Cycle time planning method for motor assembly production line based on improved mixed integer linear programming

Abstract

The present invention discloses a motor assembly production line beat planning method based on improved mixed integer linear programming, specifically: mathematically abstract the planning object and parameters of the motor assembly production line beat planning to form a mathematical expression of the motor assembly production line beat planning planning object; determine the input parameters before the motor assembly production line beat planning planning to obtain a priori knowledge variables; solve the motor assembly beat planning in Matlab to obtain a mathematical representation of the motor assembly production line beat planning result based on the improved mixed integer linear programming, and output the motor assembly production line beat planning timing diagram; draw the motor assembly production line beat planning timing table based on the improved mixed integer linear programming according to the motor assembly production line beat planning timing diagram. The present invention can be used for resource scheduling and production line beat planning of the assembly workshop of a motor manufacturing enterprise, can adapt to more different motor assembly production line scenarios, realize the overall solution of the production beat, and realize efficient and high-quality assembly.

Description

Technical Field

[0001] The present invention belongs to the technical field of motor assembly, and in particular relates to a motor assembly production line rhythm planning method based on improved mixed integer linear programming used in a motor assembly automated production line. Background Art

[0002] With the rapid improvement of automation technology, automated processing and assembly have been basically realized in the manufacturing process of motors, especially in the assembly stage. At present, large domestic motor manufacturing companies have basically adopted intelligent automatic assembly production lines based on multi-robot collaboration, visual servo image processing, data acquisition and analysis and other technologies, which has greatly improved the assembly performance of motor products, such as assembly consistency and assembly accuracy. However, due to the limitation of the movement speed of assembly robots and the fact that the assembly rhythm of most automated workshops has not been optimized, the current motor assembly production lines generally have problems such as resource waste, low efficiency, and unstable assembly process, which leads to increased enterprise costs and delayed delivery. Therefore, proposing a rhythm optimization method for motor assembly production lines with waste of resources and low efficiency is of great significance for enterprises to quickly achieve cost reduction and efficiency improvement, and it is also a hot issue currently concerned by the motor industry.

[0003] The patent with application number ZL202111675458.3 discloses a variable beat collaborative control component for a driving motor rotor core shaping production line, including a belt conveyor, an angle servo motor arranged on one side of the belt conveyor, and a beat control dial arranged on the motor shaft of the angle servo motor, the beat control dial including a fan-shaped block vertically fixed to the upper end of the motor shaft for shifting the outer circle of the driving motor rotor core from the rear to the front in the forward direction of the driving motor rotor core, and a cantilever baffle arranged on the fan-shaped block away from the arc outer edge of the fan-shaped block for blocking the forward movement of the outer circle of the driving motor rotor core from the front in the forward direction of the driving motor rotor core. An infrared detector pointing to the side opposite to the fan-shaped block is arranged on the other side of the belt conveyor opposite to the beat control dial. The angle servo motor, the infrared detector and the belt conveyor are respectively connected to the PLC controller, thereby realizing high-reliability variable beat collaborative control of the shaping production line. Although this solution can achieve beat control to a certain extent, it uses a beat control dial composed of a fan-shaped dial block and a cantilever baffle to achieve the shifting and blocking of the rotor core of the drive motor. However, the mixed integer line in the motor assembly automation production line involves many product and component specifications. The beat control method is difficult to effectively meet the assembly beat solution for the mixed assembly line and is difficult to apply to different motor assembly line scenarios. There are also problems such as waste of resources, low efficiency, and unstable assembly process, which lead to increased corporate costs and delayed delivery. Summary of the invention

[0004] In order to solve the problems of existing motor assembly automation production lines, such as waste of production line resources, low efficiency, unstable assembly process, etc., which lead to increased enterprise costs and delayed delivery, the present invention provides a motor assembly production line rhythm planning method based on improved mixed integer linear programming, which can be used for resource scheduling and production line rhythm planning of assembly workshops of motor manufacturing enterprises, can adapt to more different motor assembly production line scenarios, realize an overall solution to the production rhythm, and achieve efficient and high-quality assembly.

