JP4005595B2 - Component mounting order determination method, component mounting order determination device, component mounting machine, and program - Google Patents

Description

  The present invention relates to a component mounting order determination method in a component mounter that mounts components on a substrate.

  In mounting components on a board by a component mounting machine, the mounting order of components must be determined prior to mounting the components. In the component mounting order determination method, the component mounting order is determined for the purpose of minimizing the time from the start of mounting the components on one board to the end of mounting all the components. However, in the brute force method that generates all possible component mounting orders and finds the optimal component mounting order, the processing time for determining the optimal component mounting order increases as the number of component mounting points increases. It is difficult to determine the optimal component mounting order in a realistic time.

For this reason, a method using a greedy method has been proposed as a method for determining an optimal component mounting order in a realistic time (see, for example, Patent Document 1).
Japanese Patent Laid-Open No. 7-176891

  However, in the component mounting order determination method using the greedy method, component mounting points with the lowest cost (for example, the distance between the component mounting points) are sequentially selected to determine the component mounting order. For this reason, the mounting order is such that the moving distance between the component mounting points becomes small at the start of component mounting, but the moving distance between the component mounting points increases as the component mounting ends. There is a problem that disturbance (useless operation) occurs, and there is a problem that an optimal mounting order is not necessarily obtained.

  The present invention has been made to solve the above-described problems, and an object thereof is to provide a component mounting order determination method capable of determining an optimal component mounting order within a realistic time.

  In order to achieve the above object, a component mounting order determination method according to the present invention is a method for determining a component mounting order in a component mounter that mounts a plurality of components on a board, and includes a plurality of component mounting points and locations. For any two fixed points, the distance between the two points or the time required to mount a component at one point between the two points until the component is mounted at the other point. The cost between any component mounting point is larger than the cost between any two of the plurality of component mounting points and the predetermined points, and the cost between the predetermined points is plural. A dummy point that is smaller than the cost between any two of the component mounting points and the predetermined points is defined, positional information of the plurality of component mounting points, the plurality of component mounting points, the predetermined points, and the Get the cost for dummy points A component mounting point information acquisition step, and a traveling salesman problem for a predetermined route on the board, a dummy point, and a plurality of component mounting points that pass through each point that minimizes the sum of costs. A tour route calculating step to calculate the components, and a mounting order determining step to determine the mounting order of the components based on the tour route.

  The traveling salesman problem is used to determine the component mounting order. For this reason, the problem that the movement of the mounting points (useless operation) occurs as the component mounting end that occurs when the component mounting order is determined by the greedy method is solved, and the optimum mounting order is obtained. be able to. In addition, the traveling salesman problem can be obtained an approximate exact solution in a realistic processing time. Note that, by defining the dummy points as described above, a circuit that always connects the dummy points and the predetermined points is generated. For this reason, it is possible to generate a closed circuit having a predetermined point as a start point. Note that the component mounting order may be determined in advance by the greedy method, and the traveling salesman problem may be applied to determine the optimal component mounting order in order to improve the above-described useless operation.

  The component mounting order determination method described above further includes a cost calculating step of calculating a cost between any two of the predetermined points on the substrate, the dummy points, and the plurality of component mounting points, In the traveling route calculation step, using the cost calculated in the cost calculation step, for each of the predetermined points on the board, the dummy points, and the plurality of component mounting points, the points that minimize the sum of the costs are determined. The traveling route may be calculated using the traveling salesman problem. In the cost calculating step, the cost from the first point to the second point is considered to be equal to the cost from the second point to the first point, and the cost between any two points is calculated. It may be.

  By considering that the cost from the first point to the second point is equal to the cost from the second point to the first point, the cost calculation time can be reduced by half.

  The present invention can be realized not only as a component mounting order determination method having such characteristic steps, but also as a component mounting order determination device using the characteristic steps included in the component mounting order determination method. Or as a program that causes a computer to execute the characteristic steps included in the component mounting order determination method. Needless to say, such a program can be distributed via a recording medium such as a CD-ROM (Compact Disc-Read Only Memory) or a communication network such as the Internet.

  It is possible to provide a component mounting order determination method capable of determining an optimal component mounting order within a realistic time.

A component mounting system according to an embodiment of the present invention will be described.
FIG. 1 is an external view of a component mounting system according to an embodiment of the present invention.

  The component mounting system is a system that determines the mounting order of components on a board and mounts components on the board in accordance with the determined mounting order. The component mounting system 10 and the component mounting order determination apparatus 300 include I have.

  The component mounter 10 is a mounting device that mounts a plurality of types of components on a printed circuit board constituting an electronic device at high speed. The component mounter 10 sucks, conveys, and mounts components, and the rotary head 100 mounts various types of components. And an XY table 15 for moving the placed printed board in the horizontal plane direction.

