Analysis of road performance due to trip generation from the development of Batang Integrated Industrial Estate

The development of the Batang Integrated Industrial Estate (KIT Batang) is predicted to affect the performance of the surrounding roads. This study aims to analyze roads performance around industrial estate by referring to the Indonesian Highway Capacity Manual 1997 using a “modified four-step model” (overriding the mode choice stage). The result shows that in 2024, the presence of KIT Batang will not have a significant impact on the surrounding roads network. However in 2029, there will be a significant changes, indicated by worsening road performance—in this case referring to travel speed, service level, and volume-capacity ratio—where toll roads have shown an average level of service F. The do-something scenario is carried out by widening the toll and arterial roads in 2029. Analysis of simulation results shows improved performance on toll roads with an average service level from F to C and on arterial roads from C to B.


Introduction
In general, KIT Batang has an area of 4,300 ha.KIT Batang itself is designed to have three clusters, where currently the construction has entered phase 1 in cluster 1 covering an area of 450 ha, where the area being built is allocated as textile and battery industrial area.The supporting infrastructure at the Batang KIT phase 1 itself is targeted to operate in the first half of 2024 which includes gas, electricity, raw water and other supporting infrastructures, so that tenants can start operating in 2024 [1].
With the presence of the KIT Batang development, it will certainly have an impact on current and future traffic conditions.Therefore, further studies and research are needed regarding the performance of roads due to the development of industrial area infrastructure.

The Four-Step Model of Transportation Planning
The most popular and frequently used concept of transportation planning is the "four-step model of transportation planning".This model consists of Trip Generation-Trip Distribution-Mode Choice-and Route Choice.The trip distribution stage produces the origin-destination (OD) matrix, which can be obtained using the Furness method as in equation (1) [2].
=     and   =     (1)  = Total present movement from origin zone i to destination zone d   ,   = The growth rate of origin zone i and destination zone d At the route choice stage, the route is selected based on the distribution of trips by assuming that each driver chooses a route that minimizes their travel costs where the following generalized cost equation is used.The capacity restrained (CR) method is used in route choice.The route is chosen by considering the capacity (section) as the limit or impedance of the vehicle's trip when passing through a road section.The volume-delay function equation from the BPR formula is used as in equation (3).

Road Performance
The capacity (C) and free flow speed (FV) are measured to obtain road performance.Capacity is the maximum traffic flow through a point on the path that can be withheld per unit hour under certain conditions.Free flow speed (FV) is the speed at zero vehicle flow rate or the speed that the driver will choose if driving without being influenced by other motorized vehicles.When the number of traffic flows (Q) is known, the degree of saturation value can be determined by comparing the value of traffic flow to its capacity so that the level of service of a road section can be identified [3].

Research Methodology
This study only relies on the secondary data obtained form related authorized agencies.The modeling process carried out using a simplified four-step model in transportation planning by eliminating the mode choice stage due to secondary data obtained is already in vehicle trip units.The Semarang Raya traffic model from the FT UI Transportation Laboratory study in 2022 is used as the basis for the trips number and distribution outside the industrial area for further details on trips within the industrial estate.
Trip generation within the industrial estate is determined using trip rates by Institute of Transportation Engineers based on the land use area.The OD matrix is determined by an analogy approach using the Furness method, which will be input into the cube software during the route selection process.The results of the running process on the cube software will be used to analyze road performance due to the industrial area.Road performance analysis is carried out in 2024 and 2029.

Road Geometry Data
The geometry data of the roads are obtained from the FT UI Transportation Laboratory.The calculation of road's capacity and free flow speed refers to the Indonesian Highway Capacity Manual 1997.The calculation for urban roads have four adjustment factors consisting of carriageway width, directional split, side friction, and city size.For freeways, only the adjustment factor for carriageway width is used.

Traffic Analysis Zone
The internal zones used in the Batang KIT are 23 zones inside the industrial area (TAZ 1-23) and four zones outside the industrial area (TAZ 24-27).In addition, there are five external zones (TAZ 28-32) at each end of the road under study.

Trip Generation Calculation
The calculation of trip generation stage uses the following assumptions.

Trip Generation Rates.
The following is the trip rates for each type of land use.This value is used as a reference for calculating trip generation by using a trip rate based on known area units [4].

Building Floor Coefficient (KLB).
KLB is used as a reference for building floor area in each land use.By using six building references with industrial designation, i.e cellulose factory in Way Kanan [5], cellulose acetate and acetate ahydride mill in Karanganyar [6], PT Komatsu Remanufacturing Asia in North Jakarta [7], garment factory in Karanganyar [8], PT TACI in Cikarang [9], and PT Banten Energy International in Serang [10], the average of KLB 0.5 is obtained.