[0005] The specific technical solution adopted by the present invention to solve the above technical problems is:

[0006] The motor assembly production line cycle planning method based on improved mixed integer linear programming includes the following steps:

[0007] S100 mathematically abstracts the planning object and its parameters of the motor assembly production line rhythm planning to form a mathematical expression of the planning object of the motor assembly production line rhythm planning;

[0008] S200 determines the input parameters before planning the motor assembly production line cycle time planning, and obtains the prior knowledge variables;

[0009] S300 inputs prior knowledge variables in Matlab;

[0010] S400 uses the YALMIP toolkit and lpsolve solver in Matlab to solve the motor assembly cycle planning, obtains the mathematical representation of the motor assembly production line cycle planning results based on improved mixed integer linear programming, and outputs the motor assembly production line cycle planning timing diagram in Matlab;

[0011] S500 draws a motor assembly production line beat planning timing table based on improved mixed integer linear programming according to the output motor assembly production line beat planning timing diagram.

[0012] Preferably, in step S100, the planning object of the motor assembly production line rhythm planning is the assembly action of the smallest unit, and the planning parameters are: action i action time t i_cal , action i starts at time t i_start , action i ends at time t i_stop , the order relationship r between action i and action j ij_cal ; Among them: if the action i is before the action j after planning, then r ij_cal is 1; if action i is after action j after planning, or action i and action j have no sequential relationship, then r ij_cal is 0.

[0013] Preferably, in step S200, the input parameters before planning are: the time required for each assembly action based on manual experience t i_input, the order relationship r between actions i and j based on artificial experience ij_input , the relationship value x between action i and assembly tool k ki_tool , the interaction relationship value a between action i and action j ij_inter ; Perform the following matrix construction steps according to the input parameters:

[0014] S210 builds the assembly action time matrix T i_input =[t 1_input ,t 2_input ,…,t i_input ,…,t n_input ], t i_input represents the duration required for action i based on experience, and n is the total number of actions;

[0015] S220 builds the assembly action sequence relationship matrix R ij_input , the number in the i-th row and j-th column of the matrix represents the order relationship between action i and action j;

[0016] S230 builds the relationship matrix between assembly actions and assembly tools X ki_tool , the value of the kth row and ith column in the matrix indicates whether action i is performed by tool k;

[0017] S240 builds assembly action interaction matrix A ij_input , the i-th row and j-th column in the matrix represent the interaction relationship between action i and action j;

[0018] Among them: If action i based on artificial experience is before action j, then r ij_input is 1; if action i based on artificial experience comes after action j, or action i and action j have no sequential relationship, then r ij_input is 0; x ik_tool Indicates whether action i is completed by assembly tool k. If action i based on manual experience is performed by tool k, then x ik_tool The value is 1, otherwise it is 0; a ij_inter Indicates whether actions i and j have action interaction, that is, whether actions i and j have parts transferred on different assembly tools. If so, then a ij_inter The value is 1 if yes, otherwise 0.

[0019] Preferably, the prior knowledge variables include the assembly action time matrix T i_input , Assembly action sequence relationship matrix R ij_input , Assembly action and assembly tool relationship matrix X ki_tool , assembly action interaction matrix A ij_input .

[0020] Preferably, step S400 is as follows: using the YALMIP toolkit and the lpsolve solver in Matlab to solve the assembly cycle planning, and obtaining Then, manually judge the result after solving Whether it conforms to the actual assembly scenario, if If it does not conform to the actual assembly scenario, the input parameter values ​​before planning are modified, and the prior knowledge variables are updated, and the execution returns to step S300 to continue, so as to continue to use the YALMIP toolkit and lpsolve solver in Matlab to solve the assembly cycle planning until In line with the actual assembly scenario, the calculated assembly action time matrix T is finally obtained i_cal , Assembly action start time matrix T i_start , Assembly action end time matrix T i_stop , Assembly action sequence relationship matrix R ij_cal , and output the motor assembly production line rhythm planning timing diagram.

[0021] Preferably, the YALMIP toolkit and lpsolve solver are used to obtain When the assembly cycle is optimized, a mathematical method based on mixed integer linear programming is used. The mathematical expressions of the optimization objective function and constraint conditions are as follows:

[0022] Objective function: That is, first find t i_stop The maximum value maxt in i_stop , then maxt i_stop Perform optimization planning to find its minimum value.