FIG. 2 is a schematic diagram illustrating a positional relationship between the component supply unit and the rotary head.
As shown in the upper part of FIG. 2, the rotary head 100 includes 18 mounting heads 103 as mounting means for mounting components on a printed circuit board. The mounting head 103 is attached to a rotating base 101 that rotates without moving in the height direction so as to be movable in the height direction, and can hold a component by vacuum suction (not shown). 6 are provided.

  As shown in the lower part of FIG. 2, the component supply unit 11 includes component cassettes 111 that can sequentially provide the same components to the mounting head 103 in a horizontal row. The component supply unit 11 has a function of selecting a component to be mounted by moving and positioning the component supply unit 11 with respect to the rotary head 100 in the Z-axis direction in FIG.

  The component cassette 111 includes a reel 113 around which a component tape 110 that holds a large number of the same components as shown in FIG. 3 is wound, and a tape feeder 114 that sequentially supplies the components held on the component tape 110. The component supply unit 11 is detachable. That is, the component cassette 111 corresponds to a component that provides one type of component, and the component supply unit 11 corresponds to a set of component cassettes 111 (or a storage unit that holds the set).

  FIG. 4 is a diagram schematically illustrating the positional relationship among the rotary head, the substrate, and the component supply unit.

  As shown in the figure, the rotary shaft of the rotary head 100 does not move, and the mounting head 103 provided around the rotary shaft rotates intermittently to perform work corresponding to each position. Briefly, the mounting head 103 located at the upper portion (position B) of the suction opening 112 provided in each of the component cassettes 111 sucks the component through the suction opening 112, and the mounting head 103 is located at a position E facing the mounting head 103. The components that are attracted to the printed circuit board P when placed are mounted.

  Note that the printed circuit board P to which the component is to be mounted is placed on an XY table (not shown) that is movable in the horizontal plane direction, and the position where the component is to be mounted is determined by moving the printed circuit board P. The

FIG. 5 is a diagram illustrating an example of a component recognition camera provided at a position C in FIG.
The component recognition camera 105 is a camera for recognizing the component 106 attracted by the mounting head 103, and includes a lens barrel 105a, a small-field camera 105b, and a large-field camera 105c. In the lens barrel 105a, mirrors 107a and 107c and a half mirror 107b are provided.

  The small-field camera 105b is a camera having a field-of-view size of 6 mm, and images the part 106 reflected by the mirror 107a and the half mirror 107b. The large-field camera 105c is a camera having a field-of-view size of 28 mm, and images the part 106 reflected by the mirrors 107a and 107c.

  FIG. 6 is a flowchart showing the operation of the mounting head 103 when attention is paid to one of the mounting heads 103 provided in the rotary head 100.

  The mounting head 103 present at the suction nozzle selection position (position A in FIG. 4) selects a suction nozzle suitable for the component to be mounted next (S401). Next, the rotation base 101 is rotated until the mounting head 103 is positioned at the suction portion (position B in FIG. 4) (S402). At this position, the mounting head 103 sucks and holds the component from the suction opening 112 of the component cassette 111 using the suction nozzle (S403).

  Next, the rotation base 101 is rotated until the mounting head 103 is positioned at the component recognition camera 105 (position C in FIG. 4) (S404). Then, an image of the component picked up and conveyed by the small-field camera 105b or the large-field camera 105c provided at this position is picked up and subjected to image processing to recognize the picked-up state of the component (S405). In the recognition of the suction state, the two-dimensional shift amount between the center of the component and the center of the suction nozzle and the rotational shift amount of the component are detected. Cannot be determined). If it is determined that there is a problem in the suction state such as a missing part (YES in S406), the sucked part is discarded (S410). On the other hand, if it is determined that there is no malfunction (NO in S406), the component is corrected in the θ direction based on the rotational deviation amount detected earlier at the position D shown in FIG. 4 (S407). The two-dimensional deviation amount is corrected by finely adjusting the position of the printed board using the XY table 15. Next, the rotation base 101 is rotated until the mounting head 103 is positioned at the mounting portion (position E in FIG. 4) (S408). Then, the mounting head 103 is moved downward according to the height of the component on which the mounting head 103 is mounted, and the component is pushed into the printed circuit board P and mounted (S409).

  FIG. 7 is a diagram for explaining various modes of the component supply unit 11. FIG. 7A is a diagram showing a replacement mode. As shown in FIG. 7, the component supply unit 11 includes a left component supply unit ZL and a right component supply unit ZR. In the exchange mode, the same type of component is set in the left component supply unit ZL and the right component supply unit ZR. For example, when a component breakage occurs during production of a board using the left component supply unit ZL, the right component supply unit ZR at the standby position is moved to the suction position by the rotary head 100, and then the right component supply unit ZR. Is used to produce substrates. In the meantime, the left-part supply unit ZL is replenished with the parts that have run out. In this way, the right component supply unit ZR and the left component supply unit ZL have the same component configuration, so that the component mounter 10 can be operated continuously for a long time.

