Inspection Guidelines for Bridge Post-Tensioning and Stay Cable Systems Using NDE Methods (2017)

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.

D-1 Customizing the Decision Matrix A p p e n d i x d

D-2 INTRODUCTION Several NDE technologies that were used for the condition assessment of the full-scale PT girder and the large-scale stay cable specimens have been considered in this investigation. They were also evaluated based on several weight categories and two sets of weight factors. However, as new NDE technologies or modifications of the existing technologies emerge, it may be necessary to include these technologies in the decision matrix. In addition, if the end user decides that additional evaluation categories need to be added to the matrix or if the weight factors for these categories need to be modified, then the decision matrix considered in this study lets the user include these into the decision matrix. The procedure for customizing the decision matrix in terms of adding a new technology, a new category, or changing the weight factors are discussed in detail in this section. The step-by-step procedures are illustrated using examples. STEP-BY-STEP EXAMPLE OF CUSTOMIZING THE DECISION MATRIX The decision matrix is a multi-dimensional scoring and ranking process. In order for the decision matrix to be repeatable and general enough for inclusion of additional methods as new technologies are developed in the future, it relies on metrics that quantitatively relate the needs of a structural investigation to the performance of the NDE methods. This section provides detailed description of the developed 3D and 2D decision matrix and provides step-by-step procedure for adding a new NDE technology X and two new categories Y and Z to the existing decision matrix. Adding Technologies to the Decision Matrix NDE technologies were evaluated based on five categories: precision, accuracy, ease of use, inspection requirements, and cost. For precision and accuracy criteria, each NDE technology was given scores for identifying six different deterioration conditions (corrosion, section loss, breakage, compromised grout, voids, and water infiltration) in internal metal ducts, internal nonmetal ducts, external metal ducts, external nonmetal ducts, and anchorage regions. This leads to a multi-dimensional decision matrix as shown in Figure D-1. Thevfigure also illustrates the application of the multi-dimensional Weighted Sum Model (WSM). This matrix visually and mathematically describes how each NDE technology is given an overall score ( ) based on individual scores ( ) for its precision and accuracy. These scores are given for each of the six conditions listed above (making it a multi-dimensional decision matrix). The score of each individual method for a considered deterioration condition can then be calculated as: where index stands for the NDE technology (changes from 1 to 18 for 13 NDE technologies plus five combinations considered in this study), index stands for the defect conditions (changes from 1 to 6 for the six deterioration conditions) and index stands for criteria (changes from 1 to 2 for the two categories – precision and accuracy). For example is the score for method 1 (GPR), condition 6 (water infiltration) and category 1 (precision). Similarly and are the weight factors for precision and accuracy, respectively. Then to obtain the total score (for precision and accuracy) of method 1 (GPR) for condition 6; .

D-3 Although the precision and accuracy categories have multi-dimensional parameters including deterioration conditions, the other categories (ease of use, inspection requirements, and cost) are not affected by the deterioration condition. Therefore a 2D decision matrix is developed as shown in Figure D-2 that describes the implementation of a 2D WSM using the definitions for the “alternatives” (NDE methods) and “criteria” (ease of use, inspection requirements, and cost). This matrix visually and mathematically describes how each NDE method is given an overall score ( ) based on individual scores ( ) for its ease of use, inspection requirements, and cost. The score of each individual method for a considered technology can then be calculated as: where index i stands for the NDE method (changes from 1 to 18 for 13 NDE methods plus five combinations), and index k stands for criteria (changes from 3 to 5 for three categories – ease of use, inspection requirements and cost). For example a is the score for method 1 (GPR), and category 5 (cost). Similarly w is the weight factor for cost. Then to obtain the total score (for ease of use, inspection requirements, and cost) of NDE method 1; S w w a w . The overall score becomes S . For instance the score of Method 1 (GPR) for condition 6 (water infiltration) = S . The below two part decision matrix approach is proposed as it can be relatively straightforward for the end user to modify. For instance, suppose the end user desires a new technology or modification of a technology to be added to the NDE library used in the matrix. Using the visual representation in Figure D-1 and Figure D-2, they would simply add another set of rows to the decision matrix and provide individual scores based on the same category definitions and application assumptions used by the existing technologies. For instance, the end user desires to add a new NDE technique, say Technology X, to the decision matrix. The new NDE method, Technology X, can be added to the matrix system as method no. 19. Figure D-3 shows the updated multi-dimensional decision matrix, including NDE Technology X, for scoring the NDE method in terms of precision and accuracy. Figure D-4 shows the updated 2D decision matrix, including NDE Technology X, for scoring the NDE methods in terms of ease of use, inspection requirements, and cost.

