A simulation model is developed to integrate the different components of the new baggage check-in system. Next the same model is used to evaluate the system performance, give feedback and suggest modifications. The simulation model is also used to carry out investigation on whether the system fulfils its objectives. Some of the actual data from the current timetables are used and implemented to study how well the system can perform. This information is also used as part of the unit integration tests.
Data
Data sets are required to feed the simulation model. These include baggage arrival pattern, the process time, and the number of check-in desks.
Baggage Arrival Pattern
Baggage arrival time and pattern can be estimated by studying the train timetable of Newcastle Central Station. According to the information provided by Realtime Trains, there were a total of 204 passenger services departing from Newcastle Central Station on 25 July, 2017 (Tuesday). The first train of the day leaves at 0445 to London Kings Cross with the last one at 2300 to Sunderland. However, the number of baggage items that is expected to be transferred from the passengers on to each service is different because of the fact that not all passengers would require it.
Routing Scaling Factors
A more realistic estimate to the number of baggage items can be made when we consider the number of passenger coaches used on each train service. For example, most of the First Transpennine Express services to Manchester and Liverpool are formed of three coaches and normally all of the Virgin Trains East Coast services to London and Scotland are formed of nine coaches. Therefore, it is expected that the number of baggage items that will arrive to the system would be of a 1:3 ratio.
Nevertheless, there are some exceptional cases, especially for Arriva Cross Country Trains, their Voyager trains Class 220/221 are formed of four or five coaches each, but they can be coupled to form longer trains, giving 8–10 coaches. Moreover, some of their departures are using Class 43 High Speed Trains, which is formed of seven coaches. Therefore, it may be hard to estimate these services. As a result, it is assumed that the average number of coaches for Arriva Cross Country Trains is 4.5, an average of four and five coaches from Voyager trains, as they are majority train type serving Newcastle Central Station. Therefore, it is expected that the number of baggage that will arrive to the system would be of a 1:1.5 ratio.
Apart from Arriva Cross Country Trains, Northern Rail services should also be noted. Most of the Northern Rail services are local stopping service which mainly used for commutation, this includes Tyne Valley Line connecting Newcastle to the West Coast at Carlisle, Tee Valley Line connecting Newcastle and Middlesbrough, as well as services to Chathill, which runs only two times a day. Also, their trains are usually formed of just two cars and many stations served by them are not staffed. It may be hard to sustain the baggage transfer service especially at the starting point. There are quite a number of services that terminates at the station after Newcastle Central Station, for example Morpeth, Sunderland and Metrocentre. These relatively short journeys can hardly attract passenger to use the baggage service. As a result, it is assumed that all of the Northern Trains services are assumed with no passengers using the baggage service.
For services going to Scotland via East Coast Main Line including Edinburgh, Glasgow, Aberdeen and Inverness, no matter which operators they are run by, the baggage ratio is set to 1:2. This value was taken from the average of nine coaches Virgin Trains East Coast trains and Arriva Cross Country Voyager trains of four coaches. This works out to be 6.5 coaches, which is approximately 1:2 when compared with the reference First Transpennine Express trains.
Day and Time Scaling Factors
It can be found that the usual weekday operation has more departures than Saturdays and Sundays. For example, 29 July 2017, Saturday, has 191 departures, and 30 July 2017, Sunday, has only 151 departures, both of which are less than 204, the weekday departure number. Therefore, if the system can cope with a weekday operation, then it should also be able to handle any weekend operations with no issue.
However, in the same day, but a different time slot, even for trains that are going to the same destinations, some are busier than others. This is especially obvious for very early and late departure time when compared with morning or evening peak. As a result, the amount of baggage for the system is different. In order to solve this problem, a time scaling factor can be added to the work entry point. The summary of the factors can be found in Table 2.