  FIG. 7B is a diagram showing a preparation mode. For example, during the production of a board using the left part supply unit ZL, the mounting component of the board to be produced next is set in the right part supply part ZR. Thus, this mode corresponds to the production substrate switching in advance. This mode is the optimal mode when there are many switching of production boards.

  FIG. 7C is a diagram showing a connection mode, for example, a mode in which a board is produced using both the right part supply part ZR and the left part supply part ZL. This mode is suitable for production with many types of parts used for one substrate.

FIG. 8 is a functional block diagram showing the configuration of the component mounting order determination apparatus.
The component mounting order determination apparatus 300 is a computer that determines the mounting order of components on a board, and includes an arithmetic control unit 301, a display unit 302, an input unit 303, a memory unit 304, a component mounting order determination program storage unit 305, and a communication I. / F (interface) unit 306, database unit 307, and the like.

  The component mounting order determining apparatus 300 is realized by a general-purpose computer system such as a personal computer executing a component mounting order determining program, and is not connected to the component mounter 10 and is a stand-alone simulator (component mounting). It also functions as an order determination tool.

  The arithmetic control unit 301 is a CPU (Central Processing Unit), a numerical processor, or the like, and loads and executes a necessary program from the component mounting order determination program storage unit 305 to the memory unit 304 in accordance with an instruction from the user. Each component 302 to 307 is controlled according to the execution result.

  The display unit 302 is a CRT (Cathode-Ray Tube), LCD (Liquid Crystal Display), or the like, and the input unit 303 is a keyboard, a mouse, or the like. It is used for dialogue between the apparatus 300 and the operator.

  The communication I / F unit 306 is a LAN (Local Area Network) adapter or the like, and is used for communication between the component mounting order determination device 300 and the component mounter 10. The memory unit 304 is a RAM (Random Access Memory) or the like that provides a work area for the arithmetic control unit 301.

  The database unit 307 includes input data (such as the mounting point information array 307a and the conversion table 307b) used for the component mounting order determination process by the component mounting order determination apparatus 300, the component mounting order determined by the component mounting order determination process, and the like. Is a hard disk or the like.

FIG. 9 is a diagram illustrating an example of the mounting point information array 307a.
The mounting point information array 307a is an array composed of data with N (= n + 2) elements when the number of mounting points on the substrate 20 is n.

  Here, as shown in FIG. 10, two substrate marks 402 and 404 are formed in advance on the substrate 20, and five mounting points (mounting point 1 to mounting point 5) are provided thereon. It shall be. The five mounting points are points where components are mounted by the component mounting machine 10. The substrate mark 402 and the substrate mark 404 are imaged by a substrate recognition camera (not shown) provided in the rotary head 100 for positioning the substrate 20, and their positions are recognized. It is assumed that the position of the substrate mark 402 is recognized first, and then the position of the substrate mark 404 is recognized. In addition, a virtual point called a dummy point 406 is provided on the substrate 20. Although the dummy point 406 is provided in the vicinity of the substrate mark 404 in the drawing, it is a virtual point and may not be provided on the substrate 20 or may be at the same position as the substrate mark 404. .

  Referring to FIG. 9 again, each array element includes any of information regarding the position of the mounting point (hereinafter referred to as “mounting point information”), information regarding the position of the board mark, and information regarding the position of the dummy point. Such information is stored. FIG. 9 shows information stored in the second element Data [2] of the mounting point information array 307a (Data). In the element Data [2], mounting point information of the mounting point 2 is stored. The X coordinate “45” and the Y coordinate “18” indicating the position of the mounting point 2 on the substrate 20 and the mounting point 2 The position (Z number) “24” on the Z-axis in the component supply unit 11 of the component cassette 111 in which the component to be mounted is stored, and the speed (head speed) from when the mounting head 103 picks up the component to when it is mounted ) “6” and the movement speed “2” of the XY table are stored.

  The head speed is shown in eight stages from level “1” to level “8”, where level 1 is the fastest head speed and level 8 is the slowest head speed. That is, in the case of a heavy part or a large part, the head speed becomes slow because it cannot be mounted at a high head speed. The head speed is a speed from when the mounting head 103 picks up the component until mounting. Here, it is assumed that there is no half rotation of the mounting head 103 for simplification of processing.

