Combine harvester manufacturing quality analysis and decision-making method and system based on information fusion

Abstract

The present invention proposes a manufacturing quality analysis and decision-making method and system for combine harvesters based on information fusion. The analysis and decision-making method consists of a five-level structure of heterogeneous data grouping, component quality assessment, performance risk assessment, manufacturing quality decision, and template matching library update. The analysis and decision-making method uses information fusion technology to perform decision-level fusion on the component quality data analysis results, and uses satisfaction criteria to perform decision-making evaluation on the performance of combine harvesters, thereby providing decision support for improving the quality management level of the manufacturing process of combine harvesters.

Description

Technical Field

[0001] The invention relates to the technical field of combine harvester manufacturing, and proposes a combine harvester manufacturing quality analysis decision method based on information fusion, which is used for quality analysis in the combine harvester manufacturing process. Background Art

[0002] Today, with the high level of agricultural mechanization, agricultural combine harvesters have been increasingly widely used, and the working performance and efficiency of combine harvesters have gradually become important reference indicators for the selection of combine harvesters. In other words, combine harvesters with high manufacturing quality and low failure rate will bring greater convenience to agricultural production and will be more popular among agricultural workers. However, the manufacturing technology of agricultural machinery in my country is not well developed, and many production links have not formed a sufficiently perfect system, which poses a challenge to improving the manufacturing quality of combine harvesters.

[0003] At present, there are few existing patents dedicated to improving the manufacturing quality of combine harvesters. The manufacturing quality improvement methods proposed in the existing patents test the components of the whole machine at the same time, which has certain limitations in clarifying the manufacturing problems of combine harvesters and has limited help in improving the manufacturing quality of combine harvesters. In terms of quality decision-making, the existing patents have relatively simple analysis of product manufacturing quality and do not take into account the interrelationships of product components, resulting in errors in the decision-making results.

[0004] During the manufacturing process of combine harvesters, combine harvester manufacturers need to assemble key components of combine harvesters manufactured by various parts suppliers, including headers, conveying devices, separation devices, and cleaning devices. The manufacturing quality of these components and their parts and the final assembly quality directly affect the working performance of the combine harvester. The numerous sources of combine harvester components and the high degree of correlation put higher demands on their analysis and decision-making. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention proposes a combine harvester manufacturing quality analysis and decision-making method based on information fusion. The main functional parts, components, and systems of the combine harvester are analyzed level by level through the five-level processing structure of the quality analysis and decision-making method. The manufacturing quality and abnormal trend of the combine harvester are predicted by comparing the manufacturing quality data with the knowledge base, layer-by-layer probabilistic reasoning, and decision-level information fusion. Decision support is provided for improving the quality management level of the combine harvester manufacturing process, and prompts and suggestions are given for the production of the combine harvester.

[0006] In order to solve the above technical problems, the present invention proposes a combine harvester manufacturing quality analysis and decision-making method based on information fusion, which includes:

[0007] Step 1, obtaining the manufacturing part data of the combine harvester to be quality analyzed, grouping the manufacturing part data according to the functional structure of the combine harvester, obtaining multiple functional structure part groups, respectively serving as primary nodes, and taking each part in the manufacturing part data as a secondary node of the functional structure part group to which it belongs, and constructing a part classification tree including the primary node and the secondary node according to the belonging relationship between the primary node and the secondary node;

[0008] Step 2: Input the part classification tree into the interval matching model to obtain the hazard coefficient of each part; input the safety index of each part and the hazard coefficient of each part into the rule matching model to obtain the hazard degree of each part; collect the hazard degrees of each part as input data and input them into the probabilistic reasoning network constructed according to the failure process of the functional structure to obtain the probability of each functional device of the combine harvester; process the failure risk valuation of each functional structure according to the probability of each functional device and the satisfaction criterion to obtain the manufacturing quality analysis result of the combine harvester.

[0009] The combine harvester manufacturing quality analysis and decision-making method based on information fusion, wherein step 2 comprises:

[0010] Step 21: According to V c =|x-x0| / (x max -x min ) to obtain the risk factor V of each part C , where x max is the upper limit of manufacturing error, x min is the lower limit of manufacturing error, and the risk degree of each part is obtained according to the safety index and risk factor of each part.

