WIM Data Analysts Manual

Section 2. WIM Basics

2.1.WIM SYSTEM VS. WIM SITE

A WIM system, as used in this manual, refers to the following components:

• One controller, its computer, and associated electronics.
  • CPU
  • Sensor Interface Cards
  • Communication Interface
  • Data Storage Medium
  • Software and/or Firmware
• All roadway sensors and their leads for all lanes for which traffic data is being processed by the controller (at least one lane must have weigh sensors).
• Controller support items such as lightning protection, uninterruptable power supply, etc.

A WIM site, as used in this manual, refers to a specific roadway location at which a WIM system has been (or will be) installed. Such a site includes:

• All WIM system components.
• The power and communication service facilities.
• All wiring, conduits, pull boxes, and cabinets necessary to make the WIM system functional.
• The pavement section in which the roadway components are installed and the pavement approach and departure from the in-road sensors.

Figure 2 through Figure 6 display WIM system components and Figure 7 displays a two-lane WIM site.

Figure 2. Photo. Staggered weigh sensors (bending plates) and detection loops (single threshold system).

Figure 3. Photo. In-line weigh sensors (single load cells), a trailing axle sensor, and detection loops (single threshold system).

Figure 4. Photo. In-line weigh sensors (quartz piezos) and detection loops (double threshold system).

For single threshold weighing, each axle's right and left wheel or dual wheel is weighed once by the right and left sensors. For staggered (leading and trailing) sensors, a vehicle's speed is calculated based upon the time it takes for each axle's wheels to hit the leading and trailing sensors. For in-line (side-by-side) weigh sensors, speed can be calculated by one of two methods:

1. An axle sensor (non-weighing) may be installed downstream of the weigh sensors and speed is calculated based upon the time it takes for each axle's wheels to hit the weigh sensors and trailing axle sensor.
2. If no axle sensor is installed, the speed is calculated based upon the time between a vehicle's triggering the leading and trailing loops. This is not as accurate as using sensor-to-sensor time measurements.

For double threshold weighing, each axle's right and left wheel or dual wheel is weighed twice by the right and left sensors. The system then reports a single left weight and a single right weight for each axle. A vehicle's speed is calculated based upon the time it takes for each axle's wheels to hit the leading and trailing weigh sensors.

Throughout this manual, a WIM system's right and left weigh sensors are discussed in regard to weight data output analyses, diagnostics, calibration, etc. Such right and left sensors will be treated as single sensors even though for double threshold systems there are actually two right sensors and two left sensors.

Figure 5. Photo. Controller cabinet, front view.

Figure 6. Photo. Controller cabinet, rear view, with in-road sensor inputs to controller and lightning protection.

Figure 7. Photo. Two-lane WIM site, with Portland cement concrete (PCC) pavement installed for approach and departure.

2.2. DEFINITIONS AND TERMS

2.2.1. WIM

As defined in ASTM E 1318:

"...the process of estimating a moving vehicle's gross weight and the portion of that weight that is carried by each wheel, axle, or axle group, or combination thereof, by measurement and analysis of dynamic vehicle tire forces."

2.2.2. Site Assessment

Refers to onsite activities preceding either an onsite evaluation or calibration to verify and document that:

• The WIM system is operational.
• The sensors have no visible problems.
• The pavement condition shows no apparent deterioration.

2.2.3. Weight

Throughout this manual, the term "weight" will be used, even though it may be technically appropriate to use another term such as "load" or "force". Weights will typically be expressed in kips (k), where 1 k equals 1000 pounds (lbs).

2.2.4. GVW

Throughout this manual, the term Gross Vehicle Weight (GVW) is used to refer to the sum of all of a vehicle's wheel weights or axle weights. GVWs will typically be expressed in units of kips.

2.2.5. Calibration and Validation (Using Test Trucks Onsite)

Both calibration and validation utilize a process by which the known static axle and/or wheel weights and known axle spacings of one or more test vehicle(s) are compared with the corresponding estimates from a WIM system's reported dynamic wheel weights and axle spacings for such test vehicle(s).

The purpose of calibration is to determine and implement the WIM system settings which will result in the system's generating the best possible estimate of static axle and/or wheel weights, axle spacing distances, and vehicle speeds for the most typical truck configurations in the traffic stream over the range of speeds typical of such truck configurations.

The purpose of validation is to check a system's accuracy for conformance to an agency's specified requirements. Once a system has been initially calibrated, test trucks should be run on a routine basis (or as otherwise deemed necessary) to check the system's calibration. This is also typically referred to as a validation. If such validation indicates that the system meets accuracy requirements but that accuracy could be improved, then the calibration factors may be adjusted and additional test truck runs made to confirm that the factor adjustments produced the desired effect.

2.2.5.1. Calibration Factor

Refers to a user-defined value that is used by a WIM system to convert raw sensor readings into weights.

2.2.5.2. Calibration Factor Speed Point

Also referred to as Speed Bin, refers to a user-defined speed for which a calibration factor can be entered for a weigh sensor. Certain WIM systems provide for three or more calibration factor speed points, which allow the user to determine appropriate calibration factors over a range of vehicle speeds which will best compensate for the effects of speed. For speeds between the speed points, the system uses linear interpolation to apply calibration factors to the sensor's readings.