  The movement speed of the XY table is shown in 8 stages from level “1” to level “8”, level 1 is the fastest movement speed of the XY table 15, and level 8 is the slowest movement of the XY table 15. Is speed. That is, in the case where the component to be mounted is a heavy component or a large component, if the moving speed of the XY table 15 is too high, the mounted component is stripped off from the substrate 20 and flies away. For this reason, in the case of such a component, the moving speed of the XY table 15 is set to a large level.

  It is assumed that the information related to the position of the substrate mark and the information related to the position of the dummy point include the same information as the mounting point information.

FIG. 11 is a diagram illustrating an example of the conversion table 307b. The conversion table 307b is a table for replacing the head speed, the moving speed of the XY table 15, and the like with time. The first column shows the time t (msec) after replacement. The second column shows the distance Δxy between the mounting point i and the mounting point j. The third column shows the distance on the Z-axis between the component cassette in which the component to be mounted at the mounting point i is stored and the component cassette in which the component to be mounted at the mounting point j is stored (the absolute value of the Z number difference). Is shown. The fourth column shows the head speed when the component j is mounted. The fifth column shows the maximum value W _{max (0} → _i) of the moving speed of the XY table 15 with respect to the mounting point i from the moving speed of the XY table 15 with respect to the first mounting point.

  For example, when the head speed and the XY speed are both level 1 and the distance Δxy from the mounting point i to the mounting point j is 20, the time required for movement is obtained as 120 msec. Further, when the head speed and the XY speed at the mounting point j are both level 3, the time required from mounting the component to mounting is 140 msec.

In addition, about the value which is not described in the conversion table 307b, it calculates | requires by performing linear interpolation etc. from the already calculated | required value. For example, when the distance Δxy exceeds 103, the time t _xy (Δxy) for moving the distance Δxy is obtained by the following equation (1).

t _xy (Δxy) = 500 + (Δxy−103) /0.356 (1)
The other parameters are also replaced with time by obtaining similar interpolation equations.

FIG. 12 is a flowchart of processing executed by the component mounting order determination apparatus 300.
The arithmetic control unit 301 reads the mounting point information array 307a and the conversion table 307b from the database unit 307 (S2). According to the mounting point information array 307a shown in FIG. 9, it can be seen that the components are mounted in the order of the arrows shown in FIG.

  Next, the arithmetic control unit 301 creates a tact table based on the read mounting point information array 307a and the conversion table 307b (S4). The tact table is a table that summarizes the time required from mounting a component at the mounting point i to mounting the component at the mounting point j.

FIG. 14 is a diagram illustrating an example of a tact table. Here, the board mark 404 is the mounting point 0 and the dummy point 406 is the mounting point 6. The mounting time t _ij from the mounting point i to the mounting point j, which is each value of the tact table, is obtained according to the following equation (2).

t _ij = max {t _xy (Δxy), t _z (Δz),
t _h (h _j ), t _w (w _{max (0} → _i) )} (2)
Here, the mounting point information i related to the mounting point i described above is expressed as (x _i , y _i , z _i , h _i , w _i ), and the mounting point information j related to the mounting point j is expressed as (x _j , y _j , z). _j , h _j , w _j )
Δxy = max {| x _i −x _j |, | y _i −y _j |}
Distance between mounting points i and j on xy table t _xy (Δxy): value obtained by replacing Δxy with time based on conversion table 307b Δz = | z _i −z _j |
Based on the conversion table 307b, a distance t _z (Δz): Δz between the component cassette storing the component mounted at the mounting point i and the component cassette storing the component mounted at the mounting point j is based on the conversion table 307b. Value replaced with time (time required for component supply from the component supply unit 11 to the mounting head 103)
h _j : Head speed level of part j (level 1 (fast) to 8 (slow))
t _h (h _j ): the time from when the component calculated based on the conversion table 307b is picked up to when it is mounted (however, the moving time for half rotation of the mounting head 103 is not considered)
w _{max (0} → _i) : The maximum value (level 1 (fast) to 8 (slow)) from the moving speed of the XY table 15 with respect to the first mounting point to the moving speed of the XY table 15 with respect to the mounting point i
t _w (w _{max (0} → _i) ): value obtained by replacing w _{max (0} → _i) with time based on the conversion table 307b. Each mounting time corresponds to “cost” in the traveling salesman problem described later. To do. In the calculation process of the tact table (S4), assuming that t _ij = t _ji, so that not calculated for t _ji. Thereby, shortening of processing time can be aimed at. As “cost”, not only each mounting time but also a moving distance between two mounting points may be defined as a cost.

  Further, the mounting time from the dummy point to each mounting point (mounting point 1 to mounting point 5) is assumed to be a value (here, 10000000) larger than the mounting time between any arbitrary mounting points. Further, the mounting time between the board mark 404 and the dummy point 406 is assumed to be a value (here, 1) smaller than the mounting time between any arbitrary mounting points. In addition, these values are examples and are not limited to these.