Table 2 Summary of the scaling factor with time Early morning (before 7 am) and late night (after 8 pm) has a factor of 0.5 only as these two periods are the quietest time in the station with a limited number of departures and, therefore, passenger use. The reason 0801–1100 has a factor of 1.5 is because the baggage system is intended to provide service for long distance intercity travellers. It is more likely that the traveller would choose morning departures in order not to arrive at their destinations late. For the same reason, the period of 1801–2000 is given a factor of 0.75 when it is less likely for passengers to be travelling for long journeys. The total number of departures taken into account is 122, but not 204 because all of the exemption departures as mentioned in Sect. 5.1.1.1 are not counted, which includes all of the departures from Northern Rail as well as the 2300 Virgin Trains East Coast departure to Sunderland. After scaling the departures, the total number of departures is reduced to 119. However, this only accounts for less than 2.5% change and is acceptable related to its validity to reality.
Summary of Baggage Arrival Pattern for Same Day Delivery
After having studied the timetable, it can first be assumed that one piece of baggage from a single passenger shall be transferred in each First Transpennine Express service to Manchester. Then the number of baggage items for the other services can be estimated. The summary of a list of major destinations, operators of the train services and the relative number of estimated baggage items is shown in Table 3.
Table 3 Summary of the scaling factor with destination
The factors affecting the number of baggage arrival are used to estimate the overall factor, which is the product of destination factor multiplied by the time factor. The overall factor is then rounded up to give a number of actual baggage items to be transferred for each departure. As a result, the total number of baggage transferring each day through the service proposed came up to 267.
Baggage Check-in for Departures in the Future
Apart from passengers requiring a baggage transfer service within the same day, some other customers have been identified that may use this service, namely, those requiring to send baggage items earlier, up to 2 days prior to their actual departure. The number of customers requiring such a service can be assumed to be 10% of the total number of baggage items.
As a start, there are 267 pieces of luggage each day; therefore, it can be assumed that 30 extra pieces of baggage will arrive into the system. For design purposes it is also assumed that the baggage arrives into the system evenly during each hour from 0900 to 1800.
Check-in Procedures
A number of operations describing the check-in procedure are considered to estimate the overall check-in time: completion and verification of personal details, payments and ticket checks as well as security check. The time allocated to each operations, which is used as an input for the simulation model, is shown in Table 4.
Table 4 Check-in procedures breakdown
It can be estimated that with online check-in, the total time for the completion of check-in procedure can be reduced by 50%. The time savings come from entering the personal details into the computer system by staff as well as online payment. However, a longer time would be needed in verifying the personal details. That is the reason why 30 s is added on ticket/ID check column with online check-in. It is also noted that there is an extra fixed 1 min added to both check-in ways as buffer time which results in 5 min for walk-in check-in and 2.5 min for online check-in. If half of the customers are using online check in, then the average time for check-in would be [(4 + 1.5)/2] + 1 = 3.75 min, hence this estimate is used to run the simulation model and study the respective effect on the system.
After check-in, there is an immediate security check through X-ray machine. The time allocated for the security check is an average of 1.5 min, which includes the transporting time from the check-in desk to the machine and the time for any checks on suspicious items.
Baggage to and from Trains
There are two ways of transporting the baggage to its destination: one is using scheduled passenger trains and the other one is using a freight train. The baggage is transported to the suitable train for departure by cart either pulled using manpower if the volume is small or pulled by small vehicles similar to the ones shown in Fig. 2. The carts are filled up with the baggage to be shipped at least 10 min before every assigned train departure by the staff at storage centre. Any loaded cart can return from the trains back to the storage centre. The staff there can then sort out the baggage items on the available shelves and wait for the passengers to reclaim their bags up to 2 days after arrival. Reclaiming after 2 days can also be considered.
Passenger Trains
A possible solution is to use passenger trains, by converting some small train compartments to transport more baggage. The main benefit of this solution is that the baggage can travel together with the passenger on the same train.
There are 122 departures included in the baggage transfer service. Some services are terminated at Newcastle Central Station meaning that some baggage items will be reclaimed by the passengers at Newcastle Central. In addition there are 43 extra services on weekdays that are assumed to terminate at Newcastle Central Station and one service that is going to Sunderland, but it is originated in London Kings Cross. This means that there would be baggage items arriving from that service as well. Similarly, there is one departure to London Kings Cross, which starts from Sunderland and this service is unlikely to have baggage items arriving at Newcastle Central. All of the departure and arrival times at Newcastle Central Station are used as input information to set up and calibrate the simulation model.