[0011] In the above-mentioned combine harvester manufacturing quality analysis and decision-making method based on information fusion, the process of constructing the probabilistic reasoning network in step 2 is specifically as follows:

[0012] By combining the criticality of each part, the source of the critical factors of each functional structure is obtained as the upper node, the failure form and empirical failure probability of the core function of the functional structure are quantified as the middle node, and the degree of loss caused by the final failure of the functional structure is taken as the lower node. The probability reasoning network is constructed according to the upper node-middle node-lower node.

[0013] The above-mentioned combine harvester manufacturing quality analysis and decision-making method based on information fusion, wherein the probability reasoning network obtains the child node V according to the following formula during the reasoning process: i The conditional probability of and use the joint tree algorithm to find the probability of any node in the inference network;

[0014]

[0015] In the formula, P(V i ) is the child node V i The parent node, P i Represents the parent node vector, vector value p i Represents vector P i The value of the vector value v i Represents vector V i , thereby obtaining the conditional probability of the child node.

[0016] The combine harvester manufacturing quality analysis and decision-making method based on information fusion further includes:

[0017] Step 3: Update the satisfaction criterion according to the actual use of the combine harvester.

[0018] The present invention also proposes a combine harvester manufacturing quality analysis and decision-making system based on information fusion, which includes:

[0019] Module 1, obtaining manufacturing part data of a combine harvester to be quality analyzed, grouping the manufacturing part data according to the functional structure of the combine harvester, obtaining multiple functional structure part groups, each of which is used as a first-level node, and using each part in the manufacturing part data as a second-level node of the functional structure part group to which it belongs, and constructing a part classification tree including the first-level node and the second-level node according to the relationship between the first-level node and the second-level node;

[0020] Module 2: By inputting the part classification tree into the interval matching model, the hazard coefficient of each part is obtained; by inputting the safety index of each part and the hazard coefficient of each part into the rule matching model, the hazard degree of each part is obtained; the hazard degree of each part is collected as input data and input into the probabilistic reasoning network constructed according to the failure process of the functional structure to obtain the probability of each functional device of the combine harvester; the failure risk valuation of each functional structure is processed according to the probability of each functional device and the satisfaction criterion to obtain the manufacturing quality analysis result of the combine harvester.

[0021] The combine harvester manufacturing quality analysis and decision-making system based on information fusion, wherein the module 2 comprises:

[0022] Module 21, according to V c =|x-x0| / (x max -x min ) to obtain the risk factor V of each part C , where x max is the upper limit of manufacturing error, x min is the lower limit of manufacturing error, and the risk degree of each part is obtained according to the safety index and risk factor of each part.

[0023] In the above-mentioned combine harvester manufacturing quality analysis and decision-making system based on information fusion, the process of constructing the probability reasoning network in module 2 is specifically as follows:

[0024] By combining the criticality of each part, the source of the critical factors of each functional structure is obtained as the upper node, the failure form and empirical failure probability of the core function of the functional structure are quantified as the middle node, and the degree of loss caused by the final failure of the functional structure is taken as the lower node. The probability reasoning network is constructed according to the upper node-middle node-lower node.

[0025] The combine harvester manufacturing quality analysis and decision-making system based on information fusion, wherein the probability reasoning network obtains the child node V according to the following formula during the reasoning process i The conditional probability of and use the joint tree algorithm to find the probability of any node in the inference network;

[0026]

[0027] In the formula, P(V i ) is the child node V i The parent node, P i Represents the parent node vector, vector value p i Represents vector P i The value of the vector value v i Represents vector V i , thereby obtaining the conditional probability of the child node.

[0028] The combine harvester manufacturing quality analysis and decision-making system based on information fusion further includes:

[0029] Module 3: updating the satisfaction criterion according to the actual use of the combine harvester.

[0030] It can be seen from the above scheme that the advantages of the present invention are:

[0031] The five-level structure consisting of heterogeneous data grouping, component quality assessment, performance risk assessment, manufacturing quality decision and template matching library update can realize the grouping and labeling of combine harvester manufacturing quality data, and evaluate and predict the main functional components, main working systems and expected performance of the whole machine step by step. In this process, the knowledge base is fully utilized to focus on the components and systems with high frequency of use, high working intensity and high failure rate in the combine harvester, to ensure that the analysis and decision on the manufacturing quality of the combine harvester are in line with the actual application situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a structural diagram of the analysis and decision-making method of the present invention;

