Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
21 CHAPTER 4 METHODS FOR DATA STORAGE AND RETRIEVAL INTRODUCTION The method used for data storage depends on the type of data being collected. For example, administrative data usu- ally are collected manually. Monitoring data from old facili- ties were collected manually, whereas automated data col- lection systems are used by newer facilities. Future facilities are likely to use either automated or semi-automated data col- lection systems. Administrative Data Elements In general, most administrative information is readily avail- able to facility operators. For facilities that are owned and operated by state agencies, principal investigators are often responsible for the projects being carried out. Information regarding specific test programs may be contained in research contracts, correspondence, or other documents, whereas infor- mation pertaining to the APT facility itself and its objectives is available from facility owners/operators. However, the principal investigator often maintains data on a particular test or project, including information on sponsors and key re- search personnel. Load Application Data Elements To control the magnitudes of loads applied to the test pave- ment, some APT devices employ gravity loads with various types of wheel suspension systems; others use a hydraulic counterweight mechanism to stabilize the load magnitude. None of the existing APT facilities duplicate actual truck sus- pension characteristics. Some APT loadings are applied in a channelized fashion (that is, with no wander), while others are applied using a wander pattern that simulates the lateral distribution of high- way traffic. Various tire pressures have been used, and often the tire pressure is monitored and recorded regularly using remote pressure sensors. In some APTs, loads are applied in either a unidirectional or bidirectional mode, while in others unidirectional loadings are applied. Although loads are generally recorded using automated equipment and data storage, the magnitude or position of each and every applied APT load is not recorded. Instead, the loads are monitored on a regular basis for verification (for exam- ple, check of longitudinal or lateral position and load magni- tude). The number of wheel passes corresponding to a spe- cific loading pattern are generally recorded using automated equipment on a continuous basis. When using gravity loads, it is important to check and record the actual loading because the roughness of the test sec- tion can result in loads that are alternately larger and smaller in magnitude than the nominal gravity load. Pavement Description Data Elements Project- or test-specific information relating to pavement description (for example, structural and geometric details) is recorded under this category of data. The principal investi- gator is generally familiar with this information. However, project design and bid documents and reports, as-constructed records, technical reports, and historical archives may also contain such information. Occasionally, limited field tests are conducted to determine some of this information. Test objectives define the test pavement origin and design, pavement type, and special construction requirements. For specially constructed pavements, all the data pertaining to the type of pavement, subgrade, and intermediate layers, as- designed and as-built cross section, test bed and traffic lane dimensions, and other test-specific information (for example, PCC slab dimensions, reinforcement, and load transfer infor- mation, if applicable) should be recorded. For tests conducted on existing pavements, a search of the historical archives may be required to obtain informa- tion on pavement type, age, and traffic history. Pavement history (construction and maintenance) data are found in the as-constructed and annual maintenance record files. How- ever, if such data are unavailable, exploratory field testing by DCP, GPR, or coring may be required to determine the type of the existing pavement structure and other structural details. Project contract documents should provide information about the pavement contractor. Material Characterization Data Elements Typically, material characterization data are obtained from laboratory tests or field investigations. These should include
the source of the material, date and method of sampling, and type and method of testing. Laboratory tests may use manual or automated procedures. Automated test methods may pro- duce large amounts of data, which may necessitate a multi- stage data collection and electronic storage medium. Perma- nent deformation testsâsuch as AASHTO TP7, âTest Method for Determining the Permanent Deformation and Fatigue Cracking Characteristics of Hot Mix Asphalt (HMA) Using the Simple Shear Test (SST) Deviceââproduce large data files. The data are generally postprocessed, and only specific sample data are used for calculations; the original data files may be archived as âoff-lineâ data. The data used to calcu- late specific material characteristics (that is, strain or resilient modulus), might