Production support systems, production support methods, and programs

Abstract

A production assistance system (1) estimates the probability of each of a plurality of units, of which a mounting machine (10) is composed, being the cause of a mounting error that has occurred in the mounting machine (10). The production assistance system comprises: an acquisition unit (for example, an error count preprocessing unit (31) and a correction amount preprocessing unit (32)) that acquires, from the mounting machine (10), a mounting log that includes information about a mounting error and from which the probability is to be estimated; a mounting error cause estimation unit (33) that, on the basis of the mounting log and a first estimation model based on the relationship, in each of the plurality of units, between a mounting error count and malfunction of the unit, estimates the probability in each of the plurality of units; and an output unit (for example, a display device (40)) that outputs a result of the estimation by the mounting error cause estimation unit (33). The mounting log includes information about a production quantity in each of the plurality of unit and the mounting error count.

Description

[Technical Field]

[0001] This disclosure relates to a production support system, a production support method, and a program. [Background technology]

[0002] Patent Document 1 discloses a printed circuit board manufacturing system that makes effective use of trace information and predicts fluctuations in manufacturing quality and equipment status to appropriately set the parameters of the mounting equipment (mounting machine) in order to obtain good manufacturing quality. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2021-12979 [Overview of the project] [Problems that the invention aims to solve]

[0004] Incidentally, there are times when it is desirable to support the identification of units that cause assembly errors in assembly machines, while keeping cost increases to a minimum.

[0005] Therefore, this disclosure provides a production support system, a production support method, and a program that can help identify units that cause implementation errors while suppressing cost increases. [Means for solving the problem]

[0006] A production support system according to one aspect of the present disclosure is a production support system that estimates the probability that each of the multiple units was a cause of an assembly error that occurred in an assembly machine composed of multiple units, and comprises: an acquisition unit that acquires a first assembly log from the assembly machine that includes information about the assembly error and is the target of the probability estimation; an assembly error cause estimation unit that estimates the probability for each of the multiple units based on the first assembly log and a first estimation model based on the relationship between the number of assembly errors and the malfunction of each of the multiple units; and an output unit that outputs the estimation result of the assembly error cause estimation unit, wherein the first assembly log includes information about the number of units produced and the number of assembly errors.

[0007] A production support method according to one aspect of the present disclosure is a production support method for estimating the probability that each of the multiple units was a cause of an assembly error that occurred in an assembly machine composed of multiple units, the method comprising: obtaining a first assembly log from the assembly machine that includes information on the assembly error and is the target of the probability estimation; estimating the probability for each of the multiple units based on the first assembly log and a first estimation model based on the relationship between the number of assembly errors and the malfunction of each of the multiple units; outputting the estimated estimation result, wherein the first assembly log includes information on the number of units produced and the number of assembly errors.

[0008] A program relating to one aspect of this disclosure is a program for causing a computer to execute the above-described production support method. [Effects of the Invention]

[0009] According to one aspect of this disclosure, it is possible to realize a production support system, etc., that can help identify units that cause implementation errors while suppressing cost increases. [Brief explanation of the drawing]

[0010] [Figure 1]FIG. 1 is a block diagram showing a functional configuration of a production support system according to an embodiment. [Figure 2] FIG. 2 is a diagram showing a first example of data after preprocessing according to an embodiment. [Figure 3] FIG. 3 is a diagram showing a second example of data after preprocessing according to an embodiment. [Figure 4] FIG. 4 is a diagram showing an example of first regression coefficient information according to an embodiment. [Figure 5] FIG. 5 is a diagram showing an example of a correspondence table between tape feed accuracy and accuracy rank according to an embodiment. [Figure 6] FIG. 6 is a first flowchart showing the operation of a production support system according to an embodiment. [Figure 7] FIG. 7 is a diagram for explaining the estimation and use of a first regression coefficient according to an embodiment. [Figure 8] FIG. 8 is a diagram for explaining the estimation and use of a second regression coefficient according to an embodiment. [Figure 9] FIG. 9 is a second flowchart showing the operation of a production support system according to an embodiment. [Figure 10] FIG. 10 is a diagram showing a first example of a screen displayed by a display device according to an embodiment. [Figure 11] FIG. 11 is a diagram showing a second example of a screen displayed by a display device according to an embodiment. Embodiments for Carrying Out the Invention

[0011] A production support system according to an aspect of the present disclosure is a production support system that estimates the probability that each of a plurality of units has caused a mounting error that occurred in a mounting machine composed of the plurality of units. The system includes an acquisition unit that acquires, from the mounting machine, a first mounting log including information on the mounting error and being a target for estimating the probability, a mounting error factor estimation unit that estimates the probability for each of the plurality of units based on the first mounting log, a first estimation model for each of the plurality of units based on the number of mounting errors and the relationship between malfunctions of the unit, and an output unit that outputs the estimation result of the mounting error factor estimation unit. The first mounting log includes information on the production quantity and the number of mounting errors in the plurality of units.

[0012] As a result, the estimated probability for each of the plurality of units is output as an estimation result, so that an administrator or the like of the mounting machine can identify a malfunctioning unit by referring to the estimation result. That is, the production support system can assist an administrator or the like in identifying a malfunctioning unit by outputting the estimation result. In addition, since information on the production quantity and the number of mounting errors output by the mounting machine as standard functions are used, the probability can be estimated without adding a sensor or the like to the mounting machine. Therefore, the production support system according to an aspect of the present disclosure can assist in identifying a unit that causes a mounting error while suppressing an increase in cost.

[0013] Further, for example, the first estimation model may be a multiple regression model in which the first regression coefficient of each of the plurality of units based on the distribution of the information on the production quantity and the number of mounting errors in each of the plurality of units is used as an explanatory variable, and the probability is used as an objective variable.

[0014] As a result, the probability is estimated in consideration of a plurality of factors (a plurality of units), so that the estimation accuracy of the probability can be improved.

[0015] Furthermore, the system may further include a first coefficient estimation unit that estimates the first regression coefficient for each of the plurality of units based on, for example, the distribution of information regarding the number of production units and the number of production errors contained in a second implementation log of the implementation machine acquired before the first implementation log.

[0016] This allows the production support system to perform a series of processes, from generating the first estimation model to estimating probabilities.

[0017] Furthermore, for example, the output unit may display the probability of each of the multiple units being affected by the implementation error.

[0018] This makes it possible to visualize the probability of implementation errors, allowing administrators of the implementation machine to refer to the displayed information when determining malfunctions in multiple units. Therefore, it can effectively support the identification of units that are causing implementation errors.

[0019] Furthermore, for example, the plurality of units include a tape feeder for supplying components and a nozzle for picking up the components, and further comprises a tape feed accuracy estimation unit that estimates the tape feed accuracy based on the first implementation log and a second estimation model based on the relationship between the tape feed accuracy of the tape feeder and the amount of misalignment of the component's pick-up position, and the first implementation log may further include information regarding a control amount for controlling the nozzle.

[0020] This allows the operators of the mounting machines to identify malfunctioning units by referring to information based on the estimated tape feeding accuracy. In other words, the production support system can further assist operators in identifying malfunctioning units by outputting information based on tape feeding accuracy, while keeping costs down.

[0021] Furthermore, for example, the second estimation model may be a simple regression model that includes a second regression coefficient showing the relationship between the tape feeding accuracy and the misalignment of the component's suction position.

