At present, tool selection mainly relies on the knowledge and experience of technicians and the selection of relevant manuals and specifications. This is not only inefficient, but also restricted by personnel. A large amount of historical information cannot be reused, and information between departments cannot be shared. With the continuous emergence of new tools, new materials and new processes, how to quickly select economical and reasonable tools has become an urgent problem for engineers and technicians [1].
In this regard, domestic and foreign scholars have carried out relevant theoretical research, mainly using case-based reasoning and rule-based reasoning for tool intelligent matching. However, it is very difficult to extract the tool selection rules from the experience of technicians, related manuals and specifications, and the number of tool selection rules is very large, the matching and retrieval efficiency is too low, and the maintenance of tool selection rules is very difficult. In contrast, case libraries are easier to construct, maintain, and reasoning faster and easier to learn. However, as the tool selection scheme in the case library is continuously added, the efficiency of retrieving the tool selection history scheme will also decrease [2-3].
Aiming at the above problems, this paper proposes a tool intelligent matching method based on case-based reasoning and rule-based reasoning. The rough set theory is used to optimize the tool matching instance library, and the tool matching history plan is reduced and the redundancy is eliminated. The data information improves the efficiency of the search tool matching history scheme, and uses rule reasoning as a supplement to the case reasoning to enhance the practicability of the tool intelligent matching.
Knowledge representation of tool intelligent selection
1 Classification of basic information of tool intelligent matching
The basic information of the tool intelligent selection includes machining workpiece and tool information. The matching process must classify the two types of information to facilitate the inductive extraction of information and provide basic data support for tool intelligent matching.
When the tool is actually selected, the main feature information of the machined workpiece needs to be provided for matching with the tool selection history plan and the matching rule. According to the different characteristics expressed, the information of the processed workpiece is divided into materials, dimensions, precision, processing technology and the like.
The tool selection depends largely on the process used in the machining, and the machining process information is also included in the machining workpiece information. In order to better match the workpiece information, the tools can be classified according to the cutting process, including milling tools, turning tools, boring tools, drilling tools, etc. For milling tools, for example, milling tools can be divided into vertical milling. Knives, face milling cutters, forming cutters, keyway milling cutters, etc.
2 Knowledge representation of the tool intelligent matching example
The tool intelligent matching example is a record of the tool selection history plan, which is the basis of the example reasoning part (should include the machining requirement description and the specific selection tool). In order to more fully represent the tool matching example, it is convenient to optimize the subsequent matching example library. The matching should use the rough set decision table to represent the tool history plan. Due to the differences in the shape and processing technology of different raw materials, the attributes in the tool selection decision table are inconsistent, so it is necessary to first classify the major categories according to the shape and processing technology of the raw materials. This kind of tool matching case is used as the domain of the knowledge representation system. The basic information part of the workpiece and the tool matching result constitute the condition attribute set and the result attribute set respectively. Take the slot machining tool of the bar as an example, as shown in Table 1.
In the tool intelligent selection, the tool selection rules are mainly extracted from the experience of the craftsman, relevant manuals and specifications, and a large amount of research, collection and analysis work is required. The process of tool rule matching simulates the expert's matching process. There is a causal relationship between the front and back. It is suitable to use the production knowledge representation to represent the tool matching rules. The basic form is IF A THEN B, where A is the prerequisite for tool matching. Condition, B is the tool matching result [4]. In order to avoid ambiguity in the tool matching process, the preconditions and matching results of the tool matching can only be combined; in order to express the tool matching rules more clearly, a relational database can be used to represent Table 3 is an example library that constitutes the intelligent selection of tools. One of the rules is: IF workpiece material = cast iron ∧ machining characteristics = outer circle THEN tool material = carbide ∧ tool type = turning tool.
1 Instance optimization of tool intelligent selection
The tool library intelligent matching example library is a collection of past tool matching results and expert experience, and is the basis of tool matching case reasoning, but with the continuous expansion of tool matching history cases in the case library, the number of cases will be It will become very large, and the efficiency of retrieving the appropriate tool matching scheme will be significantly reduced. It is necessary to optimize the tool selection example library, eliminate the redundant information in the tool matching scheme, and improve the tool matching efficiency.