[0023] The constraints are:

[0024] t i_start ≥0

[0025] t i_stop -t i_start ≥t i_input

[0026] t j_start ≥t i_stop -(1-r ij_cal )×M, M is an infinite real number

[0027] r ij_cal ∈{0,1}

[0028] r ij_cal ≥r ij_input

[0029]

[0030] Preferably,

[0031] T i_cal =[t 1_cal ,t 2_cal ,…,t i_cal ,…,t n_cal ]

[0032] T i_start =[t 1_start ,t 2_start ,…,t i_start ,…,t n_start ]

[0033] T i_stop =[t 1_stop ,t 2_stop ,…,t i_stop ,…,t n_stop ]

[0034]

[0035] The present invention has the following beneficial effects:

[0036] The present invention is mainly applied to the overall solution of the production rhythm of the motor assembly automation production line within the production line range, involving the timing planning of assembly resources and assembly tool operations. For the resource scheduling and production line rhythm planning of the assembly workshop of the motor manufacturing enterprise, the present invention relies on the operation space configuration and timing planning of the autonomous mobile robot (autonomous driving vehicle) within the enterprise, and relies on enterprise information systems such as enterprise resource planning ERP (Enterprise Resource Planning) and manufacturing execution system MES (Manufacturing Execution System). It can be applied to any motor assembly production line scenario to optimize assembly resources and assembly efficiency, and achieve efficient and high-quality assembly. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flow chart of an embodiment of the method for planning the rhythm of a motor assembly production line based on improved mixed integer linear programming of the present invention.

[0038] Figure 2 It is a flow chart of the timing calculation algorithm of the motor assembly production line rhythm planning method based on improved mixed integer linear programming of the present invention.

[0039] Figure 3 It is a schematic diagram of a motor assembly production line according to the motor assembly production line rhythm planning method based on improved mixed integer linear programming of the present invention.

[0040] Figure 4It is a schematic diagram of the sequence relationship of assembly actions of the motor assembly production line rhythm planning method based on improved mixed integer linear programming of the present invention.

[0041] Figure 5 It is an assembly rhythm timing diagram of the motor assembly production line rhythm planning method based on improved mixed integer linear programming of the present invention after the motor assembly production line rhythm planning. DETAILED DESCRIPTION

[0042] Combine the following Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 The present invention is further described with Table 1.

[0043] like Figure 1 As shown, the present invention is based on the motor assembly production line beat planning method of improved mixed integer linear programming, which is as follows:

[0044] S100 mathematically abstracts the planning object and its parameters of the motor assembly production line rhythm planning to form a mathematical expression of the planning object of the motor assembly production line rhythm planning;

[0045] S200 determines the prior knowledge of the cycle planning of the motor assembly production line, that is, the input parameters before planning, and obtains the prior knowledge variables;

[0046] S300 inputs prior knowledge variables in Matlab;

[0047] S400 uses the YALMIP toolkit and lpsolve solver in Matlab to solve the assembly cycle planning, obtains the mathematical representation of the motor assembly production line cycle planning results based on improved mixed integer linear programming, and outputs the motor assembly production line cycle planning timing diagram in Matlab;

[0048] S500 draws a motor assembly production line beat planning timing table based on improved mixed integer linear programming according to the output motor assembly production line beat planning timing diagram.

[0049] S100 specifically includes: decomposing the motor assembly process into the smallest action units, which are the planning objects of the motor assembly production line rhythm planning proposed by the present invention. The planning parameters of the action units are represented by mathematical symbols: action time t i_cal , action i starts at time t i_start , action i ends at time t i_stop , the order relationship r between action i and action j ij_cal ;

[0050] Among them, r ij_calIt is a 0-1 variable, which indicates the assembly order relationship between action i and action j after calculation. If action i is before action j after calculation, the value is 1; if action i is after action j, or there is no order relationship between action i and action j, the value is 0.