[0022] This allows the production support system to further assist managers and others in identifying malfunctioning units while suppressing the increase in processing load related to estimating tape feeding accuracy.

[0023] Furthermore, the system may further include a second coefficient estimation unit that estimates the second regression coefficient for each of the plurality of units based on the distribution of the control amount for controlling the nozzle included in the third implementation log of the implementation machine, which was acquired before the first implementation log, and the measured value of the tape feeding accuracy of the tape feeder.

[0024] This allows the production support system to perform a series of processes, from generating a second estimation model to estimating tape feeding accuracy.

[0025] Furthermore, for example, the output unit may display information regarding the tape feed accuracy estimated by the tape feed accuracy estimation unit.

[0026] This allows for the visualization of information regarding tape feeding accuracy, enabling administrators of mounting machines to refer to the displayed information when determining feeder malfunctions. Therefore, it can effectively support the identification of units that are causing mounting errors.

[0027] Furthermore, for example, the mounting machine does not necessarily have to be equipped with sensors that directly measure the state of the multiple units.

[0028] This means that the mounting machine does not need to have a hard sensor, thus more reliably reducing the cost of the mounting machine.

[0029] Furthermore, for example, the plurality of units may include a head spindle, a nozzle, and a feeder, and the mounting error factor estimation unit may estimate the probability of each of the head spindle, the nozzle, the feeder, and the parts supplied by the feeder.

[0030] This allows the production support system to estimate the probability of each of the multiple units, including the head spindle, nozzle, and feeder, as well as each individual component. In other words, the production support system can assist managers in determining whether the cause of an assembly error lies with the head spindle, nozzle, feeder, or component.

[0031] Furthermore, a production support method according to one aspect of this disclosure is a production support method for estimating the probability that each of the multiple units was a cause of an assembly error that occurred in an assembly machine composed of multiple units, the method includes obtaining a first assembly log from the assembly machine that includes information about the assembly error and is the target of the probability estimation, estimating the probability for each of the multiple units based on the first assembly log and a first estimation model based on the relationship between the number of assembly errors and the malfunction of each of the multiple units, outputting the estimated estimation result, the first assembly log includes information about the number of units produced and the number of assembly errors. Furthermore, a program according to one aspect of this disclosure is a program for causing a computer to execute the above production support method.

[0032] This will produce the same effect as the production support system described above.

[0033] These general or specific embodiments may be implemented using a system, method, integrated circuit, computer program, or a non-temporary recording medium such as a computer-readable CD-ROM, or any combination of a system, method, integrated circuit, computer program, or recording medium. The program may be pre-stored on the recording medium or supplied to the recording medium via a wide-area communication network, including the Internet.

[0034] The embodiments will be described in detail below with reference to the drawings.

[0035] The embodiments described below are all comprehensive or specific examples. The numerical values, shapes, components, arrangement and connection configurations of components, steps, and the order of steps shown in the following embodiments are examples only and are not intended to limit this disclosure. Furthermore, any components in the following embodiments that are not described in an independent claim will be described as optional components.

[0036] Furthermore, each figure is a schematic diagram and not necessarily a strictly accurate representation. Therefore, for example, the scale and other aspects may not necessarily match in each figure. Also, in each figure, substantially identical components are given the same reference numerals, and redundant explanations are omitted or simplified.

[0037] Furthermore, in this specification, terms indicating relationships between identical or equivalent elements, as well as numerical values ​​and numerical ranges, do not represent only strict meanings, but also include substantially equivalent ranges, such as differences of a few percent (or about 10%).

[0038] Furthermore, in this specification, ordinal numbers such as "first," "second," etc., do not mean the number or order of components unless otherwise specified, but are used to avoid confusion and to distinguish similar components.

[0039] (Embodiment) The production support system equipped with the production support device according to this embodiment will be described below with reference to Figures 1 to 11.

[0040] [1. Configuration of the production support system] First, the configuration of the production support system according to this embodiment will be explained with reference to Figure 1. Figure 1 is a block diagram showing the functional configuration of the production support system 1 according to this embodiment.

[0041] As shown in Figure 1, the production support system 1 comprises a coefficient estimation device 20, a factor probability estimation device 30, and a display device 40. The production support system 1 is a system for supporting production using the mounting machine 10. Specifically, the production support system 1 indirectly estimates the factor probability (also described as the error factor probability or mounting error factor probability) of a unit (equipment unit) for mounting errors that occur in the mounting machine 10, without using data from a sensor that directly senses the unit. Furthermore, in this embodiment, the production support system 1 also indirectly estimates the tape feeding accuracy of the feeder, without using data from a sensor that directly senses the tape feeding accuracy of the feeder. The sensor is, for example, a sensor that measures multiple units, such as a sensor that measures the supply position of parts in the feeder, the mounting position of parts, etc.

[0042] The mounting machine 10 is an example of production equipment that constitutes a manufacturing line, and is a mounting device (component mounting apparatus) that mounts components onto an object (workpiece) such as a circuit board. The mounting machine 10 is composed of multiple units. In this embodiment, the mounting machine 10 has a drive control unit (hereinafter also referred to as a spindle), a component mounting head unit (hereinafter also referred to as a nozzle), and a component supply unit (hereinafter also referred to as a feeder) as units. The drive control unit controls the rotation and movement of the component mounting head unit. The drive control unit has a head spindle that controls the rotation of the component mounting head unit by driving a motor. The component mounting head unit mounts (mounts) components supplied to the component supply position (component suction position) of the feeder onto the circuit board. The component mounting head unit is composed of a mounting head (head) equipped with component suction nozzles (nozzles) that can pick up components from the feeder and move up and down individually. For example, multiple nozzles are attached to the head. The component supply unit has one or more feeders arranged in a row, each supplying components to the component supply position. In this embodiment, the feeder is a tape feeder.

[0043] The mounting machine 10 periodically outputs a first mounting log L1 to the factor probability estimation device 30 during production. The mounting machine 10 may, for example, output a first mounting log L1 each time it mounts one component onto the board, or it may output a first mounting log L1 each time it mounts a predetermined number of components onto the board.

[0044] The first mounting log L1 includes information on the control of each unit and information indicating production results acquired by the mounting machine 10 from the operation of picking up components from the feeder to the operation of mounting the components onto the substrate. The first mounting log L1 is a log output by the mounting machine 10 as a standard function. For example, the first mounting log L1 includes information on control quantities for controlling multiple units, information identifying multiple units used in production, and information on production results. Information on control quantities includes, for example, information on control quantities (correction quantities) for controlling the movement and rotation of a unit (e.g., a nozzle) (e.g., the suction correction quantity and recognition correction quantity described later). Information indicating production results includes information on the number of units produced (e.g., the number of suction attempts) and the number of mounting errors. The first mounting log L1 does not include data obtained by directly sensing the units (i.e., data from hard sensors).

[0045] The number of mounting machines 10 supported by the production support system 1 is not particularly limited; there may be one or multiple machines.

[0046] The components are electronic components, such as resistors and capacitors, but are not limited to these. Furthermore, the workpiece is not limited to a circuit board; any workpiece capable of undergoing the specified processing is acceptable.