  Next, the arithmetic control unit 301 creates an initial tour based on the mounting point information array 307a (S6). FIG. 15 is a diagram for explaining a method of creating an initial circuit. The initial tour circuit is created by rearranging the mounting point information array 307a in the order in which the movement speed of the XY table is fast (in order of increasing level value). Here, the XY speed level of the substrate mark 404 (Data [0]) is 0, and the XY speed level of the dummy point 406 (Data [6]) is 9. By doing so, the start point becomes the substrate mark 404 and the end point becomes the dummy point 406. In order to increase the processing speed, it is assumed that the mounting order array p that stores the element numbers of the mounting point information array 307a separately is used instead of replacing the elements of the mounting point information array 307a. p [i] indicates an element number in the mounting point information array 307a of the mounting point mounted i + 1. That is, according to the mounting order array p in FIG. 15, the components are mounted in the order of the board mark 404, mounting point 4, mounting point 5, mounting point 1, mounting point 2, mounting point 3 and dummy point 406. . FIG. 16 shows the mounting order. That is, a closed loop that returns to the dummy point 406 is created after mounting is performed in the above-described order using the substrate mark 404 as a start point.

  Next, the arithmetic control unit 301 obtains the total mounting time in the obtained initial circuit (S8). That is, the total mounting time T in the initial circuit is obtained according to the following equation (3).

T = t ₀₄ + t ₄₅ + t ₅₁ + t ₁₂ + t ₂₃ + t ₃₆ + t ₆₀ (3)
Here, as an example, the total mounting time is obtained as follows according to the tact table shown in FIG.

T = 200 + 380 + 400 + 400 + 300 + 10000000 + 1
= 10002081 (msec)
Next, the arithmetic control unit 301 obtains a tour that minimizes the total mounting time according to the processing (S10 to S18) described below.

  According to the above description, the mounting order p representing the initial circuit obtained in the initial circuit creation process (S6) is created as shown in FIG. 15, but for simplification of the following description, It is assumed here that the mounting order array p representing the initial circuit obtained in the initial circuit creation process (S6) is created as shown in FIG. For simplicity of explanation, the XY speed of the mounting point 0 (board mark 404) is level 0, the XY speeds from the mounting point 1 to the mounting point 5 are all 1, and the mounting point 6 (dummy point). It is assumed that the speed level of 406) is 9.

  First, the arithmetic control unit 301 sets the loop counter i to 0 (S10). Next, the arithmetic control unit 301 applies the 2-OPT method of the traveling salesman problem to the traveling circuit, and tries to change the mounting order (S12). Details of the 2-OPT method (S12) will be described later. When the mounting order is changed by the 2-OPT method (YES in S14), the loop counter i is reset to 0 (S16). If the mounting order has not been changed (NO in S14), the loop counter i is incremented by one (S18). The above processing (S12 to S18) is repeated until the value of the loop counter i becomes 7, that is, (i = N = n + 2) (loop A). The circuit at the time when the value of the loop counter i becomes 7 is the circuit that minimizes the total mounting time obtained by using the traveling salesman problem. That is, if the mounting order is not changed even if the loop A is repeated seven times (N times), the component mounting is determined that the circuit required at that time is the optimal circuit. Determine the order.

  Next, the 2-OPT method (S12) will be described in detail. FIG. 18 is a flowchart showing detailed processing of the 2-OPT method.

  The arithmetic control unit 301 sets i + 2 to the loop counter j (S102). When the value of the loop counter j exceeds 7, ((i + 2) mod 7) is substituted for the loop counter j. Here, “x mod y” indicates the remainder when x is divided by y.

  Next, the arithmetic control unit 301 mounts the mounting point indicated by the element p [i + 1] of the mounting order array p (hereinafter, the mounting point indicated by the element p [k] is referred to as the mounting point p [k]) and the mounting point. It is determined whether or not the XY speed of p [j] is the same (S104).

  If the XY speeds are the same (YES in S104), the total implementation time of the tour when the mounting point p [i + 1] and the mounting point p [j] are exchanged is calculated (S106).

  The arithmetic control unit 301 determines whether or not the calculated total mounting time is shorter than the total mounting time before the calculation (S108).