Freight Trains
The other solution is to use freight trains. However, freight trains used for bulk transport would not be very frequent. It is therefore unlikely that the baggage items can be transported without any modification of the current practice.
There are two directions that the baggage transfer system serves: one towards North to Scotland and one down south of Newcastle towards York (Fig. 3). Two freight trains can be introduced to transport bags along these routes to serve these two directions. It is adequate to consider that baggage items can further be served and transported to the end user when they have arrived into another railway station. It is likely that the freight train would run at night so there are fewer disruptions to regular passenger train services.
Baggage Reclaim
It is not easy to obtain a precise estimation of the number of baggage items arriving at Newcastle Central. This is because the bags arrive from numerous stations at different times. There are no data available either because there is no such system in operation in the UK at the minute. The time factor applicable to Newcastle Central Station does not quite comply with the time factors of other stations, because Newcastle Central is also a terminal station. Location of station also plays a role. We have seen that the total number of baggage items to be sent out from Newcastle Central Station was estimated to be a bit less than 300. Hence, it can be assumed that the amount of baggage arriving is similar. It is also suggested that each train has three bags to be reclaimed. With 125 trains arriving at Newcastle Central Station every day, the total number of bags comes out to 375.
The reclaim process is quicker than the check-in procedure because the staff must only check the ID of the customer and then return the baggage items from the storage area to the customer. The whole process should not take longer than 2 min on average if there are no queues.
When baggage is carried by regular passenger trains, the bags are available to be picked up 20 min after arrival. This is worked out by 8 min transporting time plus 10 min for sorting and storage at the designated point plus buffering time. Nevertheless, not all of the passengers pick up their bags straight away. It can be suggested that one passenger in every two trains leaves their baggage in the storage area overnight aiming to collect it the next day. Hence, the number of bags to be reclaimed on the same day is two and three per train. A total of 60 bags are left to be reclaimed the next day. The customers reclaiming bags are estimated to arrive at the Newcastle Central baggage collection point every 15 min from 0730 to 2200.
When baggage is carried by freight trains overnight, then the bags can only be picked up on the next day.
Storage Area
The staff working in the storage area load the carts with baggage items to be transported to the specified train and then unload the baggage items and store them in the designated area on the train. For allocation of baggage items to trains, a computer system should be implemented and used. The loading process should be completed within 10 min. It is noted from the time frame in Fig. 4 that the loading process begins before the cut-off time for the check-in. But in case there are any last minute arrivals of baggage items, a 5-min buffer time is implemented to ensure there is enough time for the completion of the loading process. Unloading is very similar to loading, hence it can be assumed that the time for both processes are the same, 10 min each.
Baggage Collection Point/Check-in Shop Design
Shop Layout
The check-in shop layout is designed as shown in Fig. 5. For the check-in side, there is a large area to accommodate customers waiting to be served. There is initially a design of three check-in desks, but this can be confirmed and changed later. Next to each counter is a luggage transporting belt, which ultimately connects to the X-ray machine at the end. If the bag passes the security check, it is then sent for storage next door.
There are two separate areas for delivery storage. A bigger one is for the same-day delivery and a smaller one is for the smaller number of customers who have checked in their bags up to 48 h prior to departure. There are also two separate areas for arrival baggage storage. These two areas function similarly, but can be separated into different origins so that the check-out process can be completed faster.
Storage Area
Within the storage area, there are racks in each column. The largest size of luggage available on the market is generally 32″, which has dimensions of 80 cm × 54 cm × 30 cm, thus it is important to apply some buffering when designing the racks for ease of placement and removal. Therefore, assuming each piece of baggage has the same size of 100 cm × 65 cm × 40 cm should be adequate. As such, the initial design of the rack has 100 cm depth and 65 cm height for each layer. There are three layers, so the top layer after placing the luggage does not exceed 2 m for health and safety reasons. The outline of the rack can be seen in Fig. 6.
The length of the rack varies depending on the size of the storage room. However, the minimum total length of racks for delivery should be over 40 m in order to accommodate 300 bags a day. This number should be multiplied by a factor of 1.5–2 for the future expansion of the service and prevention of a sudden, unforseen influx of baggage. Each position of the rack is numbered so that it can be linked to a computer system for more organised service. This reduces the chance of mistake and the baggage can be picked up and sent to the train more efficiently. The location of every baggage item is automatically assigned by a computer programme when checked-in at the front desk.
