Attachment J.5 - 9. LTPP Field WIM Guide.pdf

LTPP Field Operations Guide for SPS WIM Sites Version 1.0 July 6, 2012

Prepared by:

FHWA-LTPP Technical Support Services Contractor AMEC Earth and Environmental 12000 Indian Creek Court, Suite F Beltsville, Maryland 20705-1242

Prepared for:

Office of Infrastructure R&D LTPP Team, HRDI-30 Federal Highway Administration 6300 Georgetown Pike McLean, Virginia 22101-2296

US Department of Transportation LTPP Team Federal Highway Administration Office of Infrastructure R&D

LTPP Field Operations Guide for SPS WIM Sites July 6, 2012 Version 1.0 Page i

Table of Contents

1 INTRODUCTION

1.1 OBJECTIVE

1.2 SAFETY

1.3 ORGANIZATION OF GUIDE

2 SITE ASSESSMENT

2.1 PAVEMENT CONDITIONS

2.1.1 Profile Data

2.1.2 Distress Identification

2.1.3 Vehicle-Pavement Interaction

2.2 TRAFFIC

2.2.1 Classification Data

2.2.2 Speed Data

2.2.3 Truck Route

2.3 EQUIPMENT

2.4 REPORTING OF RESULTS

3 SITE VALIDATION - WEIGHT

3.1 BACKGROUND

3.2 PERFORMANCE TESTING PROCEDURE

3.2.1 Speed Verification and Data Collection

3.2.2 Pavement Temperature Data Collection

3.2.3 Running Trucks

3.2.4 Loading Data Collection

3.2.5 Process by Which Required Number of Test Runs Was Determined

3.3 WIM DATA ANALYSIS

3.3.1 Sample Size Influence on 95% Confidence Limits

3.3.2 Specialized Tests – Sensitivity to Speed

3.3.3 Specialized Tests – Sensitivity to Temperature

3.4 TRUCKS

3.4.1 Truck Selection

3.4.2 Measuring Trucks

3.5 REPORTING OF RESULTS

3.5.1 Failure conditions

3.5.2 Consequences of Other Than Research Quality Data

3.6 ADMINISTRATION AND SUPPORT

4 SITE VALIDATION - CLASSIFICATION

4.1 RESEARCH QUALITY CLASSIFICATION DATA

4.2 SPEED VERIFICATION AND DATA COLLECTION

4.3 REPORTING OF RESULTS

4.3.1 Consequences of Other Than Research Quality Data

4.4 ADMINISTRATION AND SUPPORT

5 PAVEMENT SMOOTHNESS

5.1 HIGH-SPEED PROFILING

5.1.1 Reporting Results

5.2 SMOOTHNESS DETERMINATION THROUGH USE OF A STRAIGHTEDGE

5.2.1 Smoothness Criteria

5.2.2 Longitudinal Smoothness Determination

5.2.3 Transverse Smoothness Determination

Version 1.0 Page ii

5.2.4 Equipment for Straight Edge Profiling

6 CONSTRUCTION

6.1 SITE LOCATION (SELECTION)

6.2 SMOOTHNESS ACCEPTANCE

6.3 EQUIPMENT ACCEPTANCE

7 WIM SYSTEM INSTALLATION AND CALIBRATION AUDIT

7.1 AUDIT SUPPORTING MATERIEL

7.1.1 Checklists for Audits

7.1.2 LTPP Traffic Sheets for Audits

7.1.3 Equipment List

7.1.4 Document List

7.2 STANDARD WIM QA AUDIT PROCEDURES

7.2.1 Personnel

7.2.2 Scheduling

7.2.3 Digital Photographs

7.2.4 Pre-visit Preparation

7.2.5 Non-Compliance Report (NCR)

7.3 WIM SYSTEM SUPPORT EQUIPMENT INSTALLATION

7.3.1 Protection of Utilities

7.3.2 Support Equipment

7.3.3 Conduit

7.3.4 Cabinet

7.3.5 Cabinet Foundation

7.3.6 Pull boxes

7.4 WIM SYSTEM EQUIPMENT INSTALLATION

7.4.1 Layout

7.4.2 Saw-cutting

7.4.3 Excavation

7.4.4 Pre-Installation Clean up

7.4.5 Installation

7.4.6 Cabling/Terminations

7.4.7 Documentation

7.4.8 Final Clean up

7.5 WIM SYSTEM EQUIPMENT CALIBRATION

7.5.1 Test Vehicles

7.5.2 Pre-Calibration Test Runs

7.5.3 System Calibration/Verification

7.5.4 Post-Calibration Test Runs

7.5.5 Classification Verification

7.6 AUDIT REPORTING

8 REFERENCES

9 SUPPORTING SOFTWARE

9.1 WIM 2 TRUCK.XLS

9.2 CLASSMACRO.XLS

9.3 WIM SMOOTHNESS INDEX SOFTWARE

9.3.1 Processing of Raw Data Files

9.3.2 Analysis of Data

10 FORMS

10.1 AIMS DATA

10.2 IDENTIFYING LTPP WIM SITES

10.3 COMMON DATA SHEET ITEMS

10.4 TRAFFIC SHEET 17 -WIM SITE INVENTORY

10.5 TRAFFIC SHEET 18 – WIM SITE COORDINATION

Version 1.0 Page iii

10.6 TRAFFIC SHEET 19 SERIES – TEST TRUCK INFORMATION

10.6.1 Traffic Sheet 19A

10.6.2 Traffic Sheet 19B-N

10.6.3 Traffic Sheets 19C-N

10.7 TRAFFIC SHEET 20 – SPEED AND CLASSIFICATION CHECKS

10.8 TRAFFIC SHEET 21 SERIES – WIM TRUCK RECORD DATA

10.9 TRAFFIC SHEET 22 SERIES - SITE EQUIPMENT ASSESSMENT

10.9.1 Traffic Sheet 22A

10.9.2 Traffic Sheet 22B