This topic uses the rough set theory to optimize the tool selection example library. Firstly, the data should be completed in the data acquisition of the tool matching example. The matching examples include the workpiece material, machining characteristics, workpiece size, tool number and other information. It covers string type, floating point type, integer type and other data types. This requires that the tool selection instance library must be discretized before the attribute reduction and value reduction of the tool selection instance library can be performed. . In order to avoid the data loss in the tool selection decision table due to discretization, and affect the correctness of the tool matching, the discretization algorithm based on attribute importance is used to discretize the tool matching instance library. Table 1 is taken as an example. The discretization process can obtain the data of Table 4. The set of conditional attribute values ​​of the tool selection decision table are {carbon steel, stainless steel, aluminum alloy}, {18, 20}, {4, 5}, {4, 5.2}, {1.6, 6.3}, the elements in the set are numbered 0, 1, or 2, respectively, in order.
The influence of the type, size and surface roughness of the workpiece material on the tool selection is inconsistent in the tool matching scheme. This requires the reduction of the redundant condition attribute of the tool selection, so the recognition-based matrix is ​​adopted. And the attribute reduction algorithm of logical operation [5], can maintain the dependency relationship between the condition attribute and the result attribute of the tool selection decision table does not change.
The specific steps of the algorithm are as follows:
(1) Calculate the discernible matrix CD of the tool selection decision table;
(2) For all the elements C ij(C ij ≠0, C ij ≠φ ) whose values ​​are non-empty sets in the discernible matrix, the corresponding disjunction logic expression Li j = ∨ai∈Ci jai is established;
(3) Combine all the disjunctive logic expressions L ij to obtain a conjunction paradigm L = ∧ Ci j0Ci j Li j ;
(4) Convert the conjunction paradigm L into the form of the disjunction paradigm, and obtain L= ∨iLi;
(5) Output tool selection decision table attribute reduction result.
In the actual selection of the tool used for the bar groove machining, the workpiece diameter and the groove depth have little effect on the tool selection. Table 4 uses the above reduction algorithm to obtain Table 5, which introduces the workpiece diameter and groove depth as redundant attributes, and the tool. The actual matching situation is the same.
The tool intelligent matching is mainly based on case-based reasoning. When the workpiece processing information is entered for historical solution retrieval, the paper first classifies according to the shape and processing technology of the processing material, finds the corresponding index in the tool matching instance library, and then finds through the index. Corresponding tool selection decision table, compare with the reduced tool selection decision table, and retrieve the qualified tool matching case; if there is no matching historical solution, the rule inference part of the tool matching carry out.
The number of cases in the tool selection example library determines its ability to provide a solution, and it is necessary to continuously expand the tool selection history plan library during use. The current workpiece information and tool selection results can be modified directly or appropriately as new cases and can then be added to the tool selection example library. However, the addition of a new tool selection scheme must be confirmed and controlled by setting permissions. Because the new tool matching scheme may conflict with the historical matching scheme, the experienced craftsman will choose the solution and remove the conflict matching scheme; or give up the new matching scheme, or modify the matching scheme again. Optional case addition. After the tool's new matching scheme is added to the tool matching instance library, the system will automatically optimize the tool matching instance library, update and generate a new tool matching reduction decision table to ensure real-time update of system data.
3 Reasoning and design of tool intelligent matching
In the tool intelligent matching process, rule reasoning is supplemented by case reasoning. When using forward reasoning strategy to select a tool for a machining situation, there is no matching historical solution, then the tool matching rule is matched, according to the workpiece information and The rules in the tool selection rule base are matched one by one, and the tools that meet the requirements are selected. In order to avoid the ambiguity of the interpretation of the tool selection rule, the premise and conclusion of the matching rule can only be the conjunction operation.
In the process of matching the machining information with the tool selection rule, if a machining condition is successfully matched with more than two tool selection rules, a tool selection rule is selected as the activation rule according to the conflict resolution strategy until all the tools are matched. The matching rule gives the tool matching result and adds the new matching plan to the historical plan library according to the tool matching plan addition method. If there is no tool matching rule that matches the machining information, it prompts to improve the tool selection rule base, expand the tool selection rules, and enhance the ability of the tool intelligent matching rule reasoning.
