326
T. Nagahata et al.
necessary to promote the so-called 3Rs: reuse, reduce and recycle. Especially, reuse,
which can reduce the energy consumption and generation of waste is emphasized.
For an efficient reuse, it is important to control the life cycle of each part. However,
it is difficult to manage the life cycle of used parts because the behavior of the users
is diverse, and it is difficult to predict the state of parts. To solve this problem, we
are developing a part agent system that manages the state of parts and supports the
maintenance action of users (Yokoki and Hiraoka 2017). A part agent system is a
system for managing the life cycle of parts by making the network agent follow the
actual parts.
The deterioration of parts varies by various factors such as use environment and
use time, and their maintenance actions are also diversified. Moreover, it is generally
difficult to observe the progress of deterioration directly. Therefore, it is necessary
to estimate the progress of deterioration by some method based on the behavior of
users and observable events, and on this basis decide the maintenance actions. In
this study, we propose a method to determine the maintenance action by creating a
causality model between the states of parts and estimating the probability of occurrence of a failure using Bayesian estimation based on the model (Nagahata et al.
2018). The causality model between the state of parts is developed by combining
the deterioration process and a deterioration simulation, and the failure probability
of parts is calculated supported on it. We then propose a method to determine the
maintenance action by utilizing the probability for the life cycle simulation (LCS).
23.2 Part Agent System
A part agent is a system in which an agent dedicated to each product or part collects
information such as the state of the product or part, manufacturing information, and
usage history, and presents it so that a user can make an appropriate judgment on the
maintenance of the product or part.
Figure 23.1 shows a conceptual diagram of the part agent. By attaching an RFID
tag (Gaetano 2005) to each part, the part agent follows the movement of the part
along the production factory, repair factory, consumer route, retail store, etc. in the
network. The correspondence is acquired through the RFID tag, and information such
as the deterioration level of the parts is acquired through the network. Based on the
information, the agent devises the most appropriate treatment for the corresponding
part at that time and proposes it to the user.
T. Nagahata et al.
necessary to promote the so-called 3Rs: reuse, reduce and recycle. Especially, reuse,
which can reduce the energy consumption and generation of waste is emphasized.
For an efficient reuse, it is important to control the life cycle of each part. However,
it is difficult to manage the life cycle of used parts because the behavior of the users
is diverse, and it is difficult to predict the state of parts. To solve this problem, we
are developing a part agent system that manages the state of parts and supports the
maintenance action of users (Yokoki and Hiraoka 2017). A part agent system is a
system for managing the life cycle of parts by making the network agent follow the
actual parts.
The deterioration of parts varies by various factors such as use environment and
use time, and their maintenance actions are also diversified. Moreover, it is generally
difficult to observe the progress of deterioration directly. Therefore, it is necessary
to estimate the progress of deterioration by some method based on the behavior of
users and observable events, and on this basis decide the maintenance actions. In
this study, we propose a method to determine the maintenance action by creating a
causality model between the states of parts and estimating the probability of occurrence of a failure using Bayesian estimation based on the model (Nagahata et al.
2018). The causality model between the state of parts is developed by combining
the deterioration process and a deterioration simulation, and the failure probability
of parts is calculated supported on it. We then propose a method to determine the
maintenance action by utilizing the probability for the life cycle simulation (LCS).
23.2 Part Agent System
A part agent is a system in which an agent dedicated to each product or part collects
information such as the state of the product or part, manufacturing information, and
usage history, and presents it so that a user can make an appropriate judgment on the
maintenance of the product or part.
Figure 23.1 shows a conceptual diagram of the part agent. By attaching an RFID
tag (Gaetano 2005) to each part, the part agent follows the movement of the part
along the production factory, repair factory, consumer route, retail store, etc. in the
network. The correspondence is acquired through the RFID tag, and information such
as the deterioration level of the parts is acquired through the network. Based on the
information, the agent devises the most appropriate treatment for the corresponding
part at that time and proposes it to the user.