be part of the âon-lineâ data in the overall database design. The collection of the âon-lineâ data may also be automated or semi-automated. Many material characterization methods are conducted manually. The data relevant to collection and handling are the same as those for materials tested with automated test meth- ods. Measurements and observations are typically recorded on worksheets. Calculations are then made, and the results are transferred to a final test report sheet. Often, the work- sheets are only kept until the test is completed and measure- ments and calculations are checked. Only a small subset of the data collected during the test process may be transferred into a database. For example, while a standard moisture- density test would produce a moisture-density curve, only val- ues for maximum dry density and optimum moisture content are extracted and entered into the database. Environmental Conditions Data Elements Automated methods are well suited for collecting envi- ronmental data. Although the intent of APT is to apply loads to a pavement over a short period of time, the test may con- tinue for months; the ambient conditions are best collected with automated methods. On-line weather stations and instrumentation will produce data streams that will likely be processed by automated means and placed into a database. Although indoor facilities may be maintained at constant temperature and moisture conditions during the conduct of a test, monitoring the conditions is normally integral to the over- all facility. Outdoor test tracks will be exposed to normal tran- sient conditions and are best monitored by automated means. However, some environmental or climatic data elements, (for example, surface temperature during special load tests) may best be measured and recorded manually. Pavement Response Data Elements Pavement responses (for example, deflection, stress, and strain) are monitored to establish relationships between response, traffic or axle loads, and pavement performance for use in predicting field performance of pavements. How- 22 ever, these relationships should consider effects of climate and aging. It is also important to differentiate between pavement sur- face responses and pavement responses at depth. Deflection tests at the surface of the pavement using an FWD or other deflection-based devices do not disturb the pavement struc- ture and therefore do not affect pavement materials. How- ever, the installation of response gages within pavement depth may affect the properties of the materials. Although pavement response data are almost exclusively recorded and stored using automated equipment, it is gener- ally not necessary to record these values under each and every applied load. Instead, pavement responses are measured on an intermittent basis to monitor both their magnitude and peri- odic changes. Pavement Performance Data Elements Pavement performance data are measured manually most often and semi-automatically on occasion. These data are usually entered and stored in the database on a periodic basis. Crack surveys usually consist of a manual measurement of cracks along the wheel track and characterization of their extent and severity; the direction of the cracks is also noted. Usually, a photographic record of each âmilestoneâ (for exam- ple, every 25,000 load repetitions) is made to document the development of cracking or other types of surface distress. Transverse cross sections (to measure rutting) or the longi- tudinal profile (to calculate roughness) of the pavement sur- face may be measured using rod and level or with laser sen- sors. Measurements are usually made at more than one line in each test section. For example, transverse profile may be mea- sured every meter in an 8-m test section, that is, nine mea- surements for each measuring sequence. One or more longi- tudinal profiles may be measured depending on wheel load wander and other factors that may influence the longitudinal profile, such as environmental effects outside the wheel path. For jointed plain concrete pavement (JPCP), pumping may be monitored visually or through photographs. Joint faulting is measured in conjunction with the profile survey or manually with the surface distress survey at each measuring sequence. DATA STORAGE AND RETRIEVAL Recent improvements in computer technology and auto- mated data collection make it easy to collect and store large amounts of data. It is impractical to maintain such data using either paper filing systems or custom software programs manipulating standard sequential files. Database systems have become the most viable means of maintaining and utilizing the large quantities of data collected by APT devices. The storage and retrieval of data encompasses both hard- ware and software. With regard to the hardware, data have been stored on devices ranging in simplicity from paper to
