Pruebas unitarias de la aplicación

Firstly, the influence of the different layout configurations considered on the recycle time of failed AGVs is investigated in this Section. It is deemed that the use of a separate maintenance site (see Figure 5.1(b) and (c)) is helpful to reduce conflict and deadlock, thereby increasing the efficiency of the recycling, although more space and extra routes will be needed to accommodate the separate maintenance site. This is because the separate maintenance site will require extra space to conduct corrective maintenance, and it needs to be connected to other stations via extra paths, as illustrated in Figure 5.1. The length of the extra paths will be normalised with the ratio to the equivalent path length between the stations. The average recycle time in the scenarios of using the 3 different layout configurations in Figure 5.1 is calculated and the calculation results are listed in Table 5.2. where the existence of the separate maintenance site is indicated by 1.

Table 5.2 Recycle time

Location indicated by Average recycle Time (hours) Extra Space (unit) Length of extra route required (unit) Figure 5.1(a) 0.13162 0 0 Figure 5.1(b) 0.12851 1 √3/2 Figure 5.1(c) 0.10075 1 3√3/4

From Table 5.2, it is found that when the maintenance site is placed in the centre (see Figure 5.1(c)), the shortest recycle time can be achieved, but at the cost of extra space and the longest extra route length. When the maintenance site shares the same site with the AGV base (see Figure 5.1(a)), the system does not require extra space and extra routes, however such a layout configuration will lead to the longest recycle time. The reason for the increased recycle time is because it is assumed that the recycling process can be started only when there is no AGV running on the recycle path. When the maintenance site is placed between the AGV base and the storage (see Figure 5.1(b)), extra space is required with the compromised values of recycle time and the length of an extra route. From these simulation results, it can be concluded that the location of

the maintenance site will directly affect the recycling, space, and route. Therefore, the maintenance site location will have significant influence on the performance and cost of a multi-AGV system and should be considered early in the design.

The influence of different maintenance strategies on the performance of the multi- AGV system is also investigated in this Section. Assume the operation time of the system is 10 hours every day and the layout configuration illustrated in Figure 5.1(b) is adopted, the number of completed missions obtained when using different maintenance strategies is calculated and the calculation results are listed in Table 5.3. In the table, T is the time interval of periodic maintenance; P is the percentage of AGVs failed within the time interval if there is no maintenance (%); N1 is the number of missions completed per year with periodic but without corrective maintenance; N2 is the number of missions completed per year with both periodic and corrective maintenance; DT1 is the percentage of downtime per year with periodic but without corrective maintenance (%);

DT2 is the percentage of downtime per year with both periodic and corrective

maintenance (%). In the calculation, the operation time of the system per day is assumed based on the average working hours of workers as the operation of AGVs is usually supervised by human. A fully automated system can run up to 10 hours a day. But this would be different in different applications.

Table 5.3 Number of completed missions

𝑻 𝑷 𝑵𝟏 DT1 𝑵𝟐 DT2 7 days 0.03 11518 30.58 11840 28.64 20 days 1.10 13213 20.36 14709 11.34 1 month 3.93 12840 22.61 15264 8.00 2 months 18.06 11028 33.53 15792 4.82 3 months 36.32 9372 43.51 15972 3.73 4 months 53.37 7983 51.88 16059 3.21 6 months 77.34 6084 63.33 16142 2.71 12 months 98.06 3280 80.23 16234 2.15

From Table 5.3, it is found that more than 98% of AGVs will fail within 12 months or after completing 3280 missions if without conducting any maintenance. This highlights the importance and necessity of conducting appropriate maintenance of the AGVs during their service. It can be imagined that the number of completed missions will increase if the AGVs can receive periodic maintenance service. This has been proved by the simulation results listed in Table 5.3. But there must exist an optimal interval for conduct periodic maintenance. In other words, both a too long and too short maintenance interval will not lead to the maximum productivity of the system. This is because too many AGVs would fail during the period if the maintenance interval is set to be too long, while the frequent periodic maintenance may cause long downtime if the maintenance interval is set to be too short. From Table 5.3, it is found that 20 days may be the best value of the periodic maintenance interval, which leads to the maximum number of completed missions (i.e. 13213) in a year. In reality, the AGV manufacturers and suppliers usually provide 2 to 6 times of planned maintenance every year. For the system analysed in this example, only 6084 missions can be successfully completed if 2 periodic maintenances are arranged in a year. This means that more than 7,000 missions cannot be completed.

Figure 5.14 Missions completed per year

Furthermore, the comparison of the corresponding values of N1 and N2 has shown that the long-term system efficiency can be further improved by using both periodic

0 2000 4000 6000 8000 10000 12000 14000 16000 18000 Num ber o f m is sio ns co m plet ed

Time interval of periodic maintenance

Without Corrective Maintenance

maintenance and corrective maintenance strategies in combination to take care of the AGVs in the multi-AGV system, although this could induce additional financial and labour costs. The advantage of using both maintenance strategies can be readily observed from the two curves in Figure 5.14, which are plotted using the N1 and N2 data listed in Table 5.3.

In summary, from the above discussion of the simulation results it can be concluded that both the location of the maintenance site and maintenance strategies have a significant influence on the performance of a multi-AGV system. Therefore, they should be optimised early at the design stage of the system.

5.11 Conclusions

This chapter has presented the methodology using CPN models to simulate the design, operation, and maintenance strategies of a multi-AGV system. From the research results described above, the following conclusions can be drawn:

1. The CPN modelling is a valid approach to simulate the design, operation, and maintenance of a multi-AGV system. The simulation results are very helpful for assessing the mission performance, evaluating routing, and planning the maintenance strategies in a particular design of multi-AGV system.

2. In the CPN models, both tokens and transitions can be allocated specific properties using different colours. Such a unique feature of the CPN makes it more powerful and flexible in performing simulation, thereby greatly simplifying the development of CPN models.

3. The application of the CPN facilitates the investigation of the influences of the location of maintenance site and the optimal interval for conducting periodic maintenance on system performance. It has been demonstrated that they do have a significant influence on the performance of a multi-AGV system and should be optimised early at the design stage of the system.

4. Long-Term high efficiency of a multi-AGV system can be achieved by using the periodic and corrective maintenance strategies together to take care of the AGVs. When only periodic maintenance strategy is adopted, the system

performance is very sensitive to the time interval for conducting periodic maintenance. It has been proved that both a too long and too short time interval will lead to low productivity of the system.

It should be noted that all of the assumed values, such as path lengths and working hours, can be modified for different applications. The AGV system considered above is relatively simple and the loading capacity of individual AGVs in the system is not considered either. For these reasons, the CPN models will be further improved in the next Chapter to investigate the operation of multi-load AGVs in a larger AGV system with more stations and bidirectional paths.