Data analysis device and method
A data analysis device, for analyzing efficiency of an assembly line operation including processes, includes: an input interface configured to acquire log data indicating a history of the processes respectively performed for each time of the assembly line operation; and a control circuit configured to generate analysis information indicating an analysis result of the history indicated by the log data, based on information generated by a probability model to calculate probability distribution for the assembly line operation. The probability model is configured to: generate process efficiency distribution indicating probability distribution of efficiency for each process in the assembly line operation, and variation factor distribution indicating probability distribution of a factor probable to vary the efficiency of the assembly line operation in each process; and generate work efficiency distribution indicating probability distribution of the efficiency of the assembly line operation, based on the process efficiency distribution and the variation factor distribution.
1 . A data analysis device for analyzing efficiency of an assembly line operation including a plurality of processes, the data analysis device comprising:
an input interface configured to acquire log data indicating a history of the processes respectively performed for each time of the assembly line operation; and
a control circuit configured to generate analysis information, based on information generated by a probability model to calculate probability distribution with respect to the assembly line operation, the analysis information indicating an analysis result of the history indicated by the log data,
wherein the probability model is configured to:
generate process efficiency distribution and variation factor distribution, the process efficiency distribution indicating probability distribution of corresponding efficiency to each process in the assembly line operation, and the variation factor distribution indicating probability distribution of a factor in each process wherein the factor is probable to vary the efficiency of the assembly line operation; and
generate work efficiency distribution, based on the process efficiency distribution and the variation factor distribution, the work efficiency distribution indicating probability distribution of the efficiency of the assembly line operation,
wherein the control circuit is configured to generate the analysis information based on a difference between the work efficiency distribution and supplementary work efficiency distribution, and the supplementary work efficiency distribution is calculated by inputting observed efficiency or factor on at least one of the plurality of processes in the history indicated by the log data into the probability model as alternative of a corresponding part of the process efficiency distribution or the variation factor distribution.
2 . The data analysis device according to claim 1 , wherein
the probability model is configured to generate the work efficiency distribution by weighting the variation factor distribution for each of the processes, based on a coefficient preset for each of the processes.
3 . The data analysis device according to claim 1 , wherein the control circuit is configured to:
detect efficiency variation based on probability by which the work efficiency distribution allows to cause the efficiency of the assembly line operation in the history indicated by the log data; and
generate the analysis information according to a detection result of the efficiency variation.
4 . The data analysis device according to claim 1 , wherein
the probability model is configured to generate the supplementary work efficiency distribution with observation of a specific factor, by using an observation value of the specific factor in the history indicated by the log data, instead of the variation factor distribution for the specific factor, and
the analysis information includes information evaluating an influence of the factor on the efficiency of the assembly line operation according to the difference between the work efficiency distribution with the observation of the specific factor and the work efficiency distribution without the observation of the specific factor.
5 . The data analysis device according to claim 1 , wherein
the probability model is configured to generate the work efficiency distribution, based on lowest efficiency distribution indicating probability distribution of lowest efficiency among corresponding efficiency to the plurality of processes respectively.
6 . The data analysis device according to claim 5 , wherein
the probability model is configured to generate the supplementary work efficiency distribution with observation of corresponding efficiency to a specific process, by using an observation value of the corresponding efficiency to the specific process among the corresponding efficiency to the plurality of processes respectively in the history indicated by the log data, instead of the lowest efficiency distribution, and
the analysis information includes information evaluating an influence of the corresponding efficiency to the specific process on the efficiency of the assembly line operation according to the difference between the work efficiency distribution with the observation of the corresponding efficiency to the specific process and the work efficiency distribution without the observation of the corresponding efficiency to the specific process.
7 . The data analysis device according to claim 1 , wherein
the assembly line operation is work to produce an article in a production line including a plurality of facilities for performing the processes respectively, and
the factor in each process includes at least one of an error stop of a facility for performing the process, or occurrence of a defective product for the article.
8 . The data analysis device according to claim 1 , further comprising a memory configured to store model data indicating the probability model,
wherein the control circuit is configured is to execute calculation as the probability model to generate the work efficiency distribution, based on the model data stored in the memory.
9 . The data analysis device according to claim 1 , wherein
the probability model is constructed by Bayesian inference based on a past history of the assembly line operation.
10 . A data analysis method for analyzing efficiency of an assembly line operation including a plurality of processes, the data analysis method comprising:
acquiring log data indicating a history of the processes respectively performed for each time of the assembly line operation; and
generating analysis information, based on information generated by a probability model to calculate probability distribution with respect to the assembly line, the analysis information indicating an analysis result of the history indicated by the log data,
wherein the probability model
generates process efficiency distribution and variation factor distribution, the process efficiency distribution indicating probability distribution of corresponding efficiency to each process in the assembly line operation, and the variation factor distribution indicating probability distribution of a factor in each process wherein the factor is probable to vary the efficiency of the assembly line operation, and
generates work efficiency distribution, based on the process efficiency distribution and the variation factor distribution, the work efficiency distribution indicating probability distribution of the efficiency of the assembly line operation,
wherein the analysis information is generated based on a difference between the work efficiency distribution and supplementary work efficiency distribution, and the supplementary work efficiency distribution is calculated by inputting observed efficiency or factor on at least one of the plurality of processes in the history indicated by the log data into the probability model as alternative of a corresponding part of the process efficiency distribution or the variation factor distribution.
11 . A non-transitory computer-readable recording medium storing a program for causing a computer to execute the data analysis method according to claim 10 .
12 . A data analysis device for analyzing efficiency of an assembly line operation including a plurality of processes, the data analysis device comprising:
an input interface configured to acquire log data indicating a history of the processes respectively performed for each time of the assembly line operation; and
a control circuit configured to generate analysis information, based on information generated by a probability model to calculate probability distribution with respect to the assembly line operation, the analysis information indicating an analysis result of the history indicated by the log data,
wherein the probability model is configured to:
generate process efficiency distribution and variation factor distribution, the process efficiency distribution indicating probability distribution of corresponding efficiency to each process in the assembly line operation, and the variation factor distribution indicating probability distribution of a factor in each process wherein the factor is probable to vary the efficiency of the assembly line operation; and
generate work efficiency distribution, based on the process efficiency distribution and the variation factor distribution, the work efficiency distribution indicating probability distribution of the efficiency of the assembly line operation, wherein
the probability model is configured to generate further work efficiency distribution with observation of a specific factor, by using an observation value of the specific factor in the history indicated by the log data, instead of the variation factor distribution for the specific factor, and
the analysis information includes information evaluating an influence of the factor on the efficiency of the assembly line operation according to a difference between the work efficiency distribution with the observation of the specific factor and the work efficiency distribution without the observation of the specific factor.