[0051] In the above step S200, the input parameters before planning are: the time required for each assembly action based on manual experience t i_input , the order relationship r between actions i and j based on artificial experience ij_input , the relationship value x between action i and assembly tool k ki_tool , the interaction relationship value a between action i and action j ij_inter ; Perform the following matrix construction steps according to the input parameters:

[0052] S210 builds the assembly action time matrix T i_input =[t 1_input ,t 2_input ,…,t i_input ,…,t n_input ], t i_input represents the duration required for action i based on experience, and n is the total number of actions;

[0053] S220 builds the assembly action sequence relationship matrix R ij_input , the number in the i-th row and j-th column of the matrix represents the order relationship between action i and action j;

[0054] S230 builds the relationship matrix between assembly actions and assembly tools X ki_tool , the value of the kth row and ith column in the matrix indicates whether action i is performed by tool k;

[0055] S240 builds assembly action interaction matrix A ij_input , the i-th row and j-th column in the matrix represent the interaction relationship between action i and action j.

[0056] like Figure 2 As shown, in the above step S300, the prior knowledge variables are input into Matlab, including the assembly action time matrix T i_input , Assembly action sequence relationship matrix R ij_input , Assembly action and assembly tool relationship matrix X ki_tool and assembly action interaction matrix A ij_input In the above step S400, the YALMIP toolkit and lpsolve solver are used in Matlab to solve the assembly cycle planning, and the result is That is Then, the result after manual judgment is Whether it conforms to the actual assembly scenario, if If it does not conform to the actual assembly scenario, the input parameter values ​​before planning are modified, and the prior knowledge variables are updated, and the execution returns to step S300 to continue, so as to continue to use the YALMIP toolkit and lpsolve solver in Matlab to solve the assembly cycle planning until In line with the actual assembly scenario, the calculated assembly action time matrix T is finally obtained i_cal , Assembly action start time matrix T i_start , Assembly action end time matrix T i_stop , Assembly action sequence relationship matrix R ij_cal , and output the motor assembly production line rhythm planning timing diagram.

[0057] Among them, the YALMIP toolkit and lpsolve solver are used to obtain When the assembly cycle is optimized, a mathematical method based on mixed integer linear programming is used. The mathematical expressions of the optimization objective function and constraint conditions are as follows:

[0058] Objective function: That is, first find t i_stop The maximum value maxt in i_stop , the last assembly action in the time sequence, and then maxt i_stop Perform optimization planning to find its minimum value.

[0059] The constraints are:

[0060] t i_start ≥0

[0061] t i_stop -t i_start ≥t i_input

[0062] t j_start ≥t i_stop -(1-r ij_cal )×M, M is an infinite real number

[0063] r ij_cal ∈{0,1}

[0064] r ij_cal ≥r ij_input

[0065]

[0066] Taking the motor assembly production line and its planning objects and parameters recorded in Table 1 as an example, a tact planning timing table of the motor assembly production line based on improved mixed integer linear programming is drawn out by the method of the present invention.

[0067] Table 1 Assembly cycle time table after motor assembly production line cycle planning

[0068] Step code subject predicate object Start time End time Action duration P1 robot Move to Parts box location 0 2 2 P2 robot Crawl Component 2 7 5 P3 robot Move to Feeding position 7 12 5 P4 robot Reset to Initial position 16 21 5 P5 Assembly robot Crawl Component 12 16 4 P6 Assembly robot Move to Position 23 26 3 P7 Assembly robot assembly Component 26 31 5 P8 Alignment module Move to Position 0 2 2 P9 Alignment module Counterpoint Component 16 19 3 P10 Alignment module Reset to Initial position 23 26 3 P11 Micro motion platform Adjustment Component 19 23 4

[0069] like Figure 3 As shown, the motor assembly production line includes three modules: a loading module 02 (including a robot and an assembly manipulator), an alignment module 03 and a micro-adjustment module 04 (micro-adjustment platform). The loading module is used to move the parts in the parts area 01 to the specified position, the alignment module uses vision for alignment, and the micro-adjustment module adjusts the posture according to the solution results.