10.9.3 Traffic Sheets 22C, 22D, 22E, 22F

10.10 TRAFFIC SHEET 23 – WIM SYSTEM TROUBLESHOOTING OUTLINE

10.11 TRAFFIC SHEET 24 SERIES – SITE PHOTO LOG

10.11.1 Traffic Sheet 24A

10.11.2 Traffic Sheet 24B

10.11.3 Traffic Sheet 24C

10.12 TRAFFIC SHEET 25 SERIES - SITE AUDIT PHOTOGRAPHS

10.12.1 Traffic Sheet 25A

10.12.2 Traffic Sheet 25B

10.12.3 Traffic Sheet 25C

10.12.4 Traffic Sheet 25D

10.12.5 Traffic Sheet 25E

10.12.6 Traffic Sheet 25F

10.12.7 Traffic Sheet 25G

10.12.8 Traffic Sheet 25H

10.13 TRAFFIC SHEET 26 SERIES – PHOTOMAPS

10.13.1 Traffic Sheet 26A

10.13.2 Traffic Sheet 26B

10.13.3 Traffic Sheet 26C

10.14 TRAFFIC SHEETS 27 SERIES - SITE AUDIT MEASUREMENTS

10.14.1 Traffic Sheet 27A

10.14.2 Traffic Sheet 27B

10.14.3 Traffic Sheet 27C

10.15 TRAFFIC SHEET 28 – NON COMPLIANCE REPORT

10.16 TRAFFIC SHEET 29 – AUDIT SUMMARY

10.17 TRAFFIC SHEET 30 – INFRARED PAVEMENT TEMPERATURE SENSOR

10.18 TRAFFIC SHEET A – STRAIGHTEDGE MEASUREMENTS – LONGITUDINAL

10.19 TRAFFIC SHEET B – STRAIGHTEDGE MEASUREMENTS – TRANSVERSE

10.20 WIM QA CHECKLIST SERIES

10.20.1 Checklist A-01

10.20.2 Checklist A-02

10.20.3 Checklist A-03

10.20.4 Checklist A-04

11 MISCELLANEOUS FORMS

Version 1.0 Page iv

LIST OF TABLES

Table 3-1 – WIM System Calibration Tolerances Table 3-2 – WIM System Calibration Tolerances – Large Sample Table 3-3 - WIM System Calibration Tolerances – Small Sample Table 3-4 Values of Students’ t Statistic for Different Sample Sizes Table 3-5 Example - All Vehicle Results vs. LTPP SPS WIM Tolerances Table 3-6 Example - Computations for LTPP SPS WIM Tolerances by Speed Range Table 3-7 Computations for LTPP SPS WIM Tolerances by Temperature Range Table 5-1 Recommended WIM Smoothness Index Thresholds Table 5-2 Long Range Index (LRI), Short Range Index (SRI) and Peak Values Table 10-1 Alternatives for Varying Numbers of Axles (Sheet 19C) Table 11-1 General Equipment for WIM Site Validation Table 11-2 Equipment Checklist for Straightedge Smoothness Testing of Newly Constructed Pavements Table 11-3 State FIPS Codes

LIST OF FIGURES

Figure 5-1 Positioning of Straightedge Figure 7-1 - NEMA Type M Cabinet Figure 7-2 – Typical Staggered Bending Plate – Dual Loop Figure 7-3 – Typical Staggered Bending Plate – Single Loop Figure 7-4 - Typical Load Cell Layout Figure 7-5 - Typical Full-Lane Quartz Sensor Layout Figure 7-6 - Typical Half-Lane Quartz Sensor Layout Figure 7-7 - Square Line Setup Figure 9-1 WIM 2 Truck on Opening Figure 9-2 CLASSMACRO spreadsheet Figure 9-3 Windows RP90L software to process raw files Figure 9-4 WIM Smoothness Index Software Figure 10-1 - WIM Sensor Measurement Figure 10-2 – Loop Distance Measurement Figure 10-3 - WIM Distance Measurement Figure 10-4 - Loop and WIM Sensor Distance Comparison Figure 10-5 - Loop Measurement Locations Figure 10-6 – Suggested Depth Check Locations

Version 1.0 Page 1

1 INTRODUCTION

The Long-Term Pavement Performance (LTPP) program has been an important source of pavement data collection for the last 20 years. This effort has been led by the Federal Highway Administration (FHWA), in partnership with state departments of transportation, American Association of State Highway Transportation Officials (AASHTO), National Academy of Science, industry, and academia.

Periodic updates on products delivered through LTPP are disseminated to FHWA partners, who are the users of the products. Users provide comments to the FHWA team on their successful application of products and on issues encountered during their implementation. Mid-way through the program, FHWA assembled an LTPP program assessment team to address concerns related to the LTPP program and its database. This field operations document provides guidance to correct deficiencies identified by the FHWA LTPP program assessment team related to the quality and quantity of traffic data collected by the LTPP program for the Specific Pavement Study (SPS) test sites.

