Introduction

The current world is following strategies to exit from energy crisis. In Bangladesh changing the vehicle fuel to the Compressed Natural Gas (CNG) is the priority enforcement. Recently, the number of CNG fuel based vehicles in Bangladesh is growing rapidly due to low price and availability of CNG. This change needs some planned substructures. One of the most important substructures is the CNG refueling stations distributions. Necessity of making a systematic process to locate CNG stations is vital because at present there is no fixed system to locate CNG stations throughout the country.
Waiting or queuing is a common phenomenon in service organization like CNG filling stations. In general, the vehicles do not like to wait in the filling station. But reduction of the waiting time usually requires extra investment. To decide whether or not to invest, it is important to know the performance of the existing distribution system. So we need models and techniques to analyze such situations. Though the queuing analysis is performed in different hospitals and dental clinics in Bangladesh but first time this research treats single server queuing system to analyze the performance of the existing CNG refueling stations. Our attention is restricted to models with single queue and single server. The results find that there is an unplanned location of CNG filling stations in two intercity networks and suggest the optimum number of CNG refueling stations in that intercity network.
Vehicles run in Sylhet-Dhaka road (Length: 265 km).

The overall scenario of four CNG filling stations in Sylhet-Jaflong network is shown in the Table. The results of Table 4 show that no stations are fully utilized. This support the results found in section 4.1 that there is excess CNG stations in this particular road.

Analysis for Sylhet-Dhaka Network
We did the same calculation in Sylhet-Dhaka road. The length of the road is 265 km. But the four strokes and leguna do not travel the total length. They travel a limited distance in the local city. The average travelling distances are about to 80 km. The numbers of vehicles are mentioned below.
There are a number of buses or trucks run by diesel/petrol on that road. The number of buses or trucks is approximately:
Total number of buses = 2900 (run by diesel)
Total number of trucks = 3300 (run by diesel)
The fuel demand for Sylhet-Dhaka Network is estimated to 404539.55 m3 and the number of required CNG station on the basis of calculated demand in Sylhet-Dhaka road is equal to twenty four stations.
According to the existence of twenty one CNG stations in the road so there is need of locating and constructing three other stations. We see that among the 21 CNG station, 18 are located in the part 85 km from Dhaka side and 3 are located in the remaining 180 km part (Bhairan-Sylhet road). For this reason we also estimate the optimum number of CNG station in Bhairab-Sylhet road.

Model:

Queuing model in non standard mode happens due to limitations of the particular layout, the customer for getting the service from servers is faced with restrictions. For analyzing these systems some parameters like waiting time and average length of queuing in such non standard systems, should be calculated. In such non-standard systems, time delay can exist in the service. These models can be seen in fueling stations or assembly and production line, so fueling stations are one of the biggest and the most efficient queuing line in the country and with considering to great amount of investment for settling the station has the special importance that will be analyzed.
Selection of branches with equal probability is done. After choosing the branch, we cannot change it. Two servers are serving in series and in certain condition. When both servers are free, the customer refer to the first server and when the first server is busy, that customer refer to the second server and if that is busy, he will refer to the next server and this trend will continue to the server.

Actual or the equivalent system Schematic of proposed CNG Station
Single Server queuing system:

A queuing system can be simple described as customers arriving for service, waiting for service if it is not immediate and if having waited for service, leaving the system after being served.
Simulation:

A Flow-chart is given below:

Recommendation:

As we said we have departure and substitution constraints in non-standard queuing systems duo to the layout constraint. There is waste of valuable serving capacity because customers leave the system or occupy free servers with delay. In fact this results in delayed serving. For example assume first car in all branches are being served while all the other automobiles have been served but they are not able to leave the system. Know assume one of these branches with 2 servers i.e., k = 2 and have departure constraint. If each of servers serve n customers in a time frame of we can consume by each server to serve each of n customers. It is evident from the figure that although second server can serve customers faster but it have to wait and is free unnecessarily. This is true in state of substitution constraint as well but instead of second server, first one has to wait this time. It will be an improvement in performance of fuel stations if we are able to reduce these delayed servings. These improvements can be in terms of: * Reduction of waiting time in queue and system * Shorter queues * Increase of systems’ customers * Induction in pace of serving * Customer satisfaction improvement * Increase of presented service in a given

In spite of high investment needed for construction of a fuel station, improper location, layout of nozzles, low width of aisles, insufficient space for cars to utilize free nozzles and etc, will cause high waiting time which is main reason of longer queues and high fuel waste duo to the running engines in queues. This fact reveals the necessity of studying these systems. The proposed model has many applications in a wide range of queuing systems. This queuing model with nonstandard state leads to the large queue length and large waiting time. We want to present a scenario in which the system remains empty immediately for the existing cars in the queue after each service accomplishment. Therefore, we could remove the wasting time between each two consecutive services, which is the main defect of nonstandard queuing systems. It seems that we cannot increase the productivity of a gas station by simply adding more pumps in the system. In order to study the behavior of our system, we have used Show Flow simulation tool to simulate a scenario. It is evident that we need to know some information about system to be able to analyze it like arrival rate and serving status or we have to use sampling to estimate them. However the success of our estimation is dependent on some factors. These factors are classified into major groups. First, are those concerning location of samplings and second are those related to time of samplings. After these factors have been determined, locations and times for sampling, minimum sample size, arrival rate and serving rate were clarified by the help of Design of Experiments (DOE).

Conclusion:

The optimum number of CNG stations in two intercity networks is estimated using the fuel demand method. An unplanned location of CNG stations in those two intercity networks in Bangladesh is found. Again, using a single server queuing system, some performance measures of selected existing CNG stations are estimated. This study finds that utilization of servers is 100% in one road where locating of more CNG stations might avoid the long queue. Again utilization of servers is not full in another road even excess number of stations is already located. Suitable locating of CNG fuelling station would result in saving resources and expenses, reduction of delay time and increasing of the security.