[0033] Figure 2The following is a workflow diagram of the analysis and decision-making method of the present invention;

[0034] Figure 3 A schematic diagram of a Wien circle. DETAILED DESCRIPTION

[0035] The main purpose of the present invention is to propose a combine harvester manufacturing quality analysis and decision-making method based on information fusion, analyze and make decisions on the manufacturing quality of the combine harvester, and judge the manufacturing level of the combine harvester, aiming to improve the manufacturing quality of the combine harvester and its working performance, so as to achieve the technical effect of improving the yield rate of the combine harvester. In order to achieve the above purpose, the present invention is implemented according to the following technical scheme:

[0036] The analysis and decision-making method consists of a five-level structure of heterogeneous data grouping, component quality assessment, performance risk assessment, manufacturing quality decision and template matching library update: heterogeneous data grouping receives combine harvester manufacturing quality data from different sources (such as part size data of combine harvester functional components and component assembly data) and groups and marks them; part-level and component-level quality assessment compares the grouped data with the combine harvester manufacturing standards to obtain a preliminary manufacturing risk assessment; performance risk assessment performs information fusion on the basis of quality assessment to determine the comprehensive impact of low-level risks on the manufacturing quality of combine harvesters; manufacturing quality decision makes statistical decisions to determine the final manufacturing quality of combine harvesters; template matching library update compares the decision conclusions of the analysis decision method with the use of combine harvesters to learn the working characteristics of combine harvesters.

[0037] The reason for the need for information fusion is that some of the components of the combine harvester need to be processed or semi-processed by the whole machine factory, and some are purchased directly from the parts factory. The data sources are complex. The components and parts information used in the final assembled machine may come from the whole machine factory or from the parts factory. It is a complex combination process. In other words, there are many possible sources of performance problems in the combine harvester. Therefore, it is necessary to align the data from these different sources, and then first fuse the qualification of each part at the component level, and finally fuse the failure probability of the components at the whole machine level.

[0038] The heterogeneous data grouping, component quality assessment, performance risk assessment, and manufacturing quality decision in the technical solution together constitute a hybrid fusion structure. The heterogeneous data grouping groups and marks the combine harvester data packets, and then obtains preliminary risk assessments according to the harvesting table, conveying device, threshing and separation device, and cleaning device in the component quality assessment. Then, in the performance risk assessment structure, component-level information fusion is performed in the form of sub-processors, and finally, decision-level information fusion is performed at the whole machine level in the manufacturing quality decision.

[0039] The heterogeneous data grouping structure in the technical solution is presented as a Venn circle structure within the scope of the data set: the combine harvester is transmitted to the grouping structure in the form of a data packet, forming a first-level Venn circle, and then the grouping structure splits the data packet into the main functional component level according to the source of the data, forming four second-level Venn circles: the header, the conveying device, the threshing and separation device, and the cleaning device. At the same time, these four devices are the first-level nodes of the classification tree, denoted as U 1n , where n represents the number of the device: 11 Represents header, U 12 Represents conveyor, U 13 Stands for the separation device, U 14 Represents the cleaning device. Next, the data flow will continue to be grouped downward according to the four first-level nodes: the parts contained in the four functional systems are used as their respective second-level nodes, so that several second-level nodes will be obtained, recorded as U 2nk , where n represents the number of the first-level node to which the second-level node belongs, k represents the number of the part under the first-level node, and the maximum value of k is the number of parts to be tested in the first-level node. After the same data packet is grouped according to the grouping method, the production batch information m is added to the part number to obtain the final part number U 2nkm , realize data grouping and prepare for subsequent decision-level information fusion.

[0040] The component quality assessment structure in the technical solution consists of an interval matching model and a rule matching model: the interval matching model is used to compare and detect part size data and component assembly data with manufacturing standards to obtain the risk coefficient VC; the rule matching model is used to compare and detect the processing results of the interval matching model with the template matching library to obtain the component risk degree D.