2.2.5.3. WIM Error

Is the difference between a test truck's static weights and the corresponding WIM reported weights as derived from the test truck's dynamic readings.

2.2.6. LTPP

Refers to the Long-Term Pavement Performance (LTPP) program (http://www.fhwa.dot.gov/pavement/ltpp/), a 20-year study of in-service pavements across North America. Its goal is to extend the life of highway pavements through various designs of new and rehabilitated pavement structures. The LTPP program evaluates different pavement materials under different traffic loading, environmental and subgrade soil conditions. Different pavement maintenance practices are evaluated as well. The LTPP program was established in 1987 under the Strategic Highway Research Program (SHRP), and is now managed by the Federal Highway Administration (FHWA).

2.2.7. LTPP SPS TPF Study

Refers to the LTPP Specific Pavement Study (SPS) Traffic Pooled Fund Study, TPF-5(004). Phase I of this study consists of assessing, evaluating, and calibrating WIM systems used to collect traffic data at the SPS sites across the country. Phase II consists of the installation and maintenance of new WIM equipment as necessary to ensure high-quality data collection.

This Manual cites a number of LTPP and LTPP SPS TPF Study documents such as the LTPP Field Operations WIM Guide, the WIM Model Specifications (See Appendix A) and the LTPP Classification Scheme. These documents are extensive and contain valuable information for WIM equipment and site maintenance.

2.2.8. Traffic ETG

Refers to the Transportation Research Board (TRB) Expert Task Group (ETG) on LTPP Traffic Data Collection and Analysis. The Traffic ETG is composed of individuals with significant experience and involvement in the collection and/or analysis of truck traffic data. The Traffic ETG provides advice and guidance to the staff of the LTPP program regarding the reliability and precision of traffic data, among other things.

2.3. VEHICLE CLASSIFICATION VERSUS VEHICLE TYPE

As used in this manual, vehicle classification refers to the identification of vehicles according to FHWA's 13 Class Scheme as described in the Traffic Monitoring Guide (http://www.fhwa.dot.gov/ohim/tmguide/). However, individual classes within this scheme include vehicles with different axle configurations and operating characteristics that need to be uniquely identified by a WIM system's classification algorithm. Additionally, the ability to perform analyses on vehicles with similar axle configurations and operating characteristics, regardless of FHWA classification, can be of great benefit in performing data analyses. Vehicle type is used in this manual to refer to vehicles with similar axle configurations and operating characteristics. A few examples of vehicle types follow.

Class 7 includes all trucks on a single-frame with four or more axles. For trucks with "variable load suspensions" or "lift axles" (as shown in Figure 8), only the axles in contact with the pavement are counted to determine classification.

Figure 8. Photo. Class 7, single-unit truck with four of its five axles in contact with pavement.

Class 8 includes several common three- and four-axle single-trailer configurations. Figure 9 displays a two-axle tractor with a single axle semi-trailer and Figure 10 displays a three-axle tractor with a single axle semi-trailer. For this method of defining a truck combination type, the first value is the number of axles on the power unit (tractor or straight truck), the "S" signifies a semi-trailer, and the following value is the number of axles on the trailer.

Figure 9. Photo. Class 8, Type 2S1.

Figure 10. Photo. Class 8, Type 3S1.

Class 9 includes five-axle single-trailer trucks. Figure 11 displays the three-axle tractor and two-axle semi-trailer, which is by far the most predominant Class 9 type. Figure 12 displays the same type but with a "spread" tandem on the trailer. If this axle spread exceeds eight feet it is not a true tandem axle and is considered to be two individual axles. Figure 13 displays a three-axle straight truck pulling a two-axle full trailer. As such, there is no "S" preceding the value defining the trailer's number of axles.

Figure 11. Photo. Class 9, Type 3S2.

Figure 12. Photo. Class 9, Type 3S2 with "spread" rear tandem.

Figure 13. Photo. Class 9, Type 32.

Class 10 includes six-axle single trailer trucks. Figure 14 displays the most common configuration, the Type 3S3 which has a semi-trailer with a tridem axle.

Figure 14. Photo. Class 10, Type 3S3.

Class 11 includes five-axle multi-trailer trucks. Figure 15 displays the most common configuration, the Type 2S12. The first value defines the number of axles on the power unit, the "S1" defines the single axle semi-trailer, and the last value defines the second trailer as a two-axle full trailer.

Figure 15. Photo. Class 11, Type 2S12.

Class 12 includes six-axle multi-trailer trucks. Figure 16 displays the most common configuration, the Type 3S12.

Figure 16. Photo. Class 12, Type 3S12.

Class 13 includes multi-trailer trucks with seven or more axles for which there are a large number of possible axle configurations. Although there are exceptions, most agencies do not find it necessary to uniquely define these by type since they account for a very low percentage of the truck traffic stream. Some states allow very heavy mining or timber hauling "trains" with many axles, which they may find beneficial to capture by type for analyses. Some states allow Longer Combination Vehicles (LCVs), which do have consistent configurations as displayed in Figure 17.