[0047] The mounting machine 10 does not necessarily have sensors (hard sensors) that directly measure the status (e.g., operation) of the drive control unit, component mounting head unit, and component supply unit. Such sensors include, for example, sensors that are not standard equipment on the mounting machine 10 but are installed as an add-on. Whether or not such sensors are standard equipment can be confirmed from the mounting machine 10 catalog, etc.

[0048] The coefficient estimation device 20 performs a process to estimate each regression coefficient using the second mounting log L2, which is acquired in advance from the mounting machine 10. "In advance" means before the above estimation process is performed (for example, before the first mounting log L1 is acquired). The second mounting log L2 is a mounting log for pre-estimation used to estimate each regression coefficient in advance. The second mounting log L2 contains data similar to the first mounting log L1 and includes information on the control of each unit acquired in the mounting machine 10 from the operation of picking up components from the feeder to the operation of mounting the components onto the substrate. The second mounting log L2 is a log output by the mounting machine 10 as a standard function. For example, the second mounting log L2 includes information on control quantities for controlling multiple units, the number of mounting errors, information identifying the multiple units used in production, and information on production results. Information on control quantities includes, for example, information on correction quantities for correcting the movement and rotation of units (for example, nozzles) (for example, the suction correction quantity and recognition correction quantity described later). Information regarding production performance includes the number of units produced (e.g., number of suction attempts) and the number of mounting errors. Note that the second mounting log L2 does not include data obtained by directly sensing the unit (i.e., data from hard sensors). The second mounting log L2 may also be used as a third mounting log.

[0049] The coefficient estimation device 20 comprises a miscount preprocessing unit 21, a correction amount preprocessing unit 22, a nonlinear regression coefficient estimation unit 23, and a linear regression coefficient estimation unit 24. The coefficient estimation device 20 can be implemented using a CPU (Central Processing Unit) and memory, etc. Furthermore, the processing by each functional block of the coefficient estimation device 20 is usually implemented by a program execution unit such as a processor reading and executing software (programs) recorded on a recording medium such as ROM.

[0050] The error count preprocessing unit 21 acquires the second implementation log L2 from the implementation machine 10, extracts data from the second implementation log L2 that will be used in the nonlinear regression coefficient estimation unit 23 to estimate the first regression coefficient, and outputs the extracted preprocessed data to the nonlinear regression coefficient estimation unit 23. Figure 2 shows a first example of the preprocessed data according to this embodiment. Figure 2 shows an example of the preprocessed data output from the error count preprocessing unit 21 to the nonlinear regression coefficient estimation unit 23.

[0051] As shown in Figure 2, the pre-processed data includes "index", "feeder serial number", "head spindle number", "nozzle serial number", "part serial number", "part size", "number of suction attempts", and "number of mounting errors".

[0052] The index is a number assigned to each combination of "feeder," "head spindle," "nozzle," and "part size." In other words, the post-preprocessing data is aggregated data for each combination of "feeder," "head spindle," "nozzle," and "part size."

[0053] The feeder serial number is the identification information (e.g., identification number) for each feeder.

[0054] The head spindle number is an identification number used to identify the head spindle that controls the rotation of the nozzle.

[0055] The nozzle serial number is the identification information (e.g., identification number) for each individual nozzle. A single component mounting head unit is equipped with multiple nozzles, and each of these nozzles is assigned a different identification number.

[0056] The component size indicates the width of the component used in production. The "1.0" shown in Figure 2 indicates, for example, that the width of the component when viewed from above is 1 mm. The component width is the length in the direction perpendicular to the component's movement direction (Y direction) in the feeder (X direction). The X direction is the direction in which the substrate is transported in the mounting machine 10.

[0057] A component serial number is identification information (e.g., an identification number) used to identify a component supplied by a feeder.

[0058] The number of suction trials indicates the number of times the nozzle has attempted to pick up a component. This number represents the number of times the nozzle performed component suction during a predetermined period or for a predetermined number of units (e.g., 1 lot) of production.

[0059] The number of mounting errors indicates the number of mounting errors that occurred out of the total number of suction attempts. Mounting errors include failures in component suction by the nozzle and failures in mounting components to the substrate.

[0060] For example, index "1" indicates that when a component with identification information "CP0001" and size "1.0" was mounted 1000 times using a nozzle with identification information "NZ00001" whose rotation was controlled by a head spindle with identification number "1", 12 mounting errors occurred. Thus, the pre-processed data includes information identifying the unit on which the component was mounted, information identifying the component to be mounted, and information regarding production results.

[0061] As will be explained in detail later, the inventors of this application have found a relationship between the number of mounting errors and unit malfunctions. In other words, the inventors have found that unit malfunctions can be estimated from the number of mounting errors. Therefore, the number of mounting errors is essential information in the post-preprocessing data output from the error count preprocessing unit 21. A malfunction refers to an alert that does not necessarily involve the stopping of the mounting machine 10, and is a state in which it is presumed that at least one of the feeder, head spindle, and nozzle has failed or deteriorated.

[0062] Referring again to Figure 1, the correction amount preprocessing unit 22 acquires the second implementation log L2 from the implementation machine 10, extracts data from the second implementation log L2 that will be used in the linear regression coefficient estimation unit 24 to estimate the second regression coefficient, and outputs the extracted preprocessed data to the linear regression coefficient estimation unit 24. Figure 3 is a diagram showing a second example of the preprocessed data according to this embodiment. Figure 3 shows an example of the preprocessed data output from the correction amount preprocessing unit 22 to the linear regression coefficient estimation unit 24.

[0063] As shown in Figure 3, the preprocessed data includes "index", "feeder serial number", "head spindle number", "nozzle serial number", "part serial number", "part size", "number of suction trials", and "each correction amount".

[0064] Each correction amount includes a recognition correction amount and an adsorption correction amount. The recognition correction amount and adsorption correction amount include correction amounts related to the X coordinate and Y coordinate, as well as the mean, median, and standard deviation of the correction amounts. The recognition correction amount indicates the correction amount for the nozzle's position in the X and Y axes after the nozzle adsorbs a component. For example, the recognition correction amount indicates the adjustment amount for the nozzle's position in the X and Y axes to mount a component to a target mounting position on the substrate. For example, the recognition correction amount is a value based on the difference between the position where the nozzle adsorbed the component and the adsorption target position.

[0065] The suction correction amount indicates the amount of positional correction performed on the nozzle in the X-axis and Y-axis directions when the nozzle picks up a part. For example, the suction correction amount indicates the amount of adjustment made when the nozzle position is adjusted so that the nozzle can pick up the part at its suction target position (e.g., the center position on the top surface of the part). For example, the suction correction amount is a value based on the difference between the suction target position of the part supplied by the feeder and the suction center of the nozzle.

[0066] Furthermore, for both the recognition correction amount and the adsorption correction amount, it is sufficient that at least one of the mean, median, and standard deviation is included in the preprocessed data. In addition, each correction amount can be identified from the image captured by the imaging device provided on the mounting machine 10.

[0067] Furthermore, the correction amount preprocessing unit 22 may calculate adsorption position displacement statistics, which will be described later, based on the recognition correction amount and the adsorption correction amount. Adsorption position displacement statistics may be included in the preprocessed data.

[0068] As will be explained in detail later, the inventors of this application have found that there is a relationship between each correction amount and tape feeding accuracy. In other words, the inventors have found that tape feeding accuracy can be estimated from each correction amount. Therefore, each correction amount is essential information in the post-processed data output from the correction amount preprocessing unit 22.