[0041] The performance risk assessment structure in the technical solution evaluates the functional risks of the main functional components of the combine harvester in an information fusion manner. The working process is carried out in a probabilistic reasoning manner and mainly includes the following steps:

[0042] S10, receiving the processing results of the component quality assessment processing structure, taking the main source of risk factors represented by too small a disengagement gap as the upper node; quantifying the core functional failure form and empirical failure probability of the functional component (such as the threshing drum) as the middle node; taking the loss degree caused by the final failure of the functional component (such as the crushing rate of the crop) as the lower node, and completing the determination of the reasoning network node;

[0043] S20, according to the failure process of the functional components, a probability reasoning network is constructed according to the upper node-middle node-lower node, wherein the upper node serves as the parent node of the middle node and the middle node serves as the parent node of the lower node;

[0044] S30. In the reasoning process, since it is necessary to obtain the failure probability of each layer of nodes, the child node V is calculated based on the parent node probability of the child node. i The conditional probability is calculated as follows:

[0045]

[0046] In the formula, P(V i ) is the child node V i The parent node, P i Represents the parent node vector, vector value p i Represents vector P i The value of the vector value v i Represents vector V i The value of , thereby obtaining the conditional probability of the child node;

[0047] S40, using the joint tree algorithm in probabilistic reasoning to calculate the probability of any node in the reasoning network.

[0048] The manufacturing quality decision structure in the technical solution processes the failure risk estimation of each functional device according to the satisfaction criterion: the functional index estimation of the combine harvester is obtained by querying the template matching library, and the manufacturing level of each functional device of the combine harvester is determined. ＊ , where a ＊ =MaxP{Q(a i ,θ j )≤a}, a is a given satisfactory risk level, Q(a i ,θ j ) is the risk of option i under state j, P{Q(a i ,θ j )≤a} is the cumulative probability that the risk value of each evaluation scheme is not higher than the target value; the final manufacturing level A of the whole machine is calculated according to A=W*a ＊ 割台 +X*a ＊ 输送装置 +Y*a ＊ 脱粒分离装置 +Z*a ＊ The cleaning device is calculated to complete the fusion, where W is the weight of the harvesting platform, X is the weight of the conveying device, Y is the weight of the threshing and separation device, Z is the weight of the cleaning device, and W+X+Y+Z=1. At the same time, the weight is updated online according to the template matching library.

[0049] The template matching library update structure in the technical solution consists of a template acquisition module, a template storage module and a template comparison module: the template acquisition module receives the actual usage of the combine harvester and makes a query request for the decision situation to the template storage; the comparison module compares the decision situation and usage of the combine harvester, and proposes modification suggestions for the existing template matching criteria; the template storage module stores the decision situation of the combine harvester and the template matching criteria.

[0050] The interval matching model in the technical solution receives the manufacturing data of the parts and the qualified interval of the quality data, and obtains the risk factor VC of the test item: V c =|x-x0| / (x max -x min ), where x is the manufacturing value of the test item, x max is the upper limit of manufacturing error, x min is the lower limit of manufacturing error, when V c When >1, the risk factor of the test item is 1.

[0051] The processing results of the rule matching model in the technical solution are determined by the safety index calculation module and the risk calculation module: the safety index calculation module takes the severity S and exposure rate E of the component as input and obtains the safety index H through fuzzy logic; the risk calculation module calculates the safety index H according to the formula D = V c * The criticality D of the parts is calculated by arcsinH.

[0052] In order to make the above features and effects of the present invention more clearly understood, embodiments are given below and described in detail with reference to the accompanying drawings.

[0053] The present invention includes Figure 1 The five-level structure shown includes heterogeneous data grouping, component quality assessment, performance risk assessment, manufacturing quality decision making, and template matching library updating.

[0054] Specifically, the heterogeneous data grouping structure as the first-level structure of the method will first receive the manufacturing quality data of the combine harvester, which includes the main size data of important parts, assembly data and performance data of important functional components, and assembly data of the whole machine.

[0055] When the data is transmitted to the heterogeneous data marshalling structure, it will be grouped and marked for subsequent fusion and recording. The combine harvester transmits the data to the marshalling structure in the form of data packets, forming a Figure 3 The first-level Venn circle shown in the figure, and then the grouping structure splits the data packets into the main functional component level according to the different sources of the data, forming the following Figure 3 The four secondary Wien circles shown are: header, conveyor, separation device and cleaning device. At the same time, these four devices are the first-level nodes of the classification tree, denoted by U1n , where n represents the number of the device: 11 Represents header, U 12 Represents conveyor, U 13 Represents threshing and separation device, U 14 Represents the cleaning device. Next, the data flow will continue to be grouped downward according to the four first-level nodes: the parts contained in the four functional systems are used as their respective second-level nodes, so that several second-level nodes will be obtained, recorded as U 2nk , where n represents the number of the first-level node to which the second-level node belongs, k represents the number of the part under the first-level node, and the maximum value of k is the number of parts to be detected in the first-level node. After the same data packet is grouped according to the grouping method, the production batch information m is added to the part number to obtain the final part number U2nkm. At this point, the grouping mark of the parts has been completed, and the size data of the parts and the related assembly data will be directly linked to the part number, which is convenient for troubleshooting in the analysis and decision-making process.