Industrial Area Occupancy Rate.
It is necessary to assume occupancy rate of industrial estates in year 2024 and 2029 by using eight industrial estate references, i.e Jababeka Cikarang Industrial Estate [11], MM2100 Bekasi Industrial Estate [12], Bukit Semarang Baru Industrial Estate, Terboyo Industrial Estate Semarang, Candi Industrial Estate Semarang [13], JIEP Pulogadung Industrial Estate [14], Medan Industrial Estate [15], Bekasi EJIP Industrial Estate [16].With reference to the year KIT Batang was founded in 2023, the assumption that the industrial area is occupied after 1 year of operation (2024) is 3% and after 6 years of operation (2029) is 18% by using interpolation.

Vehicle Composition.
Five types of vehicles were used in this study, i.e motorcycles (MCY), cars (CAR), light goods vehicles (LGV), medium goods vehicles (MGV), and heavy goods vehicles (HGV).By referring to references, the composition of vehicles produced for each type of land use is obtained.Land use itself is divided into seven types, including industrial, residential, institutional, medical, recreational, ports/terminals, and office.

Trip Generation based on Vehicle Type.
Referring to all the assumptions used, trip generation values are obtained in 2024 and 2029 for each type of vehicle across all land use types.
The trip distribution outside the industrial estate is sourced from the Semarang Raya traffic model from the FT UI Transportation Laboratory study in 2022.For the trip distribution from/to the industrial estate, it is only expected that TAZ 4, 5, 13, 15, 20, and 23 will generate trips to another TAZ.The trip distribution from/to industrial estate refers to the proportion of trip generation from each TAZ.
For the scenario with industrial park, the Tegineneng Industrial Estate was used as a reference for trip distribution from/to industrial estate by knowing the additional number of vehicle trips due to the presence of industrial estates.The number of trips after the presence of industrial estates in 2022 will change to 1,495 times the number of trips before development.As many as 33.13% of trips are the result of the existence of the industrial area [17].

Trip Assignment
The trip assignment stage is carried out using a generalized cost approach where this value is determined based on equation (2) where trips between zones of traffic analysis (TAZ) will look for trips with the cheapest value.The following table presents the weight values or operational cost coefficients for all types of vehicles (A).Furthermore, to obtain vehicle travel time, calculations are carried out using the capacity restrained (CR) approach as contained in equation (3).In this study, the coefficient values α and β are used by default from BPR, which are respectively 0.15 and 4 which are used for all types of road sections.

Validation
Model validation is carried out by comparing the vehicle travel time in the model to the actual vehicle travel time.By using mean average percentage error (MAPE), the accuracy of the models performed on the CUBE software is known as in Table 5.

Trip Distribution
To analyze the trip distribution, desire lines can be used which describe the relationship between the origin and destination of a movement.The thicker the desire lines, the greater the number of trips and vice versa.

Non-Goods Vehicle.
As an example, we use the results of desire lines on motorcycle mode (MCY).It can be seen that in 2024 the presence of industrial estates will not have such a significant impact.However, in 2029 there will be a significant increase if there is an industrial area.

Road Performance
In 2024, industrial estate has not had a significant impact on a road performance.But in 2029, it has had a significant impact on a road performance with a decrease in speed and an increase in degree of saturation or volume-capacity ratio (VCR).

Recapitulation.
The following is a recapitulation of road performance.It can be seen that the performance on toll road and primary arterial road 2/2 UD on average has entered poor and almost bad performance.
= Total cost of type k vehicles to traverse the segment in unit time (minutes)  = Vehicle travel time to cross the segment (minutes)  = Weight, coefficient of type k vehicle operating costs (minutes/km)  = Section length (km)   = Type k vehicle toll fees (Rp); on a closed system toll road the rate is converted first according to the length of the section (Rp/km x the length of the section in km = Rp)   = Type k vehicle time value (Rp/minute) = Vehicle travel time in traffic flow conditions of V (minutes)  0 = Vehicle travel time in free flow conditions (minutes)  = Traffic volume (pcu/hour)  = Road section capacity (pcu/hour) ,  = Coefficient

5. 1 . 2 .
Goods Vehicle.As an example, the results of desire lines are used in the heavy goods vehicle (HGV) mode.Similar results are also seen for non-goods vehicle modes, except that freight vehicles have a greater impact on trips from/to industrial areas than non-goods vehicles in both years.

5. 2 . 1 .
Travel Speed.The following is an illustration of the speed of the road where the green color indicates a faster speed and the red color indicates the opposite.

Figure 5 .
Figure 5. Travel Speed on Road Segment.5.2.2.Volume-Capacity Ratio (VCR).The following is an illustration of the VCR of the road where the green color indicates a faster speed and the red color indicates the opposite.

Table 1 .
Road Geometry Data a calculated per direction Source : FT UI Transportation Laboratory (2022)

Table 2 .
Trip Rates for Each Type of Land Use

Table 3 .
Trip Generation based on Vehicle Type

Table 4 .
Operational Cost Coefficient, A

Table 5 .
Model Validation of Travel Time

Table 6 .
Road Performance Recapitulation