complex optical disks and flash memory cards. With regard to software, data storage has ranged from written information (for example, tables) filed in folders and stored in cabinets to electronic text files and spreadsheets for small data amounts to dedicated databases for large data amounts. Hardware The most familiar form of data storage and retrieval is paper; observations are recorded on paper and stored for later use and analysis. These data are most likely transferred to an electronic form before analyses are conducted. The main advantage of paper storage is ease of use, but such data are hard to work with. Paper storage is appropriate for very sim- ple data sets that do not involve a large number of repetitive calculations. Much of the data are collected in some electronic form on electronic storage media. Storage media range in simplicity from floppy disks for small databases to flash cards to hard drives and optical disks for larger databases. The current floppy disk (89-mm micro floppy) operates on the principles of magnetic recording using magnetic heads for data storage and retrieval on a single rotating magnetic disk. Because of their limited capacity 1.44 megabytes MB and extremely low data transfer rate (0.06 MB/sec), floppy disks are useful only for storing small data files. However, they do offer universal compatibility and low cost. Hard disk drives contain several spinning disks that are read from, and written to, using separate read and write heads that float above the disks with a separation in the order of 10 to 20 microns. These drives are sealed permanently to protect the disks and heads from dust particles. Over the past few years, the fixed hard disk drive technology has improved; drives with larger storage capacity are becoming less expensive. There are a number of other removable mag- netic storage media devices with different sizes [for example, 40 MB Iomega Click! Drive, 100 MB and 250 MB Zip drives, and 1 gigabyte (GB) and 2 GB Jaz drives]. Flash memory cards are electronically programmable and non-reprogrammable solid-state data storage devices that use flash memory chips to store data. Entire sections of the microchip are erased (or flashed) at once. These cards lose power when they are disconnected, but the data are retained for long periods of time or until the microchip is rewritten; these are normally used in laptop PCs and digital cameras. Many types and configurations of these cards are available with memory ranging from 1 MB to 1 GB. Compact diskâread only memory (CD-ROM) technology was introduced as CD-digital/audio (CD-DA). CD-ROM fea- tures include standard design and physical structure of the disk, data format, and error correction code schemes. A CD-ROM is 12 cm in diameter with a 1.5-cm hole and is 1.2 mm thick. CD-ROMs can store up to 650 MB of data (74 minutes of play time for CD-DA); they are a highly reli- 23 able means of data storage with good data protection from damage, both inside and outside the CD-ROM drive. An advancement in CD-ROM technology is the inexpen- sive CD-ROM/CD-RW combination drive for personal com- puters, which can easily copy data to recordable CDs (CD-R) or a rewriteable CDs (CD-RW) that look almost like a CD- ROM. These CDs are inexpensive and mobile writeable stor- age media. The DVD (digital versatile disk) is an optical storage sys- tem that, like a CD, has read-only, recordable, and rewrite- able versions. DVDs are likely to replace CDs in the future; current DVD drives are compatible with CD media. DVDs can store up to 17 GB of data, compared with the 650-MB capacity of CD-ROMs. Software Data that are recorded on paper and stored in file cabinets can be retrieved manually. Depending on the importance and amount of data collected, electronic storage in text files, spreadsheets, or dedicated databases may be warranted. Electronic text (*.txt) files are simple ASCII files that can be read by most word processing, spreadsheet, and database software programs. They are platform independent and can be read on IBM compatible PCs, Apple PCs, and mainframes. A shortcoming of text files is that they cannot incorporate text attributes, such as bold and underlined characters. Rich text format (*.rtf) files can retain formatting and can be opened by major word processors in both IBM and Apple environments. Data stored in text files can have data fields that are sepa- rated, or delimited, by a comma, a tab, or a space. Each row represents a data record. Such data-delimited text files can be read into a word processor, spreadsheet, database, or special- ized statistical package for further manipulation and analysis. Spreadsheet programs can be used to store and manipulate fairly large data sets, constrained only by available memory and PC processor speed. A spreadsheet allows the user to orga- nize information into both columns and rows. Each cell of the spreadsheet, defined as the unique intersection point of a col- umn and a row, can contain a label, a value, or a formula. A label provides descriptive information, a value is a number, and a formula manipulates values and labels. Though spread- sheets have been used as databases for small amounts of data, they are generally difficult to verify and audit and do not pro- vide good tools for managing data, whether in terms of con- solidation or searching for specific details. When used as data- bases, spreadsheets are unable to display one record (row) at a time and do not allow a multiple-report format. Relational links to other tables and data are also not supported. Dedicated databases that arrange information in tables and records are best suited for large-scale data storage, manipu- lation, and retrieval. Traditional databases are organized as fields, records, and files. A field is a single piece of informa- tion; a record is a complete set of fields; and a file is a col- lection of records. A âdatabaseâ or a database management