Figure 17. Picture. Class 13, Longer Combination Vehicles (LCVs).

Table 1 displays the basic class scheme that was recommended for use in the LTPP study by the Traffic ETG. The intent of this scheme is to include the most common vehicle types found nationwide and to be supplemented with additional vehicle types unique to certain regions. It is important to note that the axle spacing and weight parameters of any desired scheme must be set up as an algorithm specifically formatted for use by a particular WIM system.

Note that although the "LTPP Classification Scheme for SPS WIM Sites" displayed in Table 1 is currently in use, it is considered a work in progress subject to revisions and enhancements. For more information regarding this document contact ltppinfo@dot.gov.

Table 1. Example Class Scheme.

2.4. TASKS OF THE OFFICE DATA ANALYST

The primary tasks of the Office Data Analyst are as follows:

• Identify any missing or invalid data.
• For missing or invalid data, attempt to determine:
  • Cause
    • Ongoing
    • Intermittent
    • Isolated, one time event(s)
  • Extent
    • What part of the day's data should be flagged as invalid?
      • All data by time frame
      • One lane only
      • Classification counts valid, but not loading data
  • Course of action to correct ongoing or intermittent problem
    • Fix remotely from office
    • Call for field visit
• Monitor each WIM system for maintenance of calibration.

2.4.1. Causes of Missing or Invalid Data

Missing data for all lanes would most likely be due to a power outage at the system's controller or the controller being otherwise shut down. The primary causes of invalid data are as follows:

• System component malfunction
  • Failure vs. intermittent
• Improper system settings
  • System component operational variables, such as
    • Loop timeout
    • Weigh sensor thresholds
  • System processing variables, such as
    • Classification algorithm
    • Calibration factors
• Site conditions
  • Rough pavement
  • Roadway geometry
• Traffic conditions
  • Congestion
  • Lane closure or alignment shift
  • Changing lanes when crossing sensors
• Weather
  • Effects on system components, such as
    • Moisture intrusion
    • Bending plate scale pit filled with ice
  • Effects on traffic, such as
    • Crosswind
    • Traffic not within lane/shoulder striping due to heavy snow on pavement
  • Weather web site address
    • http://www.nws.noaa.gov/

In regard to system settings, classification algorithms and calibration factor values have no bearing, within reason, on a system's proper operation. Determination of classification algorithms and calibration factor values are more of a "fine tuning" process to generate the most accurate data possible as opposed to generating valid versus invalid data (unless the values are obviously erroneous). However, settings related to component operation, such as loop timeout and weigh sensor thresholds, will determine whether vehicles and vehicle combinations are properly detected and their wheels are properly counted and weighed.

2.4.2. Factors Affecting WIM Data Quality Which Can Be Controlled

• Site conditions
  • Roadway geometry
  • Pavement stability and smoothness
• Equipment quality and performance capabilities
• Equipment installation and routine maintenance
• Pavement maintenance

2.4.3. Factors Affecting WIM Data Quality Which Can Be Somewhat Controlled (By Site Selection)

• Traffic characteristics
  • Changing lanes through site
  • Wheels on or partially on shoulder
  • Speed changes or stop and go
    • Merging vehicles
    • Congested traffic

Power Point presentations that describe site conditions and traffic operating characteristics that should be considered in determining a suitable location for a WIM system installation are available online at www.QualityWIM.com

2.4.4. Factors Affecting WIM Data Quality Which Cannot Be Controlled

• Out of round tires
• Dynamically unbalanced wheels
• Tire inflation pressure
• Vehicle suspension
• Wind
• Vehicle aerodynamic features
• Type of load, particularly liquid, and/or method of loading
• Acceleration/deceleration
• Undesirable traffic conditions due to weather or work on roadway

2.5. WHAT QUALITY OF DATA SHOULD BE EXPECTED?

No WIM system can produce perfect data, even with high quality equipment and ideal site conditions. It is expected that any data file is going to contain some invalid data. The analyst must consider the characteristics of the WIM site, the characteristics and features of the WIM system, and the traffic characteristics in determining if the system is producing the best data possible. Regardless of what the minimum data quality requirements are, any WIM system should be monitored and maintained as to produce the best possible data given the system's potential. The key is to keep bad data to a minimum, giving consideration to each WIM system's potential, and to quickly recognize, identify, isolate, and correct the cause of erroneous data.

Many data problems can be corrected from the office. Even if a problem does require a service call, the service technician's time onsite can be greatly reduced if the analyst has narrowed down the potential causes of the problem. Neither the purging of entire daily data files nor major WIM system corrective actions are necessary if only a scattering of bad data is found when performing routine data quality control checks. If the amount of bad data starts to increase and goes from random to chronic, the analyst needs to take corrective action, unless the problem can be tied to an atypical site condition (e.g. traffic or roadwork).

Contact

Mike Moravec
Office of Transportation Performance Management
202-366-3982
E-mail Mike