2. ESTUDIO DE MERCADO
2.2 DESCRIPCIÓN DEL SERVICIO
Truck
(Duals & TTST) Location
Route
No. Years of Data NCDOT Station Number
Yes ADT & AADT
ADTT &
AADTT Given Yes 1992-1996
VTRIS Station ID Yes AADT ADTT Given Yes 1997-2004
Currently, most NCDOT data are partial data sets collected from short-term (48-hour) count stations or turning movement count stations. Most of the counts are not factored or annualized and therefore not included for the development of the model. The ADT from these stations cannot be approximated equal to AADT. Thus, the data must be factored to an annualized value of AADT to convert it to the correct format to be used in trend analysis. These data sets are available from ATR counts and only data from the case study sections on I-95, US64, NC421, and NC279 are annualized (Appendix B). The results of the case studies appear in Chapter 5 and the appendices. All the other ATR data are un-factored, and they are not considered for further analysis.
VTRIS uses functional classes to distinguish different highway classifications (Figure 3-3). The VTRIS database consists of annually classified counts along rural principal arterials (interstates)
and rural principal arterial other (US routes). Separate data reflect urban principal and minor
arterials, though they are fewer in number. The types of routes classified by VTRIS are shown in Table 3-3. Arterials provide the highest level of mobility at the highest speed for long uninterrupted travel. The interstate highway system is an arterial network. Arterials generally have higher design standards than other roads and many principal arterials have multiple lanes with some degree of access control. Arterials are broken into principal and minor routes. The rural arterial network provides interstate and inter-county service so that all developed areas are within a reasonable distance of an arterial highway. The urban principal arterial system serves major metropolitan centers, corridors with the highest traffic volume, and those with the longest trip lengths. It carries most trips entering and leaving urban areas, and it provides continuity for all rural arterials that intercept urban boundaries.
Since NCDOT conducts classified counts once in about three years, because some recent years of data are unavailable for the case studies, and because most NCDOT data is unfactored, the research focus is the VTRIS database. The VTRIS dataset is the basis of the methodology to forecast truck traffic in North Carolina. The robustness of the VTRIS dataset is exploited in the development of the forecasting methodology. Data at each station for each facility type is manually extracted from VTRIS online database and entered into a spreadsheet for further analysis. (The manual extraction was tedious but seemed to be the most efficient for this research.
Figure 3-3. Highway Functional Classification
Source: A Guide to Highway Functional Classification, Montana DOT
Table 3-3. VTRIS Highway Classification
VTRIS Functional Classes Highway Classification
Rural Principal Arterial Interstate Interstate Routes only Rural Principal Arterial Other US Routes, some NC Routes and SR Routes
three different categories: Cars (classes 1-3), Duals (classes 4-7), and TTST (classes 8-13). Then the calculated average growth factor and other calculated metrics show the trend of traffic at a particular location. Total traffic growth rates vary from near zero to several percent per year depending on the urban or rural location of the highway. The VTRIS database provides classified counts for the 60 WIM stations currently in North Carolina. (More are planned.) These counts include vehicles of all classes averaged according to weights at the WIM Stations
Highway Pavement Monitoring Systems (HPMS) Data
The HPMS is a USDOT highway information system that includes data on the extent, condition, performance, use, and operating characteristics of the nation's highways. It contains descriptive information and ADTs on many public roads as a mix of universe and sample data for arterial and collector functional systems. Limited information on travel and paved miles is included in summary form for the lowest functional systems.
The major purpose of the HPMS is to support a data driven decision process within FHWA, the DOT, and the Congress. The HPMS provides data in the analysis of highway system condition, performance, and investment needs that make up the biennial Condition and Performance Reports to Congress. These Reports are used by the Congress in establishing both authorization and appropriation legislation, activities that ultimately determine the scope and size of the Federal-aid Highway Program, and determine the level of Federal highway taxation.
The most recent online HPMS data archive consists of total AADT data and does not include classified vehicle counts or any truck data. The online archive has not been updated with latest traffic counts restricting it to a period of 1993 to 2001. According to NCDOT Traffic Survey Unit, NCDOT maintains 60 weigh in motion sites in 2004, but the HPMS does not have an updated database. It includes only 15 or 16 WIM stations on the online map. A compressed file is available for download and the total AADT are identified by link numbers and not by station numbers. The link numbers do not match NCDOT station numbers or VTRIS dataset station numbers. Thus, comparing and complementing NCDOT and VTRIS data with HPMS data is infeasible. Consequently HPMS data were not used in this research.
Growth Factor Estimation
Usually, NCDOT uses linear and exponential models in their traffic forecasts (Traffic Forecasting Utility). General statistical packages can be used also (Appendix C). NCDOT determines the traffic growth factor using the first and last traffic count (AADT) in a sequence. Engineers apply judgment to choose the best growth factor for the project location being studied. On the other hand, the research methodology uses an Average Growth Factor for the range of the historical data. Instead of engineering judgment the research methodology proposes that the calculated average growth factor be compared to an expected statistical range and be adjusted only if it falls outside that range. Although the NCDOT method and the research method use two different formulas, the growth factors show similar traffic forecasts for low volume roads, but differences occur for high volume roads.