With the support of the Transportation Research Board (TRB) Expert Task Group on Traffic Data Collection and Analysis (or Traffic ETG), FHWA revised the traffic data collection procedures for SPS- 1, -2, -5, -6 and -8 experiments. The effort consisted of two principal elements: shifting the data collection from highway agencies to a national, centralized effort and standardizing data collection equipment and procedures. The effort has been successful in that state agencies have agreed to the establishment of a national pooled fund effort to support data collection and processing as well as equipment purchases. In addition to this field operations guide, there are other documents that address the pavement smoothness, equipment calibration checks, and equipment model specifications for a weigh-in-motion (WIM) system.

1.1 OBJECTIVE

This document contains the guidelines for traffic data collection at SPS sites. It covers all aspects of the process in the field. Some or all of the process may remain the responsibility of the highway agency depending on its degree of participation in the national pooled fund study for SPS Traffic Data Collection. This document is intended to be circulated and encouraged as standard practice at all SPS WIM sites whether or not the agency is a participant in the SPS Traffic Pooled Fund Study. The majority of the reporting requirements included in this document are specific to sites evaluated under FHWA direction. The Sheet 16 reporting requirement however is applicable to all sites with LTPP lane data submitted for inclusion in the LTPP database. Within this document there are three groups identified as participants in the process: FHWA, highway agencies and data collectors. Data collectors may be responsible for equipment testing, site assessment, site validation, downloading of data for LTPP use, site profiling or smoothness evaluation. The processing of data collected at these sites is documented separately in SPS-1, -2, -5 and -6 Traffic Data Processing for SPS-1, -2, -5 and -6 sites.

Data processing for SPS-8 sites is discussed in LTPP Monitored Traffic Data Processing.

Although this document uses the term WIM equipment, many of the processes and forms are also applicable to Automated Vehicle Classification (AVC) equipment. Sites instrumented solely with AVC equipment (generally SPS-8s) are also evaluated using this guide using all elements except for those that refer to equipment or processes that assess the quality of loading data.

Throughout this document the terms SPS projects, SPS sites and SPS locations refer to the SPS-1, -2, -5 and -6 experiment projects and those SPS-8 experiment projects that have been targeted for intensive

Version 1.0 Page 2 traffic monitoring activities. Intensive traffic monitoring activities implies installation of permanent traffic monitoring equipment to collect data on a continuous basis.

Although this document was developed as part of the effort to improve data quality and quantity at SPS- 1, -2, -5, -6 and -8 locations, the material in here is applicable to any LTPP traffic data collection location.

1.2 SAFETY

Traffic data collection involves work in and adjacent to the traveled way. Data collectors and observers are to follow all FHWA, state and individual company policies for work on highway rights of way. Data collectors are responsible for keeping a list of policies applicable to each individual state. Where site-specific conditions involve special actions these should be included within that list.

1.3 ORGANIZATION OF GUIDE

This guide is divided into six major operational sections. These are Site Assessment, Site Validation – Weight, Site Validation – Classification, Pavement Smoothness, WIM Equipment Installation & Calibration Auditing, and Data for Use in LTPP Activities.

Site Assessment is the inventory or fact gathering visit to collect all site information on pavement, equipment, traffic, inputs to validations and contact names. It is performed on the first visit to any test site with an operational WIM or AVC. In the absence of installed equipment at a site, it is part of the site selection process.

Site Validation for Weight is the verification of a site’s precision and bias in estimating truck weights.

It is performed by running test trucks of known weights. It also includes an annual pavement smoothness evaluation and validation for classification. It is recommended that the weight validation be conducted once a year for bending plates and load cells and twice a year for quartz piezo sites.

Site Validation for Classification is a check on the accuracy of the equipment with respect to the installed classification algorithm. Speed and spacing checks are part of the process. It is recommended that this be done at least once a year.

Pavement Smoothness is an evaluation of the pavement’s influence on WIM equipment precision. For in-service sites the evaluation is done using data from a high-speed profiler. At newly constructed sites, where lane closure is possible, a straightedge methodology is also available.

Construction outlines site selection and installation activities.

WIM System Installation & Calibration Auditing is the inspection and evaluation of SPS WIM installation activities in order to promote greater uniformity and quality in the installation, calibration and operation of the WIM systems installed at LTPP SPS test sections.

Additional sections are:

Supporting Software which discusses LTPP software tools associated with field operations.

Forms which includes the instructions for completing all of the forms and copies of each of the forms that is specific to the activities in this document. Some of the forms used are found in Guide to LTPP Traffic Data Collection and Processing (Ref. 3).

Version 1.0 Page 3

2 SITE ASSESSMENT

A site assessment is the first equipment evaluation visit to an SPS traffic data collection site. All SPS-1, -2, -5 and -6 sites require such a visit to provide a snapshot in time of site conditions whether the site is monitored by the agency or LTPP. Any SPS-8 identified for continuous monitoring will also require a site assessment. The outcome of an assessment is a report on the status of the equipment and a judgment on whether remediation or validation is the next step in the process of obtaining quality data.

As part of the site assessment, the field personnel should collect the site and equipment information and conduct a visual inspection of all WIM site support components. Field personnel should perform static electrical and electronic tests of the equipment. Pavement condition, in-road sensor condition, and any impact on truck motion should be recorded. Information on truck speeds, classification and route options for test trucks should be noted during each site assessment to allow for more efficient planning of subsequent validation activities.

The record of the assessment is documented in a number of forms. They are referenced in the sections that follow along with the data requirements for an assessment.