For example, when i = 0 and j = 2, if the total mounting time is obtained before and after the mounting order of the mounting points p [1] and p [2] is changed, the total mounting time does not change. This is due to the following reasons. That is, when considering the mounting order of the portions that affect the change in the total mounting time, the mounting time between two points from p [0] to p [1] before the replacement of the mounting order is t ₀₁ = 200 ( msec). The mounting time between two points from p [2] to p [3] is t ₂₃ = 300 (msec) from the tact table. For this reason, the mounting time from p [0] to p [3] is 500 (msec). On the other hand, the mounting time from p [0] to p [2] after changing the mounting order is t ₀₂ = 300 (msec) from the tact table, and the mounting time from p [1] to p [3] is from the tact table. t ₁₃ = 200 (msec). For this reason, the mounting time from p [0] to p [3] is 500 (msec), which is the same as before the mounting order is changed. For this reason, the total mounting time does not change.

  If the XY speeds of the mounting point p [i + 1] and the mounting point p [j] are not the same (NO in S104) or the total mounting time is not shortened (NO in S108), the value of the loop counter j is incremented by one. Increment (S114). After that, while satisfying the condition that the value of the loop counter j is smaller than (i + n + 1), the processing from S104 is repeated (loop B). When j exceeds N, “j mod N” is used instead of j. When (i + n + 1) exceeds N, “(i + n + 1) mod N” is used instead of (i + n + 1). Shall be used. When the processing of loop B is finished, the 2-OPT method (S12 in FIG. 12) is finished.

  If the total mounting time after the mounting point replacement is shorter than the total mounting time before the mounting point replacement (YES in S108), the arithmetic control unit 301 updates the total mounting time to the total mounting time after the mounting point replacement. (S110). Further, the arithmetic control unit 301 exchanges the mounting order by exchanging elements of the mounting order array p (S112). As will be described later with reference to FIG. 20, replacement of mounting points corresponds to replacement of closed-loop paths. When a path is replaced, a closed loop is not established unless the direction of the path existing in the middle of the replaced path is reversed. For this reason, when switching the mounting order, the process of reversing the path direction is also performed. This completes the 2-OPT method (S12 in FIG. 12) in the traveling salesman problem.

  For example, when i = 0 and j = 4, between the mounting points before replacement of mounting point p [1] (here mounting point 1) and mounting point p [4] (here mounting point 4) The mounting time is as shown in the lower table of FIG. Further, the mounting time between the mounting points after the mounting point 1 and the mounting point 4 are exchanged is as shown in the lower table of FIG. FIG. 20A and FIG. 20B are diagrams respectively showing the traveling circuits before and after the mounting points are replaced. That is, switching the mounting point 1 and the mounting point 4 corresponds to switching the path from the mounting point 0 to the mounting point 1 and the path from the mounting point 4 to the mounting point 5.

Considering only the mounting time that affects the total mounting time, the mounting time from mounting point 0 to mounting point 1 before changing the mounting point is t ₀₁ = 200 (msec) from the tact table, and mounting from mounting point 4 The mounting time at point 5 is t ₄₅ = 380 (msec) from the tact table. Therefore, the total mounting time is 580 (msec). On the other hand, the mounting time from mounting point 0 to mounting point 4 after replacement of mounting points is t ₀₄ = 200 (msec) from the tact table, and the mounting time from mounting point 1 to mounting point 5 is t ₁₅ = from the tact table. 100 (msec). For this reason, these total mounting time is 300 (msec), and it turns out that it is shorter than before mounting point replacement. As shown in FIG. 19B, the element p [2] existing between p [1] and p [4] is exchanged with the mounting point p [1] and the mounting point p [4]. ] And p [3] are also exchanged. That is, the value of p [2] changes from 2 to 3, and the value of p [3] changes from 3 to 2. As shown in FIG. 20, by replacing p [1] and p [4], the path from p [1] to p [2] existing between them, from p [2] This is because the direction of the path from p [3] and the path from p [3] to p [4] is reversed. Even if the direction of the path is reversed, the tact table is obtained as t _ij = t _ji in the tact table calculation process (S4) described above. For this reason, as far as the path direction is changed, it does not affect the change in the total mounting time.

  20 can be summarized as an example in which, even if the path direction is reversed, the mounting time between the mounting points constituting the path does not change in FIG. 20B. The final goal is to reduce the total closed-loop implementation time. Therefore, if this condition is satisfied, the mounting order may be changed as shown in FIG. 20 (a) to FIG. 20 (b). Further, FIG. 20B shows that when two mounting points are exchanged, the direction of the path must be exchanged so that a closed loop is appropriately formed.

  FIG. 21 is a diagram illustrating a processing result of the mounting order determination process performed by the component mounting order determination apparatus 300 described above. FIG. 21A is a diagram showing the mounting order p and the mounting time between mounting points. FIG. 21B is a diagram showing a mounting point circuit on the actual board and the mounting time between the mounting points. According to the mounting order array p, a circuit is generated in the order of mounting points 0, 4, 2, 3, 1, 5, 6 and 0, and the total mounting time is 10001001. Here, when the total implementation time of the closed loop shown in FIG. 20B is calculated, it is 10001251, and it can be seen that the total implementation time of the closed loop shown in FIG. 21B is shorter.