D-4 Figure D-1. 3D decision matrix for scoring NDE methods in terms of precision and accuracy.

D-5 Figure D-2. 2D decision matrix for scoring NDE methods in term of ease of use, inspection requirements, and cost. As an example if the end user wants to calculate the overall score of the new NDE Technology X (method no. 19) for detecting water infiltration defects (condition 6); from Figure D-3 the total score for precision and accuracy: S w w , and from Figure D-4 the total score for ease of use, inspection requirements and cost: S wa w w . Then the overall total score of Technology X for detecting water infiltration defects =S . It would also be necessary for this new or updated technology to have a testing procedure that describes the technology’s abilities given the variable physical parameters (see the general testing procedure template in Appendix C).

D-6 Figure D-3. Adding new Technology X to the 3D decision matrix for scoring NDE methods in terms of precision and accuracy.

D-7 Figure D-4. Adding new Technology X to the 2D decision matrix for scoring NDE methods in term of ease of use, inspection requirements, and cost. Adding Category to the Decision Matrix If it is desirable to add or modify a new category for scoring (e.g., applicability for humid environments), a new set of columns can be added to the decision matrix as well as scoring guidelines and definitions for the new categories. Depending on the definition of the new criteria it should be added either to the 3D or to the 2D decision matrix. If the new criteria, say Category Y, is applicable to deterioration conditions, like accuracy does, it should be added to the multi- dimensional decision matrix as shown in Figure D-5. However if the new category, say Category Z, is not applicable to the deterioration conditions but is applicable only for evaluating the NDE Ease of Use, w 3 Insp. Req's w 4 Cost w 5 1 GPR a1,3 a1,4 a1,5 S1 2 IRT a2,3 a2,4 a2,5 S2 3 ECT a3,3 a3,4 a3,5 S3 4 MFL a4,3 a4,4 a4,5 S4 5 MMFM Permanent a5,3 a5,4 a5,5 S5 6 MMFM Solenoid a6,3 a6,4 a6,5 S6 7 IE a7,3 a7,4 a7,5 S7 8 UST a8,3 a8,4 a8,5 S8 9 USE a9,3 a9,4 a9,5 S9 10 SPV UPV a10,3 a10,4 a10,5 S10 11 LFUT a11,3 a11,4 a11,5 S11 12 Sounding a12,3 a12,4 a12,5 S12 13 EIS a13,3 a13,4 a13,5 S13 14 GPR/USE a14,3 a14,4 a14,5 S14 15 GPR/IE a15,3 a15,4 a15,5 S15 16 MFL/ Sounding a16,3 a16,4 a16,5 S16 17 MFL/IE a17,3 a17,4 a17,5 S17 18 IRT/USE a18,3 a18,4 a18,5 S18 19 X a19,3 a19,4 a19,5 S19 Category Score N D E M et ho d

D-8 technology, it should be added to the 2D decision matrix as shown in Figure D-6. Individual scores for every existing NDE technique would need to be computed based on the definitions of these new categories. Furthermore, the individual scores would be provided based on the defined deterioration conditions. Each testing procedure would then need to be modified so as to include the effect of this parameter. The ranked list of NDE techniques listed in the condition assessment flowcharts will also have to be updated accordingly. As an example if the end user wants to add one new category to the multi-dimensional decision matrix (Category Y in Figure D-5 with weight, w ), and another category to the 2D decision matrix (Category Z in Figure D-6 with weight, w ) the formulation of total score changes as follows: • The score of each individual NDE method for a particular deterioration condition in terms of multi-dimensional categories can be calculated as: w • The score of each individual NDE technology in terms of the 2D categories can be calculated as: As an example if the end user wants to calculate the overall score of the new NDE Technology X (method no. 19) for detecting the water infiltration defects (condition 6) that includes the new Categories Y and Z; from Figure D-5 the total score for precision and accuracy: S w w w and from Figure D-6 the total score for ease of use, inspection requirements and cost: S w w w w . Then the overall total score of Technology X for detecting water infiltration defects =S .