The first form to be completed is a LTPP Traffic Sheet 18, WIM Site Coordination. This sheet has agency contact information and site-specific operational conditions. The second form to be completed is a LTPP Traffic Sheet 17, WIM Equipment Inventory. These two sheets contain all the information necessary to complete a subsequent validation. Other forms that may be completed in the course of an assessment are LTPP Traffic Sheet 16, Site Calibration Summary; LTPP Traffic Sheet 20, Speed and Classification Studies; and LTPP Traffic Sheet 22A, Site Equipment Assessment with appropriate addenda. Each of these sheets may be found within Section 10 (Forms) of this document. They are generally referred to in the document as Sheet # (number) with their traffic data association being implied. There are other LTPP data sheets with the same numbers. None of them are applicable to LTPP traffic data collection.

2.1 PAVEMENT CONDITIONS

The smoothness of a WIM site can affect its ability to produce high quality data. An assessment will include checking the pavement smoothness using either the output from a high-speed profiler or a straightedge. It also includes a qualitative distress survey.

2.1.1 PROFILE DATA

Data collected with a high-speed profiler is the recommended method for evaluating smoothness of a WIM section. Guidelines on how to profile LTPP WIM sections are contained in the LTPP Manual for Profile Measurements and Processing (Ref. Error! Reference source not found.). Any profile data used for an assessment should be collected within a year of the site assessment visit.

The results of the profile smoothness evaluation using the LTPP WIM Smoothness Index software are to be reported in any assessment report. The current index thresholds are found in Section 5.1.1 (Reporting Results). A description of the software is found in Section 9.3 (WIM Smoothness Index Software).

This visit can also indicate if the 1000 foot (305 m) WIM section overlaps one of the SPS project sections, reducing the ability to improve smoothness by grinding or reconstruction. This length also

Version 1.0 Page 4 applies to instances where a 400 foot PCC slab has been installed in an AC pavement for sensor installation.

If it is not possible to run a high-speed profiler at the site, the straightedge methodology discussed in Section 5.2 (Smoothness Determination through Use of a Straightedge) may be used as an option. This option requires traffic control during the smoothness evaluation.

2.1.2 DISTRESS IDENTIFICATION

The distress survey provides a qualitative measure of surface anomalies that may affect truck motions across the sensors. This includes such items as potholes, patches, rutting, and asphalt-concrete transition.

The Distress Identification Manual for the Long Term Pavement Performance Program (Ref. 9) shall be used as the standard guide for interpretation, identification and rating of observed distresses.

Sheet 17 includes space to record distress photos taken at the site. At a minimum an assessment should include two photos of the area in the immediate vicinity of the WIM sensor location, one facing upstream and the other facing downstream from the sensor location. Additional photos showing distresses or pavement conditions such as asphalt to PCC transitions that may affect data collection are also to be included.

2.1.3 VEHICLE-PAVEMENT INTERACTION

Several trucks should be observed while they pass over the site to determine if truck traffic is showing adverse characteristics such as bouncing, swerving, braking, or accelerating within one hundred thirty-five feet (forty meters) of the sensors. If possible, it should be determined whether the truck tires are in full contact with the sensors.

A visual record of vehicle motion observations for a minimum fifteen minutes or thirty-five trucks whichever is longer is made. At low truck volume sites1 no more than an hour of video is expected. The camera is to be positioned so that the entire vehicle can be observed as it crosses the sensors.

In the event that there are potential tire contact problems, video for a minimum of fifteen minutes or ten trucks at pavement level near the affected sensor(s) should be taken. At sites with low truck volumes this video is not expected. However the possibility of tire contact problems should be noted for investigation during any subsequent visits at the site.

The results are to be reported on Sheet 22A. They include a verbal description and reference to all video taken. The references include tape (or disk) names and times.

2.2 TRAFFIC

Existing traffic and operational conditions are evaluated to determine equipment performance and to identify what vehicles are appropriate for a WIM validation. An additional element is the determination of a truck route and the number of trucks that will be required to efficiently evaluate the loading data for the WIM equipment.

1 A low truck volume site is defined as twenty or fewer vehicles per hour during daylight hours in Class 4 and higher.

Version 1.0 Page 5

2.2.1 CLASSIFICATION DATA

Classification information means the ability of the equipment to correctly classify vehicles according to its installed algorithm and thus to determine the vehicle distribution of the traffic at the site. A copy of the classification scheme and its associated algorithm for the equipment should be obtained either from the agency or by download from the equipment. A classification algorithm is used to assign vehicles to a classification scheme. Unless otherwise indicated, this document assumes the Traffic Monitoring Guide (TMG) 13-bin classification scheme in its discussions of vehicle classes.

While on site, a sample of vehicles is taken. The classifications of these vehicles are observed visually and compared to the classifier output in order to determine whether the equipment is classifying vehicles correctly. This sample should include a minimum of one hundred trucks unless such a sample would require more than three hours of collection effort. A systematic selection of vehicles should be employed (i.e. every fifth passenger vehicle and all vehicles visually identified in Class 4 and higher).

Class 3 vehicles should be included in the population to determine if there is difficulty in distinguishing between Class 3 and Class 5 vehicles when classes 2 and 3 are collected separately by the equipment.

Sheet 20 is to be used for collecting and reporting this data. The results of the classification verification are used to complete the Sheet 16 classification validation section.

Classification data should exist in the LTPP traffic database to give a preliminary indication of the predominant truck types. Weight data may exist to indicate their general loading characteristics. This information is used to choose the most appropriate test trucks for subsequent weight validations. A truck class is considered for inclusion in the test truck population if it comprises 10 percent or more of the truck population. Class 5 vehicles should only be included when they are 70 percent or more of the truck population and concurrence of the agency is obtained. Use of two trucks in Class 6 and higher is preferred since these heavier trucks have a greater impact on pavement performance. The loading characteristics are determined by using gross vehicle weight (GVW) graphs for the test truck classes.