[0069] Referring again to Figure 1, the nonlinear regression coefficient estimation unit 23 estimates a first regression coefficient to estimate the unit factor probability for mounting errors based on the preprocessed data shown in Figure 2. The main elements in the component mounting process are the feeder, head spindle, nozzle, and component size. The first regression coefficient is used to identify which of the feeder, head spindle, nozzle, or component size is responsible for a mounting error included in the first mounting log L1. The first mounting log L1 is a log containing the same information as the second mounting log L2. The nonlinear regression coefficient estimation unit 23 is an example of the first coefficient estimation unit.

[0070] In this embodiment, the mounting error factor estimation unit 33 of the factor probability estimation device 30 uses "logistic regression analysis," a type of nonlinear regression analysis, as a method to determine the probability of each factor involved in the component mounting process, from the feeder, head spindle, nozzle, and component size, when a mounting error occurs. Therefore, the nonlinear regression coefficient estimation unit 23 estimates the first regression coefficient used in logistic regression analysis. Specifically, the nonlinear regression coefficient estimation unit 23 estimates the first regression coefficient of the linear predictor included in the logistic function (see Equation 2 described later) used in logistic regression analysis.

[0071] The nonlinear regression coefficient estimation unit 23 estimates the first regression coefficient for each element using the Markov chain Monte Carlo (MCMC) method from the distribution of the number of suction trials and the number of mounting errors for each combination of elements: feeder, head spindle, nozzle, and part size. The MCMC method is a technique for obtaining samples (generating random numbers) from a multivariate probability distribution and is a commonly used technique for maximum likelihood estimation of multiple parameters (in this case, regression coefficients) that constitute a statistical model, as described in this disclosure. Each element includes, for example, the feeder, head spindle, nozzle, and part (part size). The first regression coefficient is an explanatory variable in the first estimated model 33a.

[0072] Figure 4 shows an example of the first regression coefficient information according to this embodiment.

[0073] As shown in Figure 4, the following are included: "index", "feeder serial", "head spindle number", "nozzle serial", "part serial", "part size", "number of suction trials", "number of mounting errors", "bias term", and "each regression coefficient".

[0074] The bias term is used to calculate the linear predictor included in the logistic function. The bias term is a predetermined constant (e.g., β).

[0075] Each regression coefficient includes the feeder regression coefficient, which is the first regression coefficient for the feeder; the head spindle regression coefficient, which is the first regression coefficient for the head spindle; the nozzle regression coefficient, which is the first regression coefficient for the nozzle; and the part size regression coefficient, which is the first regression coefficient for the part size. For example, if the feeder is N f Book available, head spindle N s There are several, and the nozzle is N n There are several, and the part size is N c If there are N feeders, then there are also a corresponding number of first regression coefficients for each feeder. For example, if there are N feeders... f If there is a book, N f For each of the book feeders, the first regression coefficient is estimated individually. For example, the first regression coefficient for the feeder with feeder serial number "FD00001" is αf [1] The first regression coefficient for the feeder with feeder serial "FD00002" is α f [2] The same applies to the head spindle, nozzle, and component sizes.

[0076] Referring again to Figure 1, the linear regression coefficient estimation unit 24 estimates a second regression coefficient for estimating the tape feeding accuracy based on the preprocessed data shown in Figure 3. The tape feeding accuracy indicates the positional accuracy when the feeder (tape feeder) supplies components to the component supply position by tape feeding.

[0077] The linear regression coefficient estimation unit 24 estimates a linear regression model and its second regression coefficient (linear regression coefficient) that associates the statistical quantities (adsorption position misalignment statistics) obtained from each correction quantity shown in Figure 3 with the measured values ​​of the adsorption position misalignment (for example, measured values ​​of the tape feeding accuracy of the tape feeder), based on the distribution of these values. For example, the linear regression coefficient estimation unit 24 plots the adsorption position misalignment statistics on the horizontal axis and the feeding accuracy (measured values ​​of the inspection device) on the vertical axis, and estimates the second regression coefficient based on the distribution of the plot. The inspection device is configured to include a hard sensor that directly measures the tape feeding accuracy. The linear regression coefficient estimation unit 24 is an example of a second coefficient estimation unit.

[0078] For example, the linear regression coefficient estimation unit 24 estimates the second regression coefficient in the X direction based on the distribution of the X-direction statistics (adsorption position displacement statistics in the X direction) obtained from the correction amount in the X direction (X coordinate) among the correction amounts shown in Figure 3, and the measured values ​​of the adsorption position displacement in the X direction (for example, the measured value in the X direction of the tape feeding accuracy of the tape feeder). Also, for example, the linear regression coefficient estimation unit 24 estimates the second regression coefficient in the Y direction based on the distribution of the Y-direction statistics (adsorption position displacement statistics in the Y direction) obtained from the correction amount in the Y direction (Y coordinate) among the correction amounts shown in Figure 3, and the measured values ​​of the adsorption position displacement in the Y direction (for example, the measured value in the Y direction of the tape feeding accuracy of the tape feeder).

[0079] The second estimation model 34a generated by the linear regression coefficient estimation unit 24 is a simple regression model that includes a second regression coefficient that shows the relationship between the quantities related to tape feeding accuracy and component suction position misalignment in each of the multiple units.

[0080] The linear regression coefficient estimation unit 24 calculates the adsorption position displacement statistic based on the recognition correction amount and the adsorption correction amount. For example, the linear regression coefficient estimation unit 24 calculates the adsorption position displacement statistic based on the mean or median of the recognition correction amount and the mean or median of the adsorption correction amount. The adsorption position displacement statistic may also be calculated using the standard deviation. The adsorption position displacement statistic is calculated for both the X and Y coordinates. In other words, the second regression coefficient is estimated for both the X and Y coordinates.

[0081] The factor probability estimation device 30 outputs estimated results for the causes of mounting errors and tape feeding accuracy based on the first and second regression coefficients estimated by the coefficient estimation device 20 and the first mounting log L1. The first mounting log L1 is a log of the measurement target (estimation target) acquired when production is being carried out by the mounting machine 10. The factor probability estimation device 30 outputs estimated results for the causes of mounting errors and tape feeding accuracy using the first mounting log L1 acquired during production by the mounting machine 10, in parallel with production by the mounting machine 10. As a result, the factor probability estimation device 30 can help administrators etc. to determine unit malfunctions during production by the mounting machine 10, making it possible for administrators etc. to perform maintenance before unit failure occurs.

[0082] The factor probability estimation device 30 comprises a pre-processing unit for the number of errors 31, a pre-processing unit for the correction amount 32, an implementation error factor estimation unit 33, a tape feed accuracy estimation unit 34, and aggregation units 35-38. The factor probability estimation device 30 can be implemented using a CPU and memory, etc. Furthermore, the processing by each functional block of the factor probability estimation device 30 is usually implemented by a program execution unit such as a processor reading and executing software (programs) recorded on a recording medium such as ROM.