[0070] In Table 1, the assembly actions used in the motor assembly production line have a sequence relationship. These sequence relationships can be reflected in the timing diagram. For example, if action 1 and action 2 have a deterministic sequence relationship, and action 1 is executed before action 2, then in the timing diagram, action 1 is before action 2. Figure 4 As shown, if action 1 and action 5 have no deterministic sequential relationship, then in the timing diagram, action 1 and action 5 can be executed simultaneously.

[0071] like Figure 5 The present invention takes the motor assembly production line and its planning objects and parameters recorded in Table 1 as an example to plan the assembly beat timing chart. The present invention takes the motor assembly production line and its planning objects and parameters recorded in Table 1 as an example to plan the assembly beat timing chart is also recorded in Table 1.

[0072] The matrix appearing in the present invention is explained below:

[0073] 1. Assembly action time matrix T i_input :If there are n actions in the assembly, the assembly action time is expressed by a one-dimensional matrix T i_input =[t 1_input ,t 2_input ,…,t i_input ,…,t n_input ] to indicate.

[0074] 2. Assembly action sequence relationship matrix R ij_input : Assume that there are n actions in the assembly process of a part. Since action i has no sequential relationship with itself, the matrix diagonal is empty. The number in the i-th row and j-th column in the matrix represents the sequential relationship between action i and action j.

[0075]

[0076] 3. Assembly action and assembly tool relationship matrix X ki_tool :If x ki_tool =1, it means that the tool for action i is k; if x ki_tool = 0, it means that the tool for action i is not k. Using the matrix X ki_toolIndicates the relationship between an action and a tool.

[0077]

[0078] K is the total number of tools.

[0079] 4. Assembly Action Interaction Matrix A ij_input : During the assembly process of the motor assembly line, the two actions that interact with each other must have a deterministic sequence relationship. The matrix A ij_input Represents the interaction between any actions in the production line. The number in the i-th row and j-th column represents the interaction between actions i and j. ij_inter =1, indicating that action i and action j have a sequential relationship, and action i comes before action j. Since action i has no sequential relationship with itself, the matrix diagonal is empty.

[0080]

[0081] The above four matrices are prior knowledge variables.

[0082] 5. Planning the assembly action time matrix T i_cal : is a one-dimensional matrix with n elements.

[0083] T i_cal =[t 1_cal ,t 2_cal ,…,t i_cal ,…,t n_cal ]

[0084] 6. Planning the assembly action start time matrix T i_start : It is the start time matrix of each assembly action after being arranged in sequence, and is a one-dimensional matrix with n elements.

[0085] T i_start =[t 1_start ,t 2_start ,…,t i_start ,…,t n_start ]

[0086] 7. Planning the end time matrix of assembly action T i_stop : It is the end time matrix of each assembly action after being arranged in sequence, and is a one-dimensional matrix with n elements.

[0087] T i_stop =[t 1_stop ,t 2_stop ,…,t i_stop ,…,t n_stop ]

[0088] 8. Planning the assembly action sequence relationship matrix Rij_cal : expresses the assembly order relationship between action i and action j after planning. The number in the i-th row and j-th column of the matrix represents the assembly action order relationship between action i and action j. ij_cal , when action i precedes action j, r ij_cal =1; when action i comes after action j or the order of i and j is uncertain, then r ij_cal =0.

[0089]

[0090] Those skilled in the art should understand that the above examples and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may be subject to various changes and improvements, which are all within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the appended claims.