[0056] Specifically, the component quality assessment structure consists of an interval matching model and a rule matching model. The interval matching model receives the part size data and component assembly data, and makes a query to the template matching library to obtain the qualified interval of the data, and then c =|x-x0| / (x max -x min ) to obtain the risk factor VC of each test item (part), where x max is the upper limit of manufacturing error, x min is the lower limit of manufacturing error, x is the manufacturing value of the test item, x0 is the manufacturing standard value of the test item, when the calculated risk factor V c When it is greater than 1, the risk factor of the inspection item is set to 1. In this way, the risk factor of the part will be transferred to the rule matching model according to the previous mark.

[0057] The processing result of the rule matching model is determined by the safety index calculation module and the hazard calculation module. In the safety index calculation module, the model mainly considers two factors: the severity S of the component and the exposure rate E. Here, the severity refers to the degree to which the working performance and working safety of the combine harvester are damaged once the risk becomes a reality; the exposure rate is the probability that the combine harvester will be affected when the risk occurs. In the process of using fuzzy logic control, the input severity S is represented by SX, and the value of X is 0 / 1 / 2 / 3 / 4, from low to high. The higher the level, the greater the degree of damage; the input exposure rate is represented by EX, and the value of X is 0 / 1 / 2 / 3 / 4, from low to high. The higher the level, the greater the probability of being affected; the output safety index H is represented by HX, and the value of X is 0 / 1 / 2 / 3 / 4, from low to high. The higher the level, the more serious the impact caused by the fault. The hazard calculation module receives the hazard coefficient and safety index of the components and calculates them according to D=V c *arcsinH calculates the criticality D of the component. The criticality D is a value in the interval [0,1] and will be substituted into the formula as the probability of the upper node in the following formula.

[0058] Specifically, the performance risk assessment structure evaluates the functional risks of the main functional systems of the combine harvester in a probabilistic reasoning manner. The reasoning process includes the following steps:

[0059] S10, receiving the processing results of the component quality assessment processing structure, taking the main risk factor sources of the functional components as the upper nodes; quantifying the failure forms and empirical failure probabilities of the core functions of the functional components as the middle nodes; taking the loss degree caused by the final failure of the functional components as the lower nodes, and completing the determination of the reasoning network nodes;

[0060] S20, according to the failure process of the functional component, a probability reasoning network is constructed according to the upper node-middle node-lower node, wherein the probability value of the upper node is the danger level D of the component;

[0061] S30. During the reasoning process, the child node V i The conditional probability is calculated as follows:

[0062]

[0063] In the formula, P(V i ) is the child node V i The parent node, P i Represents the parent node vector, vector value p i Represents vector P i The value of the vector value v i Represents vector V i The value of , thereby obtaining the conditional probability of the child node;

[0064] S40, using the joint tree algorithm to find the probability of any node in the reasoning network. The probability of the middle node is calculated by substituting the probability of the upper node into the formula, and the probability of the lower node is calculated by substituting the probability of the middle node into the formula; the probability of the lower node is finally used as the final result of the performance risk assessment structure (that is, the failure risk estimation of each functional device).

[0065] Specifically, the manufacturing quality decision structure processes the failure risk valuation of each functional device according to the satisfaction criterion: the functional index valuation of the combine harvester is obtained by querying the template matching library, and the manufacturing level of each functional device of the combine harvester is determined. * , where a * =MaxP{Q(a i ,θ j )≤a}, a is a given satisfactory risk level, Q(a i ,θ j ) is the risk of option i under state j, P{Q(a i ,θ j )≤a} is the cumulative probability that the risk value of each evaluation scheme is not higher than the target value; the final manufacturing level A of the whole machine is calculated according to A=W*a * 割台 +X*a * 输送装置 +Y*a * 脱粒分离装置 +Z*a * 清选装置 It is calculated that W is the weight of the header, X is the weight of the conveying device, Y is the weight of the threshing and separation device, and Z is the weight of the cleaning device, and W+X+Y+Z=1. At the same time, the weight is updated online according to the template matching library.