Conclusion
The intelligent selection of tools based on rough set realizes the fast and accurate matching of tools, avoids the inspection of a large amount of data and tool selection errors, is no longer subject to the experience of the craftsman, improves the work efficiency, and enables the historical solution to be reused. Tool matching options can be shared between departments, so they have important practical value. (end)
In this regard, domestic and foreign scholars have carried out relevant theoretical research, mainly using case-based reasoning and rule-based reasoning for tool intelligent matching. However, it is very difficult to extract the tool selection rules from the experience of technicians, related manuals and specifications, and the number of tool selection rules is very large, the matching and retrieval efficiency is too low, and the maintenance of tool selection rules is very difficult. In contrast, case libraries are easier to construct, maintain, and reasoning faster and easier to learn. However, as the tool selection scheme in the case library is continuously added, the efficiency of retrieving the tool selection history scheme will also decrease [2-3].
Aiming at the above problems, this paper proposes a tool intelligent matching method based on case-based reasoning and rule-based reasoning. The rough set theory is used to optimize the tool matching instance library, and the tool matching history plan is reduced and the redundancy is eliminated. The data information improves the efficiency of the search tool matching history scheme, and uses rule reasoning as a supplement to the case reasoning to enhance the practicability of the tool intelligent matching.
Knowledge representation of tool intelligent selection
1 Classification of basic information of tool intelligent matching
The basic information of the tool intelligent selection includes machining workpiece and tool information. The matching process must classify the two types of information to facilitate the inductive extraction of information and provide basic data support for tool intelligent matching.
When the tool is actually selected, the main feature information of the machined workpiece needs to be provided for matching with the tool selection history plan and the matching rule. According to the different characteristics expressed, the information of the processed workpiece is divided into materials, dimensions, precision, processing technology and the like.
The tool selection depends largely on the process used in the machining, and the machining process information is also included in the machining workpiece information. In order to better match the workpiece information, the tools can be classified according to the cutting process, including milling tools, turning tools, boring tools, drilling tools, etc. For milling tools, for example, milling tools can be divided into vertical milling. Knives, face milling cutters, forming cutters, keyway milling cutters, etc.
2 Knowledge representation of the tool intelligent matching example
The tool intelligent matching example is a record of the tool selection history plan, which is the basis of the example reasoning part (should include the machining requirement description and the specific selection tool). In order to more fully represent the tool matching example, it is convenient to optimize the subsequent matching example library. The matching should use the rough set decision table to represent the tool history plan. Due to the differences in the shape and processing technology of different raw materials, the attributes in the tool selection decision table are inconsistent, so it is necessary to first classify the major categories according to the shape and processing technology of the raw materials. This kind of tool matching case is used as the domain of the knowledge representation system. The basic information part of the workpiece and the tool matching result constitute the condition attribute set and the result attribute set respectively. Take the slot machining tool of the bar as an example, as shown in Table 1.
In the tool intelligent selection, the tool selection rules are mainly extracted from the experience of the craftsman, relevant manuals and specifications, and a large amount of research, collection and analysis work is required. The process of tool rule matching simulates the expert's matching process. There is a causal relationship between the front and back. It is suitable to use the production knowledge representation to represent the tool matching rules. The basic form is IF A THEN B, where A is the prerequisite for tool matching. Condition, B is the tool matching result [4]. In order to avoid ambiguity in the tool matching process, the preconditions and matching results of the tool matching can only be combined; in order to express the tool matching rules more clearly, a relational database can be used to represent Table 3 is an example library that constitutes the intelligent selection of tools. One of the rules is: IF workpiece material = cast iron ∧ machining characteristics = outer circle THEN tool material = carbide ∧ tool type = turning tool.
1 Instance optimization of tool intelligent selection
The tool library intelligent matching example library is a collection of past tool matching results and expert experience, and is the basis of tool matching case reasoning, but with the continuous expansion of tool matching history cases in the case library, the number of cases will be It will become very large, and the efficiency of retrieving the appropriate tool matching scheme will be significantly reduced. It is necessary to optimize the tool selection example library, eliminate the redundant information in the tool matching scheme, and improve the tool matching efficiency.
This topic uses the rough set theory to optimize the tool selection example library. Firstly, the data should be completed in the data acquisition of the tool matching example. The matching examples include the workpiece material, machining characteristics, workpiece size, tool number and other information. It covers string type, floating point type, integer type and other data types. This requires that the tool selection instance library must be discretized before the attribute reduction and value reduction of the tool selection instance library can be performed. . In order to avoid the data loss in the tool selection decision table due to discretization, and affect the correctness of the tool matching, the discretization algorithm based on attribute importance is used to discretize the tool matching instance library. Table 1 is taken as an example. The discretization process can obtain the data of Table 4. The set of conditional attribute values ​​of the tool selection decision table are {carbon steel, stainless steel, aluminum alloy}, {18, 20}, {4, 5}, {4, 5.2}, {1.6, 6.3}, the elements in the set are numbered 0, 1, or 2, respectively, in order.