These may be obtained from LTPP traffic data.

If there is a conflict in vehicle distribution patterns between the database and the on-site observations, the on-site observations control the selection of vehicle types. Loading characteristics are then determined using the truck selection criteria in Section 3.4 (Trucks).

Speed measurements are to be taken with a laser or radar gun during the classification sampling operation for comparison with the equipment output. A laser gun is preferred.

2.2.2 SPEED DATA

Speed data serves two purposes, a diagnostic if problems exist in classifying trucks and to identify the appropriate range of test truck speeds for validation purposes. Precision of speed data impacts the investigation of classification discrepancies and spacing errors.

The speed data should be requested from the agency prior to an assessment since it is not an element collected or required by the LTPP program. If it cannot be obtained in advance of the visit, it will be collected on site. Speed measurement accuracy is preferred to be within 0.25 mph, but must be no less accurate than 1 mph.

The site speed limit is entered on Sheets 17 and 22. Sheet 17 contains the recommended test truck speed ranges. Sheet 22 has the 15th and 85th percentile truck speeds.

Version 1.0 Page 6

2.2.3 TRUCK ROUTE

A route that can be successfully navigated by a 5-axle tractor-trailer is identified. The route is traveled at least three times to determine a typical travel time. The time and a written description of the route are included in the Comments section of the Sheet 17. In addition, a map showing the route is put in the Sheet 17 in the Site Map section.

2.3 EQUIPMENT

The items expected for equipment review are included on Sheet 22 and the associated addenda.

The addenda are differentiated by sensor type.

Field personnel will perform a visual inspection of all WIM site support components, including the interior and exterior of the controller cabinet, service mast, solar panels, visible conduit and all other system support structures and equipment

Field personnel will perform the equipment static electrical and electronic tests. A series of static electronic readings of the WIM system components as well as supporting equipment are taken and recorded on Sheet 22. Proper safety measures are taken to ensure that all test equipment is used properly.

Field personnel will identify and record deficiencies involving the supporting equipment that will require repair. All items whose present or deteriorating condition will eventually adversely affect the operation and/or accuracy of the WIM equipment are to be documented. Any unsafe conditions existing at or near the site will also be noted.

The physical layout of the sensors and cabinet along with distances between them in relation to the travel lanes and the LTPP lane is included as a sketch in the section labeled Site Layout of the Sheet 17.

2.4 REPORTING OF RESULTS

A memo (letter report) is prepared after each site assessment visit. For assessments conducted under FHWA direction it is sent to the LTPP Traffic Team Lead with copies to the LTPP Team Leader, the appropriate Regional Support Contractor (RSC) and the relevant agency contacts. It is forwarded within 2 weeks of the assessment.

The memo should contain the following at a minimum:

♦ A WIM site assessment was done at the SPS X located on route X at mile point X, direction on date. The LTPP Lane is identified as lane X in the WIM controller. .

♦ This site is/is not recommended for validation.

♦ The site is equipped with type of sensors and type of electronics.

♦ The equipment is/is not in working order.

♦ The pavement smoothness will not/ may or may not/ will affect validation efforts.

The mandatory attachment is Sheet 17. For sites that are operational and collecting data a Sheet 16 is also included.

Version 1.0 Page 7

For sites that are probably not going to successfully produce the quality data LTPP needs, an analysis and supporting data must be included. The following conditions can define a site that is “not recommended” for a validation:

♦ Non-working equipment ♦ Speed errors exceeding ± 1 mph with overall errors in classification exceeding five percent ♦ Classification data that is not research quality for trucks in Class 6 and higher ♦ Smoothness that may or will affect validation efforts

Version 1.0 Page 8

3 SITE VALIDATION - WEIGHT

Scale accuracy is central to the acceptance of quality traffic weight data. Site validation is the process by which trucks with known static weights are run over a site without adjusting any of the WIM and AVC equipment’s operational parameters. This gives an estimate of the precision and bias of the most recent data provided with the installed sensors and algorithms. After any required calibrations are concluded to reduce bias or improve precision, a final set of runs is made. These runs produce an estimate of the precision and bias of the data subsequently provided by the sensors until the next validation or termination of scale operation.

Scales that meet the tolerances in Table 3-1 below are considered to produce research quality loading data for LTPP. It is the expectation of the LTPP program that research quality data will be produced at all SPS-1, -2, -5, -6 and -8 locations. Failure to meet axle spacing tolerances is not grounds for denying a research quality label for loading data.

These tolerance levels are taken from ASTM E-1318-02. They apply with respect to the mean and standard deviation of the errors. The ASTM criteria of no more than five percent of the errors exceeding the tolerances for each element tested is not applied for determination of site performance for LTPP.

LTPP does not consider wheel load errors. The results of applying the ASTM criteria should also be reported for informational purposes.

Table 3-1 – WIM System Calibration Tolerances

SPS-1, -2, -5, -6 and -8 Sites 95 Percent Confidence Limit of Error Loaded single axles +20 percent Loaded axle groups +15 percent Gross vehicle weights +10 percent Axle spacing length + 0.5 ft

Any site validation for weight must also include a validation for classification as discussed in Section 4 (Site Validation-Classification).

3.1 BACKGROUND

Since collection of accurate traffic loading rates throughout the year is necessary to provide the loading rate estimates needed for SPS test section research, the WIM systems used in this effort must meet the criteria under a variety of conditions. Historically, many WIM systems have had problems accurately weighing vehicles when environmental conditions change from those present when the equipment was calibrated. Changes in pavement strength at the scale location (often caused by changes in pavement temperature or moisture content) are known to cause problems with WIM system sensor accuracy.