[0066] i refers to the differences in crop varieties, crop maturity, and natural height of the combine harvester. In fact, it refers to the operating objects and operating scenes of the combine harvester. j refers to the difference in the operating gear or operating speed of the combine harvester. (The above two points refer to the industry standard NYT2090-2011 Combine Harvester Test Conditions) That is, the operating status. The risk of the i scheme in the j state depends on the comparison with the template matching library, so that the degree of loss obtained according to the operating scheme (i scheme), operating status (j state) and risk assessment structure is obtained and quantified based on experience. This cumulative probability refers to the probability and probability of the corresponding state j occurring during use when the risk Q of the combine harvester is less than the given satisfaction level a.

[0067] Specifically, the template matching library update structure consists of a template acquisition module, a template storage module and a template comparison module: the template acquisition module receives the actual usage of the combine harvester and makes a query request for the decision situation to the template storage; the comparison module compares the decision situation and usage of the combine harvester, and proposes modification suggestions for the existing template matching criteria; the template storage module stores the decision situation of the combine harvester and the template matching criteria.

[0068] The following is a system embodiment corresponding to the above method embodiment. This embodiment can be implemented in conjunction with the above embodiment. The relevant technical details mentioned in the above embodiment are still valid in this embodiment. In order to reduce repetition, they are not repeated here. Accordingly, the relevant technical details mentioned in this embodiment can also be applied in the above embodiment.

[0069] The present invention also proposes a combine harvester manufacturing quality analysis and decision-making system based on information fusion, which includes:

[0070] Module 1, obtaining manufacturing part data of a combine harvester to be quality analyzed, grouping the manufacturing part data according to the functional structure of the combine harvester, obtaining multiple functional structure part groups, each of which is used as a first-level node, and using each part in the manufacturing part data as a second-level node of the functional structure part group to which it belongs, and constructing a part classification tree including the first-level node and the second-level node according to the relationship between the first-level node and the second-level node;

[0071] Module 2: By inputting the part classification tree into the interval matching model, the hazard coefficient of each part is obtained; by inputting the safety index of each part and the hazard coefficient of each part into the rule matching model, the hazard degree of each part is obtained; the hazard degree of each part is collected as input data and input into the probabilistic reasoning network constructed according to the failure process of the functional structure to obtain the probability of each functional device of the combine harvester; the failure risk valuation of each functional structure is processed according to the probability of each functional device and the satisfaction criterion to obtain the manufacturing quality analysis result of the combine harvester.

[0072] The combine harvester manufacturing quality analysis and decision-making system based on information fusion, wherein the module 2 comprises:

[0073] Module 21, according to V c =|x-x0| / (x max -x min ) to obtain the risk factor V of each part C , where x max is the upper limit of manufacturing error, x min is the lower limit of manufacturing error, and the risk degree of each part is obtained according to the safety index and risk factor of each part.

[0074] In the above-mentioned combine harvester manufacturing quality analysis and decision-making system based on information fusion, the process of constructing the probability reasoning network in module 2 is specifically as follows:

[0075] By combining the criticality of each part, the source of the critical factors of each functional structure is obtained as the upper node, the failure form and empirical failure probability of the core function of the functional structure are quantified as the middle node, and the degree of loss caused by the final failure of the functional structure is taken as the lower node. The probability reasoning network is constructed according to the upper node-middle node-lower node.

[0076] The combine harvester manufacturing quality analysis and decision-making system based on information fusion, wherein the probability reasoning network obtains the child node V according to the following formula during the reasoning process i The conditional probability of and use the joint tree algorithm to find the probability of any node in the inference network;

[0077]

[0078] In the formula, P(V i ) is the child node V i The parent node, P i Represents the parent node vector, vector value p i Represents vector P i The value of the vector value v i Represents vector V i , thereby obtaining the conditional probability of the child node.

[0079] The combine harvester manufacturing quality analysis and decision-making system based on information fusion further includes:

[0080] Module 3: updating the satisfaction criterion according to the actual use of the combine harvester.