  FIG. 22 is a diagram showing the relationship between the mounting order array p, the mounting point information array Data, and the mounting order shown in FIG. The value of each element of the mounting order array p shown in FIG. 22A indicates the index of the mounting point information array Data shown in FIG. 22B, and between the mounting order array p and the mounting point information array Data. Indicates an element of the mounting point information array Data indicated by each element of the mounting point array p. That is, the value 0 of p [0] indicates the index of the element of the mounting point information array Data of the mounting point that is mounted first. Therefore, Data [0] is indicated. Data related to the position of the substrate mark 404 is stored in Data [0]. The value 4 of p [1] indicates the index of the element of the mounting point information array Data of the mounting point that is mounted second. Therefore, Data [4] is indicated. In Data [4], mounting point information of mounting point 4 is stored. When the mounting mark information array Data is rearranged on the basis of the value of the mounting order array p by removing the board mark 404 and the dummy point 406 from the circuit determined as described above, FIG. As shown, a mounting point information array Data that is mounted in the order of mounting points 4, 2, 3, 1 and 5 is generated.

  FIG. 23 is a diagram for explaining another method of creating the mounting point information array Data. Assume that a mounting point array p as shown in FIG. 23 (a) and a mounting point information array Data as shown in FIG. 23 (b) are given. These show that the starting point of the circuit is a dummy point. For this reason, a circuit as shown in FIG. 24 is generated. Therefore, in such a case, the mounting point information array Data as shown in FIG. 23C is created according to the same method as described in FIG. 22, and then the mounting order in the mounting point information array Data is changed. Perform replacement work. As a result, a circuit as shown in FIG. 21C is generated. By obtaining the mounting point information array Data in this way, the component is mounted from the mounting point close to the board mark 404.

FIG. 25 is a table showing experimental results of the conventional method and the method according to the present application.
In this table, the sample number is shown in the first column, the number of mounting points is shown in the second column, the number of component types is shown in the third column, and the conventional method ( The total mounting time when the parts are mounted according to the mounting order determined by the greedy method) is shown. The fifth column mounts the parts according to the mounting order determined using the traveling salesman problem that is the method of the present application. The sixth column shows the percentage of the total mounting time shown in the fourth column with respect to the total mounting time shown in the fifth column, and the seventh column shows the present application time. This shows the processing time required to obtain the mounting order according to the above method. In the conventional method and the method of the present application, the arrangement of the Z-axis component cassettes is the same and the conditions are aligned.

  According to the improvement rate shown in the sixth column, it is over 100% in all samples, and an improvement of 18.2% at maximum (sample E) is observed. In addition, the processing time required to obtain the mounting order is all less than 1 second.

  As described above, according to the present embodiment, an optimal component mounting order can be determined within a realistic time.

  The component mounting system according to the embodiment of the present invention has been described above, but the present invention is not limited to this embodiment.

  For example, in the above-described embodiment, on the premise of a component mounting machine having a rotary head, the expression for the mounting time between the mounting points (formula (2)) is obtained and the mounting order is obtained. By using another formula, the board itself does not move, but the component suction head is moved to the mounting point on the board, and the component mounting order in a component mounting machine called a modular machine for mounting the component is obtained. Good.

  In the method of determining the component mounting order shown in FIG. 12, the initial circuit is obtained by rearranging the mounting point information array 307a in the order in which the moving speed of the XY table is fast (in order of increasing level value). However (S6 in FIG. 12), the initial circuit may be obtained by the greedy method. By doing so, it is possible to perform the component mounting order determination process at a high speed by the subsequent traveling salesman problem.

  In the above-described embodiment, the dummy point 406 is provided in the vicinity of the board mark 404. However, the dummy point 406 is not necessarily provided in the vicinity of the board mark 404. A dummy point 406 may be provided in the vicinity of the component mounting point to be mounted.

  Further, the dummy point 406 may be located at the same position as a component mounting point at which a component is mounted first in advance.

  The present invention can be applied to a method for determining a component mounting order in a component mounting machine, and in particular to a method for determining a component mounting order in a rotary machine.