Additionally, vehicle dynamics change with vehicle speed. Therefore it is also necessary to ensure that the WIM systems are able to accurately weigh trucks at the range of speeds at which trucks travel at each scale site.

In this document, the term “scale” is assumed to include all components involved in the collection of vehicle weight data, unless otherwise specifically stated. The term “scale” is assumed to be synonymous with “Scale System.” The term “scale” is not restricted in meaning to those components (sensors) that

Version 1.0 Page 9 physically measure axle forces. It includes sensors, the electronics and software that interpret sensor signals, and converts that data into estimates of axle weights.

All WIM systems currently on the market use one or more “calibration factors” as part of the process for converting axle sensor information into axle and vehicle weight estimates. Generally these “calibration factors” are increased if the scale under-estimates static weights or decreased if the scale over-estimates static weights. In addition, many scales use an “auto-calibration” process that adjusts scale calibration based on monitored inputs. These adjustments are designed to take into account changing sensitivity of the scale sensors and electronics to changing environmental conditions and the aging sensors.

The goal of the validation is to verify that the scale will produce accurate vehicle weights under expected operating and environmental conditions. This includes demonstrating that the scale system does not produce outputs that are biased by temperature and/or vehicle speed.

The recommended scale performance verification process involves weighing trucks of known weight with the WIM scale, and then comparing the WIM system’s measurements with the known weights. The conditions under which vehicles are weighed by the WIM system are controlled to demonstrate that the system operates accurately under the majority of conditions expected to occur during LTPP data collection. During testing, the site should be operated in its “normal” manner. That is, if auto-calibration settings are normally “on” then they should also be “on” during the verification tests. Sites which use auto-calibration are not considered appropriate for use in LTPP data collection. Additional verification of the selected auto-calibration target and a demonstrated ability to return to calibration operating condition parameters is expected. LTPP does not provide a method for validation of sites that use auto-calibration.

The primary verification testing can be done with a minimum of two test trucks of different weights.

Once a weighing system has proved to be unaffected by vehicle speed and temperature over the course of a year, it is also permissible to verify sensor calibration by comparing the statically measured weights of vehicles pulled from the traffic stream at static scales with WIM system weights for those same trucks. This document describes only the use of test vehicles.

3.2 PERFORMANCE TESTING PROCEDURE

The scale accuracy tests are designed to verify the performance of the scale when pavement temperature and vehicle speed vary. The test plan requires the test trucks to make multiple passes over the scale at different specified speeds. Whenever possible site visits are made during times of the year when the greatest daily temperature variations are likely to be found at the site. Statistical tests are performed to examine system accuracy for the entire data collection period and for subsets of the data that represent subsets environmental or operational conditions.

The accuracy and reliability of data produced by the scale are computed by comparing the measured axle weights collected under these varying conditions with the known axle weights of the test vehicles.

Ideally validations are repeated two or three times during the year to make sure that the scale works accurately during all expected environmental conditions at that site. Additional validations may also be required if concerns about the accuracy of the scale under different temperature conditions are not fully resolved by the initial tests, or if calibration drift is detected by routine quality assurance tests.

Once the test trucks have been weighed at a certified static scale and the weights given to the WIM data collection staff, a preliminary determination of accuracy is made by making at least forty passes over the

Version 1.0 Page 10 scale. If the agency agrees and the ability to change calibration factors is available, then the system should be calibrated so that its output will have less bias and greater precision. A Sheet 16 is required to document the data quality using the initial forty runs. After calibration, a minimum of ten passes should be made to check the WIM readings against the static weights. The runs should be done using the same test trucks operated at least two different speeds. At least ten verification passes (five per truck) should be made after each calibration adjustment. The calibration process is repeated until the ten post-calibration passes confirm the scale’s accuracy is consistent with research quality data or it is determined that no further improvement can be obtained (generally three calibration iterations). All calibration adjustments should be made using procedures outlined in the scale vendor’s documentation.

3.2.1 SPEED VERIFICATION AND DATA COLLECTION

Before the test runs are started, the data collection crew should use the radar gun or laser speed measurement system to determine whether the WIM system is correctly measuring speeds of the traffic.

Speed measurement error is preferred to be within 0.25 mph, but must be no greater than 1 mph. The measurement of the speeds should be done according to the instructions provided by the equipment manufacturer. The data collection crew should also observe the reported spacings of the drive tandems on 3-axle tractors. The speed measurements are to be made of the normal traffic (cars and trucks) at the site. They do not require the test trucks.

If the average error for vehicle speeds is correct (within 1 mph), or if the average observed drive tandem spacing is between 4.2 and 4.4 feet, then test runs can begin. If both of these values are incorrect and there are frequent classification errors (in excess of five percent for trucks), the system may not be operating correctly, and the system’s input variables should be adjusted.

Once the crew has established that the system is correctly measuring vehicle speed, additional speed data does not have to be collected other than for the test vehicles. The data collection staff should continue to record the speed values output by the WIM system for the test trucks because this maintains the record of vehicle conditions needed to analyze system performance under different truck operating speeds. Speed measurements of test trucks should continue to be collected.