The influence of the type, size and surface roughness of the workpiece material on the tool selection is inconsistent in the tool matching scheme. This requires the reduction of the redundant condition attribute of the tool selection, so the recognition-based matrix is ​​adopted. And the attribute reduction algorithm of logical operation [5], can maintain the dependency relationship between the condition attribute and the result attribute of the tool selection decision table does not change.
The specific steps of the algorithm are as follows:
(1) Calculate the discernible matrix CD of the tool selection decision table;
(2) For all the elements C ij(C ij ≠0, C ij ≠φ ) whose values ​​are non-empty sets in the discernible matrix, the corresponding disjunction logic expression Li j = ∨ai∈Ci jai is established;
(3) Combine all the disjunctive logic expressions L ij to obtain a conjunction paradigm L = ∧ Ci j0Ci j Li j ;
(4) Convert the conjunction paradigm L into the form of the disjunction paradigm, and obtain L= ∨iLi;
(5) Output tool selection decision table attribute reduction result.
In the actual selection of the tool used for the bar groove machining, the workpiece diameter and the groove depth have little effect on the tool selection. Table 4 uses the above reduction algorithm to obtain Table 5, which introduces the workpiece diameter and groove depth as redundant attributes, and the tool. The actual matching situation is the same.
The tool intelligent matching is mainly based on case-based reasoning. When the workpiece processing information is entered for historical solution retrieval, the paper first classifies according to the shape and processing technology of the processing material, finds the corresponding index in the tool matching instance library, and then finds through the index. Corresponding tool selection decision table, compare with the reduced tool selection decision table, and retrieve the qualified tool matching case; if there is no matching historical solution, the rule inference part of the tool matching carry out.
The number of cases in the tool selection example library determines its ability to provide a solution, and it is necessary to continuously expand the tool selection history plan library during use. The current workpiece information and tool selection results can be modified directly or appropriately as new cases and can then be added to the tool selection example library. However, the addition of a new tool selection scheme must be confirmed and controlled by setting permissions. Because the new tool matching scheme may conflict with the historical matching scheme, the experienced craftsman will choose the solution and remove the conflict matching scheme; or give up the new matching scheme, or modify the matching scheme again. Optional case addition. After the tool's new matching scheme is added to the tool matching instance library, the system will automatically optimize the tool matching instance library, update and generate a new tool matching reduction decision table to ensure real-time update of system data.
3 Reasoning and design of tool intelligent matching
In the tool intelligent matching process, rule reasoning is supplemented by case reasoning. When using forward reasoning strategy to select a tool for a machining situation, there is no matching historical solution, then the tool matching rule is matched, according to the workpiece information and The rules in the tool selection rule base are matched one by one, and the tools that meet the requirements are selected. In order to avoid the ambiguity of the interpretation of the tool selection rule, the premise and conclusion of the matching rule can only be the conjunction operation.
In the process of matching the machining information with the tool selection rule, if a machining condition is successfully matched with more than two tool selection rules, a tool selection rule is selected as the activation rule according to the conflict resolution strategy until all the tools are matched. The matching rule gives the tool matching result and adds the new matching plan to the historical plan library according to the tool matching plan addition method. If there is no tool matching rule that matches the machining information, it prompts to improve the tool selection rule base, expand the tool selection rules, and enhance the ability of the tool intelligent matching rule reasoning.
Conclusion
The intelligent selection of tools based on rough set realizes the fast and accurate matching of tools, avoids the inspection of a large amount of data and tool selection errors, is no longer subject to the experience of the craftsman, improves the work efficiency, and enables the historical solution to be reused. Tool matching options can be shared between departments, so they have important practical value. (end)
Concerned about surprises
Tag: Tool Selection Tool Material Rough Set Theory Reduction
Previous: CNC lathe tool requirements and main tool type Next: Daily maintenance of kitchen hardware
A Shower Faucet (also spigot or tap: see usage variations) is a valve controlling the release of a liquid that enabling washing and drying in the bathroom.
Shower Faucet,Single Handle Shower Faucet,Single Lever Bathtub Faucet,Hot Cold Water Shower Head
Kaiping Jenor Sanitary Ware Co., Ltd , https://www.jenorsanitary.com