Staff and Resource Allocations
The number of check-in desks depends on the demand by potential customers and can be estimated considering the baggage arrival pattern. Figure 7 shows the number of departures and baggage for each hour. It can be seen that there is quite an even number of departures most of the time, from 0700 to 2100 weekdays. However, the baggage varies quite a lot with 9–11 am showing a spike on the graph where demands are maximum over the day. Other than that, 8–9 am and 12–8 pm shows a high usage of the system as well. This suggests that resources should be allocated accordingly. Therefore, the work shift of each check-in desk can follow this pattern as an initial input, and then after a review further adjustments with different trials can be undertaken to work out the best possible solution.
Simulation Modelling
Simulation modelling has been used to study the designed rail baggage transfer system performance. It is an evaluation technique, which helps us to understand how well the system has been designed. There is a vast amount of literature on simulation modelling that should be mentioned: [27, 47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80]. Four cases were modelled using SIMUL8. Case 1: a system of only one passenger. Case 2: freight trains being placed on the system where required, based on flight times from Newcastle Airport. Cases 3 and 4 focused on increasing track utilisation by saturating the system with freight trains. The study considered the age of the passenger, as well as the choice of transportation for different purposes of travel. For example, the average age of a leisure traveller is 47.5 years old, whereas the average age of a business traveller is 45.6 years old. Figures show that 79% of leisure trips were on the road using cars, whereas only 48% of business trips are by car [52]. These choices are influenced by the amount of luggage they travel with. Leisure trips require heavy luggage; however, business trips need less luggage.
A mesoscopic simulation modelling methodology is developed for analysing freight train operations. A rail network model was developed using SIMUL8. Two situations were examined for the study. Case 1: Freight train movement characterised with insignificant deviations from schedules. Case 2: Freight train movement characterised with significant deviations from schedules. The rail network model shows that the more structured and scheduled the network operation of freight trains with freight trains, the lower the queue in the rail network becomes. Hence, the amount of costs incurred for the company is lower, and vice versa [50].
Two potential scenarios that can occur at a railway yard were studied. Case 1: when the yard is situated between a high-speed track and a conventional track, which are both electrified. Case 2: when the yard is situated between an electrified high-speed track and a non-electrified conventional track. The first design presented was based on an interchange between an electrified high-speed line and an electrified conventional line. The second design was an interchange between an electrified high-speed line and a non-electrified conventional line. It is possible to operate a conventional train on a high-speed track through the use of an interchange yard. A secondary conclusion was that issues such as a reduced line capacity due to conventional rolling stock with a lower maximum top speed than high-speed rolling stock did occur due to operating conventional stock on a high-speed line [70]. Simulation packages with a rail focus include RailSys and OpenTrack. Other simulators and performance analysers include: XpressMP, ARENA, SIMUL8. A comprehensive discussion on simulation modelling applications for designing baggage transfer services is presented in [81] by Yeung H and Marinov M. (2017). It is suggested in that paper that SIMUL8 is the good choice of software for modelling the baggage transfer system; therefore, we shall not repeat this discussion here. Instead we now look at how SIMUL8 [51] operates and has been implemented to study the performance of the newly designed system for baggage transfer at Newcastle Central. Before we introduce the simulation model developed, we would like to remind the reader about SIMUL8’s building blocks:
-
1.
Work Entry Point: Arrival of work items to the system and the arrival pattern can be deterministic or stochastic behaviour.
-
2.
Queue: The point where the work items are waiting for the next process.
-
3.
Work Centres: System servers or machines where process is taking place here and the output of this stage will be passed to other point of system for further working, storage or direct delivery through work exit point.
-
4.
Storage Point: As a buffer or a queue to gather (semi-)finished work items for moving up to the next process
-
5.
Work Exit Point: The point where the work items leave the system and where the service or process is meant to be completed. There can be over one Work Exit Point, for example, one exit point to the delivery of items while some defective items are sent for scrap and disposed.