Speed data for the speed verification process is entered on Sheet 20. Speed data for the test trucks while load data is collected is recorded on Sheet 21

3.2.2 PAVEMENT TEMPERATURE DATA COLLECTION

Since not all test sites are expected to be equipped with accurate pavement temperature sensors, manual pavement temperature readings should be collected at each site with a validated infra-red temperature sensor. The suggested validation procedure is included in Section 10 (Forms) with the instructions for Sheet 30. The methodology is the same as that used for LTPP FWD hand-held infra-red temperature sensors (Ref. 6). Pavement temperature is preferred over air temperature as it more accurately reflects environmental impacts on the sensors. Because the use of air temperature data requires conversion to pavement temperature through generally accepted practices before doing temperature-based analyses it is not the preferred option.

If the WIM has temperature sensors that are not measuring surface temperatures the correlation between the surface temperature and the equipment’s temperature probe(s) cannot be established without additional information to verify the reasonableness of the WIM temperature measurements. In this instance the surface temperature should be the temperature collected and used in the analysis.

Version 1.0 Page 11

Manual temperature measurements should be made near but not on the sensor. The temperature should be collected immediately after the trucks pass the scale. Prior to collecting temperature data a location within two feet of a WIM sensor will be selected and marked. If the materials for the shoulder and pavement are the same OR the LTPP lane has more than one vehicle every twenty seconds, the location will be on the shoulder. If the materials are different and the LTPP lane has less than one vehicle every twenty seconds, the location will be six inches into the lane from the shoulder marking and at least one foot from a WIM sensor. The area for temperature measurement will be marked with a 6 x 6 inch square using chalk. In taking the measurement the personnel will stand facing traffic and hold the sensor approximately 30 inches (about hip height) above the pavement. The measurement is recorded on Sheet 21.

To obtain a wide temperature variation2, it may be necessary to collect data over more than 8 hours in a day. FHWA-LTPP and the agency should consult prior to the scheduling of the field tests on the timing and duration of the validation.

Where possible, at least 12 test runs should be performed for each temperature range. These should be performed to effectively cover the range of test speeds. For example, if time is available to make one additional pass per temperature condition, the additional run might be made at the speed at which the majority of trucks operate or at the speed with the fewest test runs.

It is also important to collect data after the temperature has started to decline to determine whether cooling of the upper pavement layers (while the lower layers stay warm) affects WIM output.

While every effort will be made to select a day when pavement temperatures will vary widely during the testing procedure, it is possible that during the selected test day(s) little temperature variation will occur.

Successful tests taken under these “less than desired” conditions will still be considered validation of the scale’s calibration. However, it is expected that additional tests focusing on the scale’s temperature sensitivity will be conducted at that site subject to resources and scheduling opportunities. Validating the site during different seasons each year may be necessary.

3.2.3 RUNNING TRUCKS

The required minimum test runs are:

♦ 40 initial runs at least two different speeds prior to any calibration efforts to determine the initial equipment performance ♦ 10 runs at two different speeds after each change in calibration, repeat until calibration is satisfactory ♦ 40 runs to validate the final calibration of the scale. The 10 runs performed after the last calibration change may be included in these 40 runs.

At the direction of FHWA if the site meets the loading tolerances with a bias of less than ±2 percent, the validation may be terminated after the initial 40 runs.

2 A wide temperature variation is considered 30 degrees Fahrenheit or more. Ranges less than 30 degrees Fahrenheit may produce fewer temperature groups for analysis.

Version 1.0 Page 12

The use of three different speeds is preferred. A 30 degree Fahrenheit minimum pavement temperature range is the target for varying pavement temperatures.

Once the test trucks have been weighed at a calibrated, certified static scale that meets Handbook 443 specifications, they should make repeated passes over the sensors. When approaching the WIM scale, drivers should hold the vehicle speed constant. They should report after any runs where they were not able to maintain a constant speed. (This contact should take place after they have completely crossed the scale, as radio system broadcasts can cause interference with the WIM system outputs, causing false readings.) In most case trucks will drive a continuous loop around the scale following the pre-determined truck route. Drivers should contact the data collection crew prior to unscheduled breaks or when conditions warrant a change in the test schedule.

Target test speeds should be differentiated by at least 5 mph and preferably 10 mph. For optimum differentiation in test truck speed groups a minimum of a 15-mph spread should exist between the highest and lowest speed to be tested. Using only two speeds should be considered when the difference between the 15th and 85th percentile speeds is 10 mph or less OR the 85th percentile speed exceeds the speed limit. Speeds below the 15th percentile speed for trucks should not be used. Four speeds should be considered when the difference between the 15th and 85th percentile speeds exceeds 20 mph.

Drivers should switch from high to medium to low target speed with each pass, continually rotating among assigned target speeds. This pattern will increase the sample size for each of the speed-temperature combinations.

Sufficient time should be maintained between each truck pass in order to record the data for each truck.

If trucks are run in platoons the ability to print or download the data after the set of runs or after system adjustments are made should be verified and tested before relying on post-run downloads for analysis data.

A total of forty runs is the minimum required to have an acceptable data set for analysis prior to the first calibration. If turnaround times are such that two trucks between them cannot complete forty runs in a 10-hour site visit (breaks included), additional trucks should be used.

It is important to note the following points –

♦ The test trucks should move at a constant speed across the sensors.

♦ Vehicle passes must be made at a variety of speeds.

o The trucks should not be operated at speeds above the posted limits.

o Truck operation should not cause safety problems by operating too slowly.

♦ Trucks should be as centered in the lane as much as possible when crossing the scale.