D-9 Figure D-5. Adding new Category Y to the 3D decision matrix for scoring NDE methods in terms of precision and accuracy.

D-10 Figure D-6. Adding new Category Z to the 2D decision matrix for scoring NDE methods in term of ease of use, inspection requirements, and cost. Modifying Weight Factors Two scenarios that are believed to be most common in bridge inspections are considered in this investigation: “Scenario 1: Cost-Driven Approach” and “Scenario 2: Accuracy-Driven Approach”. The first scenario expectedly emphasizes (by weight) the category “Cost,” and the second scenario emphasizes (by weight) the category “Accuracy.” The rationale behind these two scenarios is that end users will likely be choosing methods for either a large-scale bridge inspection program (where cost of the NDE technique is a primary deciding factor), or a single bridge inspection in which possible areas of deterioration are known or expected to exist (where accuracy of the NDE technique is a primary deciding factor). The associated weight factors for these scenarios are given in Table D-1.

D-11 Table D-1. Categories and weight factors for Scenario 1 and Scenario 2. Weighted Category Scenario 1 Weight Factors Scenario 2 Weight Factors C1 Precision W1 5% W1 5% C2 Accuracy W2 10% W2 60% C3 Ease of Use W3 15% W3 15% C4 Inspection Requirements W4 10% W4 10% C5 Cost W5 60% W5 10% Each category has its own weight and the end user may want to modify the weight factor due to addition of new criteria or to emphasize some of the criteria relative to others. If it is desirable to add or modify a new criteria for scoring, it could be done following the procedures described in the earlier section. Once the criteria is added the weight factors need to be modified. The WSM works by providing an overall score for each one of the “alternatives” (NDE methods) and by summing their individual scores under N “criteria” multiplied by the weight of the associated criterion. Furthermore, each of the criteria must be weighted to emphasize their desired influence among the remaining criteria, and all of the individual weights must sum to unity. For example, if the end user considers adding “Category Y” and “Category Z” and creates a similar cost-driven and accuracy-driven approach the weight factors of each criteria could be assigned as shown in Table D-2. The sum of all individual weights is unity and the weight factors still emphasize on the cost for Scenario 1 and on accuracy for Scenario 2. Table D-2. Categories and weight factors for Scenario 1 and Scenario 2 with new Category Y and Z added to the list. Weighted Category Scenario 1 Weight Factors Scenario 2 Weight Factors C1 Precision W1 5% W1 5% C2 Accuracy W2 10% W2 45% C3 Category Y W3 5% W3 5% C4 Ease of Use W4 15% W4 15% C5 Inspection Requirements W5 10% W5 10% C6 Cost W6 45% W6 10% C7 Category Z W7 10% W7 10% The end user can also create a totally new Scenario 3, which may have different weights to emphasize the desired influence of each category. As an example, Table D-3 presents a more balanced approach for cost and accuracy and put equal emphasis on both. The possible new categories are also considered in the list. Care should be taken so that the sum of all the weights must be unity. Table D-3. Categories and weight factors for a new Scenario 3. Weighted Category Scenario 3 Weight Factors C1 Precision W1 10% C2 Accuracy W2 25% C3 Category Y W3 5% C4 Ease of Use W4 15% C5 Inspection Requirements W5 10% C6 Cost W6 25%

D-12 C7 Category Z W7 10% GUIDELINES FOR SCORING METHODS BASED ON CRITERIA Each category defines a specific relationship between the usefulness of the NDE method and a potential need in the structural engineering condition assessment process. In this investigation, the range of scores vary from 0 to 10, with 0 representing the lowest possible score (least desirable), and 10 representing the highest possible score (most desirable). Each NDE method should be scored based on performance which is measured by the objective findings of actual experimental data obtained from laboratory testing or information provided by the manufacturer. In order to provide guidance for computing the individual scores (the values of parameter a in the matrix), the ranking definitions are provided in Table 5-1 – Table 5-4. Each category has descriptions of how the NDE method should be ranked (between 0 and 10). CLOSING REMARKS This section provides guidelines to the end user in adding new NDE technologies, new categories, and new weights to the various categories, to modify the decision matrix.