1 is an external view of a component mounting system according to an embodiment of the present invention. It is the schematic which shows the positional relationship of a components supply part and a rotary head. It is an external view of the reel around which the component tape holding components is wound. It is the figure which showed typically the positional relationship of a rotary head, a board | substrate, and a component supply part. It is a figure which shows an example of the components recognition camera provided in the position C of FIG. 4 is a flowchart showing an operation of the mounting head 103 when attention is paid to one of the mounting heads 103 provided in the rotary head 100. It is a figure for demonstrating the various modes of the components supply part. It is a functional block diagram which shows the structure of a component mounting order determination apparatus. It is a figure which shows an example of the mounting point information arrangement | sequence 307a. It is a figure which shows the positional relationship of the mounting point on a board | substrate 20, a board | substrate mark, and a dummy point. It is a figure which shows an example of the conversion table 307b. 10 is a flowchart of processing executed by the component mounting order determination apparatus 300. It is the figure which showed the mounting order of components typically. It is a figure which shows an example of a tact table. It is a figure for demonstrating the preparation method of an initial circuit. It is the figure which showed the mounting order of components typically. It is a figure which shows an example of the mounting point arrangement | sequence p. It is a flowchart which shows the detailed process of 2-OPT method. It is a figure which shows the mounting point arrangement | sequence p before and behind the mounting order change of components. It is a figure which shows the tour circuit before and behind the mounting order change of components. It is a figure showing the processing result of the mounting order determination process by the component mounting order determination apparatus 300. FIG. FIG. 22 is a diagram illustrating a relationship among a mounting order array p, a mounting point information array Data, and a mounting order shown in FIG. It is a figure for demonstrating the other production methods of the mounting point information arrangement | sequence Data. It is the figure which showed an example of the cyclic circuit typically. It is a table | surface which shows the experimental result of the conventional method and the method concerning this application.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Component mounting machine 11 Component supply part 15 Table 20 Board | substrate 100 Rotary head 101 Rotation base 103 Mounting head 105a Lens barrel 105b Small field camera 105c Large field camera 105 Parts recognition camera 106 Parts 107b Half mirror 107a Mirror 110 Parts tape 111 Parts cassette DESCRIPTION OF SYMBOLS 112 Adsorption opening part 113 Reel 114 Tape feeder 300 Component mounting order determination apparatus 301 Calculation control part 302 Display part 303 Input part 304 Memory part 305 Component mounting order determination program storage part 306 Communication I / F part 307 Database part 307a Mounting point information arrangement 307b Conversion table 402, 404 Substrate mark 406 Dummy point

Claims (6)

A method for determining a mounting order of components in a component mounter for mounting a plurality of components on a board,
For any two points including a plurality of component mounting points , the component is mounted on the distance between the two points or one of the two points, and then the component is mounted on the other point. Define the time required until the cost between the two points,
In order to apply the traveling salesman problem in a closed loop starting from one of the plurality of component mounting points and a board mark provided for board position recognition to determine the mounting order of the parts, the start A circuit creation step for creating a circuit by setting a dummy point that is the end point of the closed loop at or near the point; and
Component mounting point information for acquiring position information of the plurality of component mounting points, the board mark and the dummy point, and acquiring the cost between each of the plurality of component mounting points, the board mark and the dummy point. An acquisition step;
The traveling route calculation step the sum of the cost of the traveling route to the end point of the visited plurality of component mounting point from the start point the dummy point is the traveling route that minimizes, is calculated using the traveling salesman problem When,
A mounting order determining step for determining a component mounting order for the components based on the circuit. The cost from the dummy point to each mounting point shall be greater than the cost between any arbitrary mounting points
The component mounting order determination method according to claim 1, wherein: The starting point, of the substrate mark provided on the substrate, the component mounting order deciding method in claim 1, characterized in that the substrate marks are finally recognized. An apparatus for determining a mounting order of components in a component mounter for mounting a plurality of components on a board,
For any two points including a plurality of component mounting points , the component is mounted on the distance between the two points or one of the two points, and then the component is mounted on the other point. Define the time required until the cost between the two points,
In order to apply the traveling salesman problem in a closed loop starting from one of the plurality of component mounting points and a board mark provided for board position recognition to determine the mounting order of the parts, the start A traveling route creation means for creating a traveling route by setting a dummy point that is an end point of the closed loop at or near a point; and
Component mounting point information for acquiring position information of the plurality of component mounting points, the board mark and the dummy point, and acquiring the cost between each of the plurality of component mounting points, the board mark and the dummy point. Acquisition means;
Traveling route calculating means sum of the cost of the traveling route to the end point of the visited plurality of component mounting point from the start point the dummy point is the traveling route that minimizes, is calculated using the traveling salesman problem When,
And a mounting order determining means for determining a mounting order of the components based on the circuit. A component mounter for mounting components on a board,
According placement order of claims 1-3 wherein the component which is determined by the component mounting order deciding method according to any one of the component mounting apparatus, characterized in that mounting the component on the substrate. A program for causing a computer to execute the steps included in the component mounting order determination method according to any one of claims 1 to 3 .

Images