is configured to mine association rules between the controller modules by using an improved Apriori algorithm, and construct a rule base; A reasoning system construction module is configured to construct a controller module reasoning system based on the rule base, for automatically recommending controller modules according to input information; The reasoning evaluation module is configured to construct an evaluation function to comprehensively evaluate the reasoning result of the controller module reasoning system, optimize the selection of the controller module, and obtain a drive controller design scheme. The iterative updating module is configured to update the controller module database based on the drive controller design scheme obtained by the reasoning evaluation module.

Images