system (DBMS)âconsisting of a collection of programs that enable entering, selecting, and organizing data in a databaseâ is used to access information from a database. State-of-the-Art in Databases The purpose of recording and storing data is to make them available at a later date for use in analyses. The data types, data terms, units of recording, and format may vary greatly from project to project. An electronic database is a collection of information optimized for quick selection of desired data using a DBMS. The relational database, or automatic navi- gation system, is the state-of-the-art system for data storage and retrieval; it does not require the user to specify how to retrieve the data but merely what should be retrieved. Hier- archical and network models, on the other hand, require that the user understand how the data are structured within the database. Data stored in a standard relational database system can be used by more than one application. It can be loaded, analyzed, manipulated, and stored in a way that suits the format of each user. Data stored in a standard relational database system are retrieved and manipulated using standard database manipula- tion language, Structured Query Language (SQL). SQL Server is a relational DBMS that provides centralized security, data integrity and control, rich user interfaces, and a variety of off-the-shelf productivity tools. SQL Server is known for high performance and scalability and provides sup- port for very large databases. SQL Server employs a dynamic locking architecture that keeps concurrent users from inter- fering with each other during queries and updates. It imple- ments comprehensive user-level permissions on tables, views, stored procedures, and SQL commands. It also supports field- level database security features; permissions can also be applied to groups. The popularity of the relational model is due in part to its use by most microcomputer DBMS (for example, Microsoft Access) and its ease of use and understanding of the rela- tional model. In particular, it is a simple matter to train end users to retrieve data from relational databases through the use of SQL, a fundamental standardized component of any relational DBMS. Because it is a common, standardized sys- tem, a person learning to use SQL for a microcomputer data- 24 base would find little difference in retrieving data from a relational database implemented on a different hardware platform, such as a client-server workstation or a mainframe computer. The Jet database engine is a generalized piece of software that provides the ability to store data in, and retrieve data from, a range of DBMSs (for example, Microsoft Access). In other words, when Microsoft Access is used to manipulate a database, Jet is behind the scenes performing all of the real work. While it is optimized for accessing Access database files directly, Jet can attach to any database that supports the Open Database Connectivity (ODBC) interface. This means that an end user can manipulate a database in any format as long as the userâs computer has the appropriate ODBC driver. Databases that Jet can manipulate include Microsoft Access, Oracle, dBase 5, Btrieve 6.0, and FoxPro. Data Access Objects (DAOs) are the clearly defined pieces of code that provide an interface to the functionality of Jet. DAOs allow a programmer to manipulate databases within his or her working environment (for example, COBOL, C, C++). In other words, DAOs allow a programmer to design and write his or her own DBMS software programs The main purpose of a data storage system is to store and retrieve data. Factors to be considered in the selection of a data storage system include data safety, ease of use, storage capacity, cost, performance, reliability, and manageability. PROTOCOLS FOR COLLECTION OF DATA ASSOCIATED WITH APT A list of data collection protocols was prepared for data cat- egories and data elements. When materials are characterized using a nonstandard test (for example, in research situations), the test procedure should be documented in the database. Some data elements, especially those belonging to the admin- istrative data elements, will not require a standard protocol for data collection. Some of the standard protocols, where sam- pling or testing is conducted as a function of stationing along the test section, may need to be modified to accommodate ATP devices that operate over very short sections of pave- ment. The use of standard data definitions and adequate doc- umentation of test methods will ensure proper interpretation of data by users.