[0083] The functions of the error count preprocessing unit 31 and the correction amount preprocessing unit 32 are the same as those of the error count preprocessing unit 21 and the correction amount preprocessing unit 22, and the description thereof will be omitted. The error count preprocessing unit 31 acquires the first mounting log L1 from the mounter 10, generates the preprocessed data shown in FIG. 2 from the acquired first mounting log L1, and outputs the data to the mounting error factor estimation unit 33. Further, the correction amount preprocessing unit 32 acquires the first mounting log L1 from the mounter 10, generates preprocessed data including the data shown in FIG. 2 and the statistical amount of the adsorption position deviation from the acquired first mounting log L1, and outputs the data to the tape feed accuracy estimation unit 34.

[0084] Based on the preprocessed data from the error count preprocessing unit 31 and the first regression coefficient from the non-linear regression coefficient estimation unit 23, the mounting error factor estimation unit 33 estimates the factor probability of the unit for the mounting errors included in the preprocessed data. The mounting error factor estimation unit 33 estimates the factor probability of the unit in the first mounting log L1 to be measured using the first estimation model 33a based on the first regression coefficient.

[0085] The mounting error factor estimation unit 33 calculates the probability that each element (for example, each unit) is a factor for the generated error by a binomial distribution derived based on the preprocessed data and the first regression coefficient. The probability p that y mounting errors occur after N adsorptions is expressed by the following formula 1 as a binomial distribution model.

[0086]

Equation

[0087] Here, q represents the mounting error factor probability per adsorption trial for each element. When the above formula 1 is transformed, the probability q of the mounting error factor is expressed by the following formula 2.

[0088] q = 1 / (1 + exp(-z)) ···(Formula 2)

[0089] Here, z is a linear predictor, and α f is the first regression coefficient of the feeder, α n is the first regression coefficient of the nozzle, α sThe first regression coefficient of the head spindle, α c If we let be the first regression coefficient for component size, β be the bias term, and chipW be the component size, then it can be expressed by the following equation 3.

[0090] z = β + α f +α n +α s +α c / chipW ···(Formula 3)

[0091] The linear predictor z can be calculated from the pre-processed data from the error count pre-processing unit 31. Equations 2 and 3 are examples of the first estimation model 33a. The first estimation model 33a is a statistical model based on the relationship between the number of implementation errors in each of several units and the malfunction of that unit. The first estimation model 33a can also be said to be an identification model for identifying malfunctioning units from among several units.

[0092] The mounting error cause estimation unit 33 calculates the probability q of the mounting error cause by substituting the linear predictor z, obtained by substituting the first regression coefficients of the feeder, nozzle, head spindle, and component size when a mounting error occurs into Equation 3, into Equation 2. The mounting error cause estimation unit 33 calculates the probability q of the mounting error cause for each of the feeder, nozzle, head spindle, and component size.

[0093] Note that probability q is the dependent variable in the first estimation model 33a. The first estimation model 33a is a multiple regression model in which the first regression coefficient (partial regression coefficient) is the explanatory variable and probability q is the dependent variable.

[0094] The tape feed accuracy estimation unit 34 estimates the tape feed accuracy of the feeder of the mounting machine 10 based on the pre-processed data from the correction amount pre-processing unit 32 and the second regression coefficient from the linear regression coefficient estimation unit 24. A second estimation model 34a is generated including the second regression coefficient. The tape feed accuracy estimation unit 34 estimates the tape feed accuracy in the first mounting log L1 of the measurement target using the second estimation model 34a, which is based on the relationship between the tape feed accuracy of the tape feeder and the statistical amount of misalignment of the component's suction position.

[0095] The tape feed accuracy estimation unit 34 estimates the tape feed accuracy corresponding to the suction position deviation statistic calculated based on the first implementation log L1 of the measurement target, using the second regression coefficient. The tape feed accuracy is, for example, distance.

[0096] Furthermore, the tape feed accuracy estimation unit 34 estimates the accuracy rank of the feeder based on a correspondence table between tape feed accuracy and accuracy rank (see Figure 5 below). This accuracy rank is the accuracy rank at that time and may change over time.

[0097] Figure 5 shows an example of a correspondence table between tape feeding accuracy and accuracy rank according to this embodiment. X-direction feeding accuracy refers to the tape feeding accuracy in the X direction, and Y-direction feeding accuracy refers to the tape feeding accuracy in the Y direction.

[0098] As shown in Figure 5, the correspondence table associates one accuracy rank with each X-direction feed accuracy and Y-direction feed accuracy. In the example in Figure 5, the accuracy ranks are "A", "B", "C", and "D", indicating that the feed accuracy is increasing in this order.

[0099] The accuracy rank is an example of information based on tape feed accuracy.

[0100] Referring again to Figure 1, the aggregation unit 35 accumulates the error probability for each component size from the implementation error factor estimation unit 33 and outputs the error probability for each component serial number.

[0101] The aggregation unit 36 ​​stores the error probability for each nozzle from the implementation error factor estimation unit 33 and outputs the error probability for each nozzle serial number.

[0102] The aggregation unit 37 stores the error probability for each head spindle from the implementation error factor estimation unit 33 and outputs the error probability for each head spindle number.

[0103] The aggregation unit 38 stores the error probability for each feeder from the implementation error factor estimation unit 33 and the tape feed accuracy from the tape feed accuracy estimation unit 34, and outputs the error probability and accuracy rank for each feeder serial.

[0104] The aggregation units 35-38 accumulate the error probability for a predetermined production quantity (e.g., 1 lot) and calculate the probability of a single error for the predetermined production quantity based on the accumulated error probability. The aggregation units 35-38 may, for example, aggregate the error probability for a predetermined production quantity and estimate the average value as the probability of a single error. The error probability is an example of an estimated result.

[0105] The display device 40 displays various information. The display device 40 displays the estimated probability of implementation errors and the estimated tape feed accuracy by the factor probability estimation device 30. The display device 40 is a liquid crystal display device, but is not limited thereto.

[0106] The coefficient estimation device 20, the factor probability estimation device 30, and the display device 40 may be implemented as separate devices, or at least two of the devices may be implemented as a single integrated device. For example, the coefficient estimation device 20 and the factor probability estimation device 30 may be implemented as a single integrated device, or the coefficient estimation device 20, the factor probability estimation device 30, and the display device 40 may be implemented as a single integrated device.

[0107] [2. Operation of the Production Support System] Next, the operation of the production support system 1 configured as described above will be explained with reference to Figures 6 to 11. Figure 6 is a first flowchart showing the operation (production support method) of the production support system 1 according to this embodiment. Figure 6 shows the operation of estimating the first regression coefficient and the second regression coefficient. The process in Figure 6 is performed by the coefficient estimation device 20.

[0108] As shown in Figure 6, the error count preprocessing unit 21 and the correction amount preprocessing unit 22 of the coefficient estimation device 20 each acquire a second implementation log L2 for pre-estimation (S10). The error count preprocessing unit 21 and the correction amount preprocessing unit 22 function as acquisition units that acquire the second implementation log L2.

[0109] Next, the error count preprocessing unit 21 and the correction amount preprocessing unit 22 each perform preprocessing (S20). The error count preprocessing unit 21 performs preprocessing on the second implementation log L2 to generate, for example, the preprocessed data shown in Figure 2 and outputs it to the nonlinear regression coefficient estimation unit 23. The correction amount preprocessing unit 22 performs preprocessing on the second implementation log L2 to generate, for example, the preprocessed data shown in Figure 3 and outputs it to the linear regression coefficient estimation unit 24. In this embodiment, the correction amount preprocessing unit 22 outputs the preprocessed data including the adsorption position displacement statistics to the linear regression coefficient estimation unit 24.