Claims

1. A combine harvester manufacturing quality analysis and decision-making method based on information fusion, characterized in that: include: Step 1, obtaining the manufacturing part data of the combine harvester to be quality analyzed, grouping the manufacturing part data according to the functional structure of the combine harvester, obtaining multiple functional structure part groups, respectively serving as primary nodes, and taking each part in the manufacturing part data as a secondary node of the functional structure part group to which it belongs, and constructing a part classification tree including the primary node and the secondary node according to the belonging relationship between the primary node and the secondary node; Step 2: by inputting the part classification tree into the interval matching model, the hazard coefficient of each part is obtained; by inputting the safety index of each part and the hazard coefficient of each part into the rule matching model, the hazard degree of each part is obtained; the hazard degree of each part is integrated as input data and input into the probability reasoning network constructed according to the failure process of the functional structure, the failure probability of each functional device of the combine harvester is obtained; according to the failure probability of each functional device and the satisfaction criterion, the failure risk valuation of each functional structure is processed to obtain the manufacturing quality analysis result of the combine harvester; Wherein the step 2 comprises: Step 21: According to V c =|x-x0| / (x max -x min ) to obtain the risk factor V of each part C , where x max is the upper limit of manufacturing error, x min is the lower limit of manufacturing error, x is the manufacturing value of the test item, x0 is the manufacturing standard value of the test item, and the risk degree of each part is obtained according to the safety index and risk factor of each part; The specific process of constructing the probabilistic reasoning network in step 2 is: By combining the dangerousness of each part, the dangerous factor sources of each functional structure are obtained as the upper node, the failure form and empirical failure probability of the core function of the functional structure are quantified as the middle node, and the loss degree caused by the final failure of the functional structure is taken as the lower node. The probability reasoning network is constructed according to the upper node-middle node-lower node. During the reasoning process, the probability reasoning network obtains the child node V according to the following formula i The conditional probability of any node in the inference network is obtained by using the joint tree algorithm; In the formula, P(V i ) is the child node V i The parent node, P i Represents the parent node vector, vector value p i Represents vector P i The value of the vector value v i Represents vector V i The value of , thereby obtaining the conditional probability of the child node; Also includes: Step 3: Update the satisfaction criterion according to the actual use of the combine harvester.

2. A combine harvester manufacturing quality analysis and decision-making system based on information fusion, characterized in that: include: Module 1, obtaining manufacturing part data of a combine harvester to be quality analyzed, grouping the manufacturing part data according to the functional structure of the combine harvester, obtaining multiple functional structure part groups, each of which is used as a first-level node, and using each part in the manufacturing part data as a second-level node of the functional structure part group to which it belongs, and constructing a part classification tree including the first-level node and the second-level node according to the relationship between the first-level node and the second-level node; Module 2: by inputting the parts classification tree into the interval matching model, the risk coefficient of each part is obtained; by inputting the safety index of each part and the risk coefficient of each part into the rule matching model, the risk degree of each part is obtained; the risk degree of each part is integrated as input data and input into the probability reasoning network constructed according to the failure process of the functional structure, the failure probability of each functional device of the combine harvester is obtained; the failure risk valuation of each functional structure is processed according to the failure probability of each functional device and the satisfaction criterion, and the manufacturing quality analysis result of the combine harvester is obtained; This module 2 includes: Module 21, according to V c =|x-x0| / (x max -x min ) to obtain the risk factor V of each part C , where x max is the upper limit of manufacturing error, x min is the lower limit of manufacturing error, x is the manufacturing value of the test item, x0 is the manufacturing standard value of the test item, and the risk degree of each part is obtained according to the safety index and risk factor of each part; The specific process of building the probabilistic reasoning network in module 2 is as follows: By combining the dangerousness of each part, the dangerous factor sources of each functional structure are obtained as the upper node, the failure form and empirical failure probability of the core function of the functional structure are quantified as the middle node, and the loss degree caused by the final failure of the functional structure is taken as the lower node. The probability reasoning network is constructed according to the upper node-middle node-lower node. During the reasoning process, the probability reasoning network obtains the child node V according to the following formula i The conditional probability of any node in the inference network is obtained by using the joint tree algorithm; In the formula, P(V i ) is the child node V i The parent node, P i Represents the parent node vector, vector value p i Represents vector P i The value of the vector value v i Represents vector V i The value of , thereby obtaining the conditional probability of the child node; also includes: Module 3: updating the satisfaction criterion according to the actual use of the combine harvester.

Images