Claims

1. A motor assembly production line cycle planning method based on improved mixed integer linear programming, characterized by: The steps include: S100 mathematically abstracts the planning object and its parameters of the motor assembly production line rhythm planning to form a mathematical expression of the planning object of the motor assembly production line rhythm planning; S200 determines the input parameters before planning the motor assembly production line cycle time planning, and obtains the prior knowledge variables; S300 inputs prior knowledge variables in Matlab; S400 uses the YALMIP toolkit and lpsolve solver in Matlab to solve the motor assembly cycle planning, obtains the mathematical representation of the motor assembly production line cycle planning results based on improved mixed integer linear programming, and outputs the motor assembly production line cycle planning timing diagram in Matlab; S500 draws a motor assembly production line beat planning timing table based on improved mixed integer linear programming according to the output motor assembly production line beat planning timing diagram; The specific steps of S400 are as follows: Use the YALMIP toolkit and lpsolve solver in Matlab to solve the assembly cycle planning and obtain Then, manually judge the result after solving Whether it conforms to the actual assembly scenario, if If it does not conform to the actual assembly scenario, the input parameter values ​​before planning are modified, and the prior knowledge variables are updated, and the execution returns to step S300 to continue, so as to continue to use the YALMIP toolkit and lpsolve solver in Matlab to solve the assembly cycle planning until In line with the actual assembly scenario, the calculated assembly action time matrix T is finally obtained i_cal , Assembly action start time matrix T i_start , Assembly action end time matrix T i_stop , Assembly action sequence relationship matrix R ij_cal , and output the timing diagram of the motor assembly production line rhythm planning; Using the YALMIP toolkit and the lpsolve solver, we can obtain When the assembly cycle is optimized, a mathematical method based on mixed integer linear programming is used. The mathematical expressions of the optimization objective function and constraint conditions are as follows: Objective function: That is, first find t i_stop The maximum value maxt in i_stop , then maxt i_stop Carry out optimization planning to find its minimum value; The constraints are: t i_start ≥0 t i_stop -t i_start ≥t i_input t j_start ≥t i_stop -(1-r ij_cal )×M, where M is an infinite real number r ij_cal ∈{0,1} r ij_cal ≥r ij_input K is the total number of tools.

2. The motor assembly production line tact planning method based on improved mixed integer linear programming according to claim 1 is characterized in that: In step S100, the planning object of the motor assembly production line rhythm planning is the assembly action of the smallest unit, and the planning parameters are: action i action time t i_cal , action i starts at time t i_start , action i ends at time t i_stop , the order relationship r between action i and action j ij_cal ; Among them: if the action i is before the action j after planning, then r ij_cal is 1; if action i is after action j after planning, or action i and action j have no sequential relationship, then r ij_cal is 0.

3. The motor assembly production line tact planning method based on improved mixed integer linear programming according to claim 1 is characterized in that: In step S200, the input parameters before planning are: the time required for each assembly action based on manual experience, t i_input , the order relationship r between action i and action j based on artificial experience ij_input , the relationship value x between action i and assembly tool k ki_tool , the interaction relationship value a between action i and action j ij_inter ; Perform the following matrix construction steps according to the input parameters: S210 builds the assembly action time matrix T i_input =[t 1_input ,t 2_input ,…,t i_input ,…,t n_input ], t i_input represents the duration required for action i based on experience, and n is the total number of actions; S220 builds the assembly action sequence relationship matrix R ij_input , the number in the i-th row and j-th column of the matrix represents the order relationship between action i and action j; S230 builds the relationship matrix between assembly actions and assembly tools X ki_tool , the value of the kth row and ith column in the matrix indicates whether action i is performed by tool k; S240 builds assembly action interaction matrix A ij_input , the i-th row and j-th column in the matrix represent the interaction relationship between action i and action j; Among them: If action i based on artificial experience is before action j, then r ij_input is 1; if action i based on artificial experience comes after action j, or action i and action j have no sequential relationship, then r ij_input is 0; x ik_tool Indicates whether action i is completed by assembly tool k. If action i based on manual experience is performed by tool k, then x ik_tool The value is 1, otherwise it is 0; a ij_inter Indicates whether actions i and j have action interactions. If so, then a ij_inter The value is 1 otherwise, 0.

4. The motor assembly production line tact planning method based on improved mixed integer linear programming according to claim 1 is characterized in that: The prior knowledge variables include the assembly action time matrix T i_input , Assembly action sequence relationship matrix R ij_input , Assembly action and assembly tool relationship matrix X ki_tool , assembly action interaction matrix A ij_input .

5. The motor assembly production line tact planning method based on improved mixed integer linear programming according to claim 1 is characterized in that: T i_cal =[t 1_cal ,t 2_cal ,…,t i_cal ,…,t n_cal ] T i_start =[t 1_start ,t 2_start ,…,t i_start ,…,t n_start ] T i_stop =[t 1_stop ,t 2_stop ,…,t i_stop ,…,t n_stop ]

Images