[0110] Next, the nonlinear regression coefficient estimation unit 23 estimates the first regression coefficient based on the pre-processed data from the error count pre-processing unit 21 (S30). Figure 7 is a diagram illustrating the estimation and use of the first regression coefficient according to this embodiment. Note that Figure 7 omits the illustration of the error count pre-processing units 21 and 31.

[0111] As shown in Figure 7, the nonlinear regression coefficient estimation unit 23 estimates the first regression coefficient for each element using the MCMC method based on the second implementation log L2, and outputs the estimated first regression coefficient to the implementation error factor estimation unit 33 of the factor probability estimation device 30. The implementation error factor estimation unit 33 functions as an acquisition unit that obtains the first regression coefficient.

[0112] Referring again to Figure 6, the linear regression coefficient estimation unit 24 then estimates the second regression coefficient based on the preprocessed data from the correction amount preprocessing unit 22 (S40). Figure 8 is a diagram illustrating the estimation and use of the second regression coefficient according to this embodiment.

[0113] As shown in Figure 8, the linear regression coefficient estimation unit 24 estimates the second regression coefficient based on the adsorption position displacement statistics based on the second implementation log L2 and the distribution of the inspection measurement value M of the inspection device, and outputs the estimated second regression coefficient to the tape feed accuracy estimation unit 34 of the factor probability estimation device 30. The tape feed accuracy estimation unit 34 functions as an acquisition unit that obtains the second regression coefficient.

[0114] Note that the first implementation log L1, the second implementation log L2, and the inspection measurement value M shown in Figure 8 are data acquired using the same implementation machine 10.

[0115] Next, we will explain the process of performing estimations for the production of the measurement target. Figure 9 is a second flowchart showing the operation (production support method) of the production support system 1 according to this embodiment. The processes in steps S110 to S170 shown in Figure 9 are performed by the factor probability estimation device 30, and step S180 is performed by the display device 40.

[0116] As shown in Figure 9, the error count preprocessing unit 31 and the correction amount preprocessing unit 32 of the factor probability estimation device 30 each acquire the first implementation log L1 of the measurement target (S110). The error count preprocessing unit 31 and the correction amount preprocessing unit 32 function as acquisition units that acquire the first implementation log L1.

[0117] Next, the error count preprocessing unit 31 and the correction amount preprocessing unit 32 each perform preprocessing (S120). The error count preprocessing unit 31 performs preprocessing on the first implementation log L1 to generate, for example, the preprocessed data shown in Figure 2 and outputs it to the implementation error cause estimation unit 33. The correction amount preprocessing unit 32 also performs preprocessing on the first implementation log L1 to generate, for example, the preprocessed data shown in Figure 3 and outputs it to the tape feed accuracy estimation unit 34. In this embodiment, the correction amount preprocessing unit 32 outputs the preprocessed data, including the adsorption position misalignment statistics, to the tape feed accuracy estimation unit 34.

[0118] Next, the implementation error factor estimation unit 33 obtains the first regression coefficient from the nonlinear regression coefficient estimation unit 23 (S130) and estimates the error factor probability for each element (S140). As shown in Figure 7, the implementation error factor estimation unit 33 estimates the error factor probability for each element based on the pre-processed data from the error count pre-processing unit 31 (pre-processed data based on the first implementation log L1), the first regression coefficient, and the above equations 2 and 3. Figure 7 shows an example in which the implementation error factor estimation unit 33 outputs the feeder factor probability. The implementation error factor estimation unit 33 outputs the error factor probability for each component serial to the aggregation unit 35, the error factor probability for each nozzle serial to the aggregation unit 36, the error factor probability for each head spindle number to the aggregation unit 37, and the error factor probability for each feeder serial to the aggregation unit 38.

[0119] Referring again to Figure 9, the tape feed accuracy estimation unit 34 then obtains the second regression coefficient from the linear regression coefficient estimation unit 24 (S150) and estimates the tape feed accuracy of the feeder (S160). As shown in Figure 8, the tape feed accuracy estimation unit 34 estimates the tape feed accuracy of the feeder based on the suction position displacement statistics from the correction amount preprocessing unit 32 and the second regression coefficient. The tape feed accuracy estimation unit 34 estimates the tape feed accuracy in the X direction based on the suction position displacement statistics in the X direction and the second regression coefficient in the X direction, and estimates the tape feed accuracy in the Y direction based on the suction position displacement statistics in the Y direction and the second regression coefficient in the Y direction.

[0120] Referring again to Figure 9, the tape feed accuracy estimation unit 34 then estimates the accuracy rank of the feeder based on the estimated tape feed accuracy in the X and Y directions and the correspondence table shown in Figure 5 (S170). The tape feed accuracy estimation unit 34 outputs the accuracy rank to the aggregation unit 38.

[0121] Next, the display device 40 displays information from each aggregation unit 35 to 38 (S180). Figure 10 is a diagram showing a first example of the screen displayed by the display device 40 according to this embodiment. Figure 10 shows the results of determining the cause of the implementation error.

[0122] As shown in Figure 10, the display device 40 displays the error factor probability from the factor probability estimation device 30. Figure 10 shows an example of displaying the error factor probability for LOT01 to LOT03 for head spindle addresses "1" and "2". The addresses correspond to the serial numbers. The probability that the head spindle at address "1" is the cause of the mounting error that occurred in LOT01 is 10%, the probability that it is the cause of the mounting error that occurred in LOT02 is 5%, and the probability that it is the cause of the mounting error that occurred in LOT03 is 8%. Similarly, the probability that the head spindle at address "2" is the cause of the mounting error that occurred in LOT01 is 90%, the probability that it is the cause of the mounting error that occurred in LOT02 is 80%, and the probability that it is the cause of the mounting error that occurred in LOT03 is 100%. In this way, by displaying the error factor probability for each LOT on a single screen, the production support system 1 can assist the manager of the mounting machine 10 in determining which elements are normal and which are malfunctioning. Factor probabilities are displayed, for example, for each unit, but are not limited to this.

[0123] Furthermore, the display device 40 may display the average value of the error probability for each LOT (the period average shown in Figure 10). This makes it easier for administrators to determine which elements are normal and which are malfunctioning.

[0124] For example, in the case of LOT01, the sum of the failure probability of the head spindle at address "1" and the failure probabilities of the nozzle, component, and feeder corresponding to that head spindle at address "1" is 100%.

[0125] Figure 11 shows a second example of the screen displayed by the display device 40 according to this embodiment. Figure 11 shows the result of determining the tape feeding accuracy. Specifically, Figure 11 shows the accuracy rank of the tape feeding accuracy of the feeder (the rank shown in Figure 11).

[0126] As shown in Figure 11, the display device 40 displays the tape feed accuracy determination results from the factor probability estimation device 30. Figure 11 shows an example of displaying the tape feed accuracy determination results for LOT01 to LOT03 for feeder addresses "1" and "2". The addresses correspond to the serial numbers. For the feeder at address "1" (serial: FD0001), the accuracy rank for LOT01 is "B", and the accuracy ranks for LOT02 and LOT03 are "A". Similarly, for the feeder at address "2" (serial: FD0002), the accuracy rank for LOT01 is "D", the accuracy rank for LOT02 is "C", and the accuracy rank for LOT03 is "D". In this way, by displaying the accuracy rank of each LOT of the feeder on a single screen, the administrator of the mounting machine 10 can be helped to determine which feeders are functioning correctly and which are malfunctioning.