This research uses the Average Growth Factor (AGF) method because it applies all traffic counts (not just first and last counts), and it is suitable for the VTRIS database. Besides calculating AGF for the cases in this research, the data recorded at WIM stations were tested to determine statistically significant ranges of growth factors by highway type and whether the growth of truck traffic is greater than general traffic (Appendices D and E). The analysis also tested linear regression models to capture causal influences such as NC region, county population and nearby warehouses on truck traffic (Appendix F). While the analysis shows effects, it is difficult to develop a linear regression model of growth factor at the project level for the trucks versus “exogenous” causal data because of the incompatibility of the project location and the data geography. A project location is a specific highway segment, but causal data covers regions. The VTRIS data given as AADT permits developing alternative traffic growth rates and models. The Annual Growth Factor (GF) is:
In cases where some years of traffic data are missing or the traffic data varies, the annual growth
rate may be determined by the following equation which is the “first and last year” calculation used by NCDOT. It assumes that each year in the period experiences a constant growth rate.
GF = [(Tfy – Tby)/Tby] / (fy – by)
In order to include the effect of all traffic data within a time period the Average Growth Factor (AGF) may be calculated as shown below. The AGF is the growth factor proposed by this research.
AGF = Σ GF / N where:
T = traffic = AADT GF = annual growth factor
AGF = average annual growth factor t = year, t-1 = previous year
N = the number of growth factors
fy = future year of a historic or future time period by = base year of a historic or future time period
If the traffic forecast follows a linear model with a “first and last year” GF, as NCDOT often
assumes for Interstate and other major highways, the future traffic Tfy may be determined by:
Tfy = Tby + GF (fy – by) = AADTfy
This research recommends replacing the GF by the AGF to account for years of variable traffic growth.
Tfy = Tby + AGF (fy – by) = AADTfy
If the traffic forecast follows an exponential model, which NCDOT often assumes for minor roads that may experience significant traffic growth, the future traffic Tfy may be determined as
follows:
Tfy = Tby (1 + AGF)(fy – by) = AADTfy
Similar forecasting equations represent future values of DUAL and TTST trucks for each VTRIS station or other station for which factored and annualized truck traffic data are available. Thus,
and for an exponential model:
DUALfy = DUALby (1 + AGF)(fy – by)
TTSTfy = TTSTby (1 + AGF)(fy – by)
The previous truck traffic forecasting equations emphasize that truck traffic should be forecast separately from AADT before the percentages by truck category are determined.
%DUAL = (DUAL/AADT) x 100 %TTST = (TTST/AADT) x 100
To support the AGF-based truck and traffic forecasts there are 60 NC VTRIS stations (WIM stations) defined by USDOT and NCDOT. USDOT and NCDOT have scheduled more VTRIS stations for construction. They are located on different VTRIS route types and provide 365-day continuous, classified AADT and truck counts from 1997 to 2004 (the years for this research.) Thus, each VTRIS (WIM) station can provide reliable growth factors for DUAL and TTST trucks and general traffic. For the 60 stations, only 51 have data recorded in the VTRIS online database. These 51 stations consist of 19 stations on Interstates and 32 stations on US and NC arterials. Some of these stations have data only for a single year and therefore it is impossible to develop growth factors from those stations. Such stations are not included in the subsequent analysis and statistical tests. All stations, which have at least two years of data have been included for developing the models. These include 18 stations along Interstates and 27 stations along US and NC arterials.
Some stations have missing data between 1997 and 2004. In order to provide uniformity in the growth rate, the research assumes that the growth rate remains constant between the missing years. Average growth factors (AGF) for each VTRIS (WIM) station are calculated for ADT (total traffic), Cars, total trucks, and DUAL and TTST trucks.
Another important product obtained from the dataset is the percentage of Duals and TTSTs in total traffic. The composition of traffic into three categories plays an important role in the development of the model. It leads to developing default values of percent Duals and percent TTSTs in cases where a classified count at the project location is unavailable. Table 3-4, Table 3-5
and Figure 3-4 show a summary of results for Interstate routes at WIM stations defined by Strategic Highway Research Program (SHRP) station numbers. Similarly Tables 3-6 and 3-7 and Figure 3-5 show summary results for Arterials Other highways.
Descriptive Statistics of the Data
Accurately estimating current and future Dual & TTST volumes is critical because of the effects of trucks on highway congestion and pavement distress. Thus, the preferred technique for forecasting traffic volumes should include separate forecasts for AADT, Duals, and TTSTs (Appendix C) rather than just forecasting AADT and assuming a percent of Duals and TTSTs applied to the forecasts. However, the accurate estimation of historic annual percentage of Duals
Figure 3-4. Average Percentage Duals and TTSTs at WIM (SHRP) Stations on Interstate Routes, 1997 - 2004