3.2.4 LOADING DATA COLLECTION

For each vehicle pass, the data collection staff should record the WIM system’s output for axle weights and axle spacings. Each test vehicle’s speed when crossing the sensor (collected with the laser or radar

3 “Specifications, Tolerances and Other Technical Requirements for Weighing and Measuring Devices”: National

Institute of Standards and Technology (NIST) Handbook 44, U.S. Government Printing Office Superintendent of Documents, Mail Stop: SSOP Washington, D.C. 20402-9328, ISBN 0-16-046313-7.

Version 1.0 Page 13 gun) may need to be obtained for comparison with the WIM system’s output. The following data from the WIM system should be obtained:

♦ The sequence number of the test run ♦ The date of the test run ♦ The time of the test run ♦ Axle weights of the test trucks as they pass over the scale ♦ Spacing between each axle on the test truck ♦ Speed of the test truck ♦ Pavement temperature at the time of the test run

The sequence number is the reference number used by the WIM system software to indicate this particular weight record. It may also be called the “record number” or “WIM assigned vehicle number.”

If information can be downloaded (retrieved) at the completion of the runs, this number can be used to crosscheck data entered during the runs.

These measured test truck attributes can be obtained by reading the values off the WIM scale display screen. This data should also be stored permanently and retrieved later for confirmation that the data entry at the site was correct. Some WIM systems also allow these data to be printed from a portable printer or written to an ASCII file for later retrieval. An Excel spreadsheet described in Section 9.1 (WIM 2 TRUCK) has been created by FHWA-LTPP to assist in recording and analyzing the data collected. It is available from LTPP Customer Support. The preferred alternative is to use Sheet 21 if WIM 2 TRUCK is not being used.

Where differences in the weights recorded by the system and the observer occur (that is the screen of the computer on-site rounds up a value that is carried to more significant digits in the main WIM system record), the WIM system record should be used in the analysis effort.

Only records that correctly report the number of axles on a test truck should be included in the analysis.

For example a 5-axle truck record must have five consecutive non-zero axle weights and four consecutive non-zero axle spacings to be valid for analysis. When wheel load data are used, only records with the correct number of sequential non-zero wheel loads should be used. Errors in identifying test trucks do not count towards the 40 run minimum. Errors in collecting test truck data by the equipment should be noted in the validation report. The specific type of error(s) should also be described.

3.2.5 PROCESS BY WHICH REQUIRED NUMBER OF TEST RUNS WAS DETERMINED

A minimum of thirty-six test runs is required based on the desired number of speed temperature combinations. Four test runs are needed at each of three test speeds within each of three temperature ranges (4 x 3 x 3 = 36) for ideal conditions. However this is an absolute minimum. Additional test runs are encouraged whenever possible. The forty run limit was established to improve the chances of getting at least twelve runs in each temperature and each speed subset. The total number of runs should be split evenly between all trucks used. This may result in making more than the minimum number of runs. The minimum number of test runs is not reduced for sites at which fewer speed ranges or temperature ranges are anticipated.

Version 1.0 Page 14

Additional runs may also be necessary if conditions at the site prevent some planned runs from being performed at the desired speeds or if traffic conditions cause improper test conditions to exist (for example the test vehicle must brake to avoid another vehicle while approaching the sensors).

3.3 WIM DATA ANALYSIS

Once the data have been collected in the field, statistics must be computed to determine whether the WIM site meets the accuracy standards for research quality data set by LTPP. The basic statistic required for this test is the error expressed as a percentage of the known value. The percentage of error calculated from the data collected for each run is then used to compute a series of summary statistics.

These summary statistics are used to determine whether the scale is producing research quality data.

The first set of tests requires data from all of the 40 plus test runs performed at the site. It provides a general overview of system performance, given all of the environmental and vehicle speed conditions that were present during the testing. To perform these tests, the analyst should calculate the mean and standard deviation of the ‘percentage of errors’ for these variables:

♦ Front axle weights of the test trucks ♦ Weights of all non-front axle single axles ♦ Weights of all tandem axles4 ♦ Weights of any other multi-axle groups by axle group type ♦ Weights of all single axles ♦ Weights of all multiple axle groups as a set ♦ Gross vehicle weights

The percentage error is computed as the equipment reported value less the static value and the result divided by the static value.

For each of the variables, the analyst should compute the 95% confidence limits for the percentage of error using (X + 1.96σ) and (X – 1.96σ) where:

X = mean value of the percentage of error for the statistic σ = standard deviation of that percentage of error

This value (for each variable) should then be compared with the values shown in Table 3-2. If any of these values exceeds the associated range in Table 3-2, the WIM system does not produce research quality data. If all of the loading values are within the ranges in Table 3-2, the system produces research quality data.

4 A tandem axle is an axle group in which the consecutive vehicle axles are at least 3.3 ft (1 m) and not more than 8 ft (2.4 m) apart. If the axle spacing of a pair of axles on the same suspension exceeds the maximum distance each axle is considered a single axle and the tandem called a split tandem.

Version 1.0 Page 15

Table 3-2 – WIM System Calibration Tolerances – Large Sample

For SPS WIM Sites X + 1.96 σ Single axles +20 percent Tandem axles (and other axle groups) +15 percent

The analyst should also compute the mean and standard deviation of the error for the following variables:

♦ Drive tandem axle spacings ♦ All axle spacings

The 95% confidence limits for these variables should be computed using the same method as for the other variables and compared against the spacing value in Table 3-2. Not meeting the axle spacing criterion does not impact the decision on the research quality label for the loading data.

The scale being tested may pass the research quality test outlined above, but it still may not truly be able to produce research quality data for vehicles traveling at different speeds or under different temperature conditions. Subsets of data grouped by speed or temperature are also analyzed to determine of the subset is of research quality.