[0127] Furthermore, when the display device 40 displays the feeder estimation results, the error factor probability and accuracy rank may be displayed on the same screen. Also, the display device 40 may display the estimated value of tape feed accuracy in place of or along with the accuracy rank.

[0128] As described above, by estimating the probability of error factors and tape feeding accuracy, the state of the mounting machine 10 (the state of each unit) can be estimated with high accuracy without increasing costs (without adding hard sensors). This enables pinpoint maintenance, thereby reducing equipment performance loss (for example, short-term or long-term downtime of the mounting machine 10) at a low cost.

[0129] (Other embodiments) The above describes production support systems, etc., according to one or more embodiments, based on embodiments. However, this disclosure is not limited to these embodiments. Without departing from the spirit of this disclosure, various modifications to these embodiments that a person skilled in the art could conceive, or forms constructed by combining components from different embodiments, may also be included in this disclosure.

[0130] For example, in the above embodiment, the implementation error cause estimation unit was described as estimating the probability of implementation errors using logistic regression analysis, but other nonlinear regression analyses may also be used to estimate the probability of implementation errors. The implementation error cause estimation unit may estimate the probability of implementation errors using, for example, polynomial regression analysis, support vector regression analysis, etc.

[0131] Furthermore, although the above embodiment describes an example where the feeder is a tape feeder, other feeders such as bulk feeders may also be used. In this case, the factor probability estimation device does not need to include a tape feeding accuracy estimation unit.

[0132] Furthermore, the coefficient estimation device and the factor probability estimation device according to the above embodiment may be terminal devices located in a factory or the like where the mounting machine is installed, or they may be server devices located remotely from the factory.

[0133] Furthermore, the first and second regression coefficients according to the above embodiment may be estimated only once, or they may be estimated periodically, and the first and second regression coefficients used in the factor probability estimation device may be updated periodically. In addition, the first and second regression coefficients may be estimated after a predetermined event occurs, such as replacing a unit with a new one or repairing a malfunction.

[0134] Furthermore, the first implementation log and the second implementation log according to the above embodiment are also referred to as equipment logs.

[0135] Furthermore, although the above embodiment describes an example in which tape feeding accuracy is estimated based on suction position misalignment statistics, tape feeding accuracy may also be estimated based on, for example, either the recognition correction amount or the suction correction amount.

[0136] Furthermore, in the above embodiments, each component may be implemented by being composed of dedicated hardware or by executing a software program suitable for each component. Each component may also be implemented by a program execution unit such as a CPU or processor reading and executing a software program recorded on a recording medium such as a hard disk or semiconductor memory.

[0137] Furthermore, the order in which each step in the flowchart is performed is illustrative for the purpose of specifically illustrating this disclosure, and may be in a different order. Also, some of the above steps may be performed simultaneously (in parallel) with other steps, and some of the above steps may not be performed.

[0138] Furthermore, the division of functional blocks in the block diagram is just one example; multiple functional blocks can be implemented as a single functional block, a single functional block can be divided into multiple parts, or some functions can be moved to other functional blocks. In addition, the functions of multiple functional blocks with similar functions can be processed in parallel or time-sharing by a single piece of hardware or software.

[0139] Furthermore, the production support system according to the above embodiments may be implemented as a single device or as a plurality of devices. When the production support system is implemented as a plurality of devices, the various components of the production support system may be distributed among the plurality of devices in any manner. When the production support system is implemented as a plurality of devices, the method of communication between the plurality of devices is not particularly limited and may be wireless communication or wired communication. In addition, wireless communication and wired communication may be combined between the devices.

[0140] Furthermore, each component described in the above embodiments may be implemented as software, or typically as an integrated circuit (LSI). These may be individually integrated onto a single chip, or some or all of them may be integrated onto a single chip. Here, we refer to them as LSIs, but depending on the degree of integration, they may also be called ICs, system LSIs, super LSIs, or ultra LSIs. Moreover, the method of integrated circuit implementation is not limited to LSIs; it may also be implemented using dedicated circuits (general-purpose circuits that execute dedicated programs) or general-purpose processors. After LSI manufacturing, a programmable FPGA (Field Programmable Gate Array) or a reconfigurable processor that allows for the reconfiguration of the connections or settings of circuit cells inside the LSI may be used. Furthermore, if an integrated circuit implementation technology that replaces LSIs emerges due to advances in semiconductor technology or other derived technologies, it is naturally possible to integrate the components using that technology.

[0141] A system LSI is a highly functional LSI manufactured by integrating multiple processing units onto a single chip. Specifically, it is a computer system consisting of a microprocessor, ROM (Read Only Memory), RAM (Random Access Memory), and other components. The ROM stores the computer program. The system LSI achieves its function by operating according to the computer program, with the microprocessor performing its operations.

[0142] Furthermore, one aspect of this disclosure may be a computer program that causes a computer to perform each characteristic step included in the production support method shown in either Figure 6 or Figure 9.

[0143] Furthermore, for example, the program may be a program to be executed by a computer. Also, in one aspect of this disclosure, such a program may be recorded on a computer-readable non-temporary recording medium. For example, such a program may be recorded on a recording medium and distributed or made available. For example, by installing the distributed program on a device having another processor and having that processor execute the program, it becomes possible to have that device perform the above-mentioned processes.

[0144] (Note) Based on the above description of embodiments, the following technologies are disclosed.

[0145] (Technology 1) A production support system that estimates the probability that each of the multiple units was a contributing factor to an assembly error that occurred in an assembly machine composed of multiple units, An acquisition unit that acquires a first implementation log from the implementation machine, which includes information regarding the aforementioned implementation error and is the target for estimating the probability of the aforementioned implementation error, An implementation error factor estimation unit estimates the probability for each of the multiple units based on the first implementation log and a first estimation model based on the relationship between the number of implementation errors and the malfunction of each of the multiple units, The system includes an output unit that outputs the estimation results of the implementation error cause estimation unit, The first implementation log includes information regarding the number of units produced and the number of implementation errors. Production support system.

[0146] (Technology 2) The first estimation model is a multiple regression model in which the first regression coefficients for each of the multiple units, based on the distribution of information regarding the number of units produced and the number of implementation errors for each of the multiple units, are used as explanatory variables, and the probability is used as the dependent variable. The production support system described in Technology 1.

[0147] (Technology 3) The system further includes a first coefficient estimation unit that estimates the first regression coefficient for each of the plurality of units based on the distribution of information regarding the number of production units and the number of production errors contained in a second implementation log of the implementation machine, which was acquired before the first implementation log. Production support system as described in Technology 2.

[0148] (Technology 4) The output unit displays the probability for each of the multiple units related to the implementation error. A production support system described in any of the following technologies: 1-3.

[0149] (Technology 5) The plurality of units include a tape feeder for supplying components and a nozzle for picking up the components. The system further includes a tape feeding accuracy estimation unit that estimates the tape feeding accuracy based on the first implementation log and a second estimation model based on the relationship between the tape feeding accuracy of the tape feeder and the amount of misalignment of the component's suction position. The first implementation log further includes information regarding the control quantity for controlling the nozzle. A production support system described in any of the following technologies (1-4).