-
6.
Work Items: The object that are brought to the work entry point for further process along the simulation modelling system.
-
7.
Resources: Any machines, employee, operator, signal etc. that requires fulfilling the tasks and processes at different work centres.
In the case of baggage transfer service, the work item is defined as the baggage to be or that has been transferred. The resources would be anything that is in the system including check-in desk, staff member, X-ray machine for security, carts for movement of baggage around the station, storage racks etc. The entry point is the check-in desk at the station where customers drop off their bags or the arrival of bags from other destinations prior to customer pick up. The queue is where the luggage is waiting to be checked in on arrival. The storage point is the racks behind the check-in point where the luggage is waiting to be shipped or picked up by customers. The work exit point is the point where the baggage leaves the check-in centre for shipment or pick-up.
There are a number of trials performed to study the system performance in different ways.
Check-in of Baggage
The first part of the modelling is for the check-in process. The aim of this simulation is to test if the check--in system is able to handle the baggage and to find the best possible resource allocation in terms of staffing check-in desks. Also, it can be used to study the effect on walk-in check-in and online check-in by inputting different sets of parameters. The clock has been set to run from 04:00 for 19 h daily until 23:00. This covers all of the check-in process as the last train with baggage transfer service leaves at 22:46. The warm-up period is set to 0 as there should not be any process that must be done prior to serving customers. The target completion time for the whole check-in process is ideally 15 min and 20 min as the bottom line. The general model for SIMUL8 is set like below in Fig. 8.
There are two work inputs, one representing the baggage for passengers with same-day departure, and the other is for passengers of future departures. They are followed by a queue where they wait to be served by the staff at the check-in desk. There are three check-in desks, which work in different shifts for better utilisation of the system. After the check-in process, the baggage is sent for a security check through conveyer belts, followed by storage prior to shipment.
Figure 9 shows the data input for the check-in desks. They follow different time shifts, and the best shifts are worked out after different trials in the simulation.
Transport of Baggage within Station
The next part of the modelling is for the movement of the baggage within the station to and from the train. The aim of this simulation is to find the best possible resource allocation in terms of staff for moving the bags. It can be used to evaluate whether passenger trains or freight trains provide a better service in terms of cost and time.
The clock has been set to run from 00:00 for 24 h daily because the last train to arrive is beyond 2am and the first train to depart is 4 am so it is more convenient to be set running round the clock. The warm-up period is set to 0 as the system is running 24 h a day non-stop. The target check-in process is 15 min, and this allows 15 min for loading to the carts and transporting to the designated train. The general model for the software is set as below in Fig. 10.
The time set for moving from storage point to the train platform is set to be 10 min, which includes 2 min buffering time. Eight minutes is enough for movement across any platform at the station although some take much shorter time due to distance. Loading and unloading to and from the train takes 1 min each as it has been studied in the timetable that some of the services would only stop at the station for 2 min. The return time to the storage point is set to be 8 min and does not need any buffering time. The general idea for the flow of path is illustrated below in Fig. 11.
There are three types of train services, which set Newcastle Central Station as: 1-First Stop; 2-Last Stop; 3-Mid-Stop. These are assigned to different label values in the software as in Fig. 12.
Each relates to the processes that must be undergone. 1—Omits the unloading process from train; 2—Omits the loading process to train; 3—Takes the full process. This affects how the routing path is set in the simulation modelling as shown in Fig. 13.
Baggage Reclaim
The third part of the simulation model is about the baggage reclaiming process. It is very similar to the check-in process, but without a security check and the time for check-out is much shorter than check-in. There are three types of customers, but basically they are just the same except for the arrival pattern. The overall modelling setup can be seen in Fig. 14.
The clock has been set to run 24 h as the last train arrives after 2am. The average process time for each check-in desk is set to be 2 min.
Storage Area
The last part of the simulation is a very simple setup as the aim of this model is to find the number of staff needed to complete the loading and unloading procedures of the carts. The general model can be seen in Fig. 15. Because the loading and unloading processes both take the same time, in order to keep the model simple, only one activity was set. The time for the activity was set fixed 10 min and the clock was set to run 24 h due to the train schedule.