[0150] (Technology 6) The second estimation model is a simple regression model that includes a second regression coefficient showing the relationship between the tape feeding accuracy and the misalignment of the component's suction position. Production support system as described in Technology 5.

[0151] (Technology 7) The system further includes a second coefficient estimation unit that estimates the second regression coefficient for each of the plurality of units based on the distribution of the control amount for controlling the nozzle included in the third implementation log of the implementation machine, which was acquired before the first implementation log, and the measured value of the tape feeding accuracy of the tape feeder. Production support system as described in Technology 6.

[0152] (Technology 8) The output unit displays information regarding the tape feed accuracy estimated by the tape feed accuracy estimation unit. A production support system as described in any of Technical 5-7.

[0153] (Technology 9) The aforementioned mounting machine does not have sensors to directly measure the status of the multiple units. A production support system described in any of the technologies 1-8.

[0154] (Technology 10) The aforementioned plurality of units include a head spindle, a nozzle and a feeder, The assembly error cause estimation unit estimates the probability of each of the head spindle, the nozzle, the feeder, and the parts supplied by the feeder. A production support system described in any of the following technologies (1-9).

[0155] (Technology 11) A production support method for estimating the probability that each of the multiple units was a contributing factor to an assembly error that occurred in an assembly machine composed of multiple units, The first implementation log, which includes information regarding the aforementioned implementation error and is the target for estimating the probability, is obtained from the implementation machine. Based on the first implementation log and the first estimation model which is based on the relationship between the number of implementation errors and the malfunction of each of the multiple units, the probability for each of the multiple units is estimated. Output the estimated result, The first implementation log includes information regarding the number of units produced and the number of implementation errors. Production support methods.

[0156] (Technology 12) A program for causing a computer to execute the production support method described in Technical 11. [Industrial applicability]

[0157] This disclosure is useful for support systems and the like that assist in production using mounting machines. [Explanation of Symbols]

[0158] 1. Production support system 10 Implementing machine 20 Coefficient Estimation Device 21, 31 Error count pre-processing 22, 32 Correction amount preprocessing unit 23. Nonlinear regression coefficient estimation unit (first coefficient estimation unit) 24 Linear regression coefficient estimation unit (second coefficient estimation unit) 30-factor probability estimation device 33. Implementation Error Factor Estimation Unit 33a First Estimated Model 34 Tape feed accuracy estimation unit 34a Second Estimated Model 35, 36, 37, 38 Tallying Department 40 Display device L1 First Implementation Log L2 Second Implementation Log (Third Implementation Log) M Inspection Measurement Value

Claims

1. A production support system that estimates the probability that each of the multiple units was a contributing factor to an assembly error that occurred in an assembly machine composed of multiple units, An acquisition unit that acquires a first implementation log from the implementation machine, which includes information regarding the aforementioned implementation error and is the target for estimating the probability of the aforementioned implementation error, An implementation error factor estimation unit estimates the probability for each of the multiple units based on the first implementation log and a first estimation model based on the relationship between the number of implementation errors and the malfunction of each of the multiple units, The system includes an output unit that outputs the estimation results of the implementation error cause estimation unit, The first implementation log includes information regarding the number of units produced and the number of implementation errors in the plurality of units. The first estimation model includes a first regression coefficient for each of the plurality of units, based on information regarding the number of units produced and the distribution of the number of assembly errors for each of the plurality of units. The implementation error cause estimation unit estimates the probability in each of the plurality of units based on the first implementation log and the first regression coefficient of each of the plurality of units. The production support system further includes a first coefficient estimation unit that estimates the first regression coefficient for each of the plurality of units based on the distribution of information regarding the number of units produced and the number of assembly errors included in the second assembly log of the assembly machine, which was acquired before the first assembly log. The first coefficient estimation unit updates the first regression coefficient for each of the multiple units if any of the units included in the multiple units in the mounting machine are replaced or repaired. Production support system.

2. The output unit displays the probability for each of the multiple units related to the implementation error. The production support system according to claim 1.

3. The plurality of units include a tape feeder for supplying components and a nozzle for picking up the components. The system further includes a tape feeding accuracy estimation unit that estimates the tape feeding accuracy based on the first implementation log and a second estimation model based on the relationship between the tape feeding accuracy of the tape feeder and the amount of misalignment of the component's suction position. The first implementation log further includes information regarding the control quantity for controlling the nozzle. The production support system according to claim 1 or 2.

4. The second estimation model is a simple regression model that includes a second regression coefficient showing the relationship between the tape feeding accuracy and the misalignment of the component's suction position. The production support system according to claim 3.

5. The system further includes a second coefficient estimation unit that estimates the second regression coefficient for each of the plurality of units based on the distribution of the control amount for controlling the nozzle included in the third implementation log of the implementation machine, which was acquired before the first implementation log, and the measured value of the tape feeding accuracy of the tape feeder. The production support system according to claim 4.

6. The output unit displays information regarding the tape feed accuracy estimated by the tape feed accuracy estimation unit. The production support system according to claim 3.

7. The aforementioned mounting machine does not have sensors to directly measure the status of the multiple units. The production support system according to claim 1 or 2.

8. The aforementioned plurality of units include a head spindle, a nozzle and a feeder, The assembly error cause estimation unit estimates the probability of each of the head spindle, the nozzle, the feeder, and the components supplied by the feeder. The production support system according to claim 1 or 2.

9. The plurality of units include a tape feeder for supplying components and a nozzle for picking up the components, The first implementation log further includes information regarding a control quantity for controlling the nozzle, The production support system further includes a tape feeding accuracy estimation unit that estimates the tape feeding accuracy based on the first implementation log and a second estimation model based on the relationship between the tape feeding accuracy of the tape feeder and the amount of misalignment of the component's suction position. The estimation results include information for displaying, on the same screen, the probability of each of the multiple tape feeders and information regarding the tape feeding accuracy of that tape feeder. The production support system according to claim 1 or 2.

10. A production support method for estimating the probability that each of the multiple units was a contributing factor to an assembly error that occurred in an assembly machine composed of multiple units, The first implementation log, which includes information regarding the aforementioned implementation error and is the target for estimating the probability, is obtained from the implementation machine. Based on the first implementation log and the first estimation model which is based on the relationship between the number of implementation errors and the malfunction of each of the multiple units, the probability for each of the multiple units is estimated. Output the estimated result, The first implementation log includes information regarding the number of units produced and the number of implementation errors in the plurality of units. The first estimation model includes a first regression coefficient for each of the plurality of units, based on information regarding the number of units produced and the distribution of the number of assembly errors for each of the plurality of units. In estimating the probability in each of the plurality of units, the probability in each of the plurality of units is estimated based on the first implementation log and the first regression coefficient of each of the plurality of units. The aforementioned production support method is Based on the distribution of the production quantity information and the number of assembly errors included in the second assembly log of the assembly machine, which was obtained before the first assembly log, the first regression coefficient for each of the multiple units is estimated. If a unit included in the plurality of units in the mounting machine is replaced or repaired, the first regression coefficient of each of the plurality of units is updated. Production support methods.

11. A program for causing a computer to execute the production support method described in claim 10.

Images