A s s e t M a n a g e m e n t

Transcription

1 Integrated Engineering Design Management (IEDM) Master Program, Faculty of Engineering, Cairo University IEDM 62 Infrastructure and Asset Management Building Asset Management Ahmed Elhakeem Associate Professor, AASTMT PhD. Constr. Engr. & Mgt. 26 University of Waterloo, Canada MASc. Constr. Engr. & Mgt. 22 University of Waterloo, Canada BSc. Civil Engineering 996 Helwan University, Cairo, Egypt By Ahmed Elhakeem, PhD Associate Professor, Construction & Building Engineering Dept. - AASTMT 22 nd of September 25 nda A s s e t M a n a g e m e n t Infrastructure Asset Management Building Asset Management Asset Management Asset & Management THE Problem THE Challenges IMS Generalization/ Customization Building Asset Management Main Facets Research, Ideas, Models

2 Asset & Management b Assets and Organizations Assets are valuable items (lands, houses, cars, ) owned by an Organization Management Decide for Actions related to owned Assets (sell, repair, renovate, ) Country E G Y P T Ministry Health Education. Governorate Giza Cairo. Region North Nsr City.. Asset (e.g. School) School (y) School (x).. Asset Management Asset Management Asset management is an Organization level management. Organization can be small or very large, such as ministries and governments. Organization tries to identify its valuable Assets and plan to keep them in good shape and performance to serve the mission they created for. Asset Management tries to achieve the best performance for the organization throughout proper management of its assets. The principles to manage any type of assets are the same (same Management System). The management process is very challenging as all assets compete to receive Available/Limited Budgets to enhance their performance and hence the organization performance. An asset manager should prepare a Priority Plan based on Current and Future Conditions of assets and after examining the consequence of possible Actions (Repair/No-Repair decisions) on the Overall Performance. 2

3 The Problem Simply: The Crisis of Infrastructure Assets From Early 2 to 25 (USA NOW) US Assets!!, What about Egyptian Assets The Challenge Simply: Which, When & What In order to manage and distribute limited budgets among competing assets, it is important to answer simultaneously three interrelated questions. Within a certain planning horizon (e.g. a 5- year plan), and a predefined budget/targeted performance, Which assets to consider in the plan; When to act (now, next year, the year after, ); & What kind of act to perform (minor, moderate, ormajor act) The three questions are interrelated for example, the more major act to do (What) the less assets you can consider (Which) due to limited budgets. The Challenge Simply: Which, When & What IMS Infrastructure Management System I = B,B,S, or W I M S Asset Management (Infrastructure Asset Management) IMS Repair, Rehabilitation and Replacement Decisions Operation stage (NOT construction stage) Organization level (NOT project level) Manageable units (elements or components of the building represent organizations assets) Short Tactical Long 3

4 Condition 9/23/25 IMS Generalization/Customization Long Tactical Limited Budget to Maintain Safety and Operability Which components to repair When to repair What kind of repair Type of Assets Generalizat ion/ Customization Short Bidding, Planning, Scheduling, Resource Mgt., Building Asset Management Buildings vs. Others Building Management System Facets of BMS e.g., certain window Which, When & What Prioritization and Fund Allocation Network-DSS Excellent Now Planning Horizon $ Future Which When What Inspection Assessment Prediction Repair Options LC Analysis New Now x Unacceptable level time BUILNGS : Diversity, Interrelated, Technicality Condition Assessment Deterioration Prediction Optional Repair Strategies 4

5 Facets of BMS Facets of BMS $$$ Improvement Condition Assessment Deterioration Prediction Optional Repair Strategies Condition Assessment Decisions Visual Inspection Photographic / Optical Non-Destructive Evaluation Smart Sensors Time consuming, subjective, & costly. Facets of BMS Facets of BMS $$$ Improvement Deterioration Prediction Optional Repair Strategies Lack of historical data Diverse components Multiple-component interactions Static, Inaccurate, predicts condition not causal problems. Improper definition of repair, and the cost / improvement calculations. Minor, Moderate,,Replacement Undefined meaning, cost, & impact. Example light repair - - Cost = 28% of New, & - Improves condition from D to C. 5

6 Facets of BMS $$$ Improvement How to support: Better Condition Assessment Condition Assessment Deterioration Prediction Optional Repair Strategies Deterioration/Performance Perdition Action Modeling and Representation (e.g., Time consuming, subjective, & costly. Static, Inaccurate, predicts condition not causal problems. Improper definition of repair, and the cost / improvement calculations. Repair Action) Network Optimization Condition Assessment How to improve RECAPP 22 Condition Assessment Select Component Deficiency list for the selected component USER INPUT of Distre ss Subjective and Needs Experts 6

7 Visual Guidance System Sample Interactive Visual Database Condition Assessment sample Condition Assessment survey Aggregated reports for windows Aggregated Reports Types Deficiencies Symptoms Possible Deficiencies Weights for Aluminum windows (%) Broken / Detached Seals Frame Related Deficiencies 25 5 Glazing Problems 25 Hardware Problems 2 5 Finishing Problems 7 25 j d i W ij S ij S = % Severities Visual Guidance 7

8 Investigation Investigation System : User Interface Picture with = 2... Scroll to a comparable picture. Picture with = 8 Comments Pictures with =, 2 are stored. The component is matched to the closest picture; e.g., picture for = 6. Fast, simple, & accurate. Requires 5 to pictures per component (based on scale). Direct condition assessment without details about deficiencies. System 2: User Interface Please select the severity of each defect Calculations: = 63 i.e.,. Condition = Fair 2. Check: This component should match the picture n. below with a Fair condition. Good Fair Poor Critical Comments Four pictures are stored for each component at different condition states. The pictures confirm the appropriateness of the user specified severities. Requires 4 pictures per component. Provides detailed deficiency evaluation. More Accurate. Requires personal judgment to select the severities. Investigation Condition Assessment Visual Guidance System 3: User Interface Comments Pls. scroll each defect till a comparable picture is shown Picture changes with user selection of each deficiency and each severity level. 5% n. Scroll to comparable picture. Picture for def. 2 & severity 5% Calculations: = 63 i.e., Condition = Fair Requires many pictures for each component (about 25). Provides accurate and detailed deficiency evaluation. Fast and most accurate. Personal judgment is almost eliminated. 3 8

9 Implementation Program Other Research Improved Condition Assessment with 3D-CAD Technology 34 Traditional Sequential Proposed Parallel Inspection Experts on the go Time Consuming Expensive Subjective Head Office Caretakers Fast Cheap Head Office

10 Portable Field Inspection Proposed Office Assessment HEAD OFFICE Inspection Pictures (from Site) 3D Designed Actual View of Real Building ON SITE Connected Digital Camera School 3D Walkthroughs Superimpose The system directly attaches pictures of each inspected component to its respective CAD objects. TO OFFICE Tablet PC Visual Guidance (Less Subjective) The building comes to the experts, not the opposite Office Assessment VISUAL GUIDANCE SYSTEM IN OFFICE EXPERT Visual Guide MA TCH Please Enter 6 Severity Level: 6 3D-CAD Animation of Building Pictures VISUAL GUIDANCE Pictorial database of similar components at different Condition states. Accurate Severity Level 39

11

12 2 Deterioration Model Is it Custom or General Deterioration Prediction All LG SNT split AC units has a five-year service life Meaning, All LG SNT split AC units deteriorate similarly Generalization Deterioration Model Argument Deterioration Model Argument No Reason What if they are in the same country, city, & building What if they are in the same building, direction, altitude, use,. No Reason N No Reason Generalization Customization 2

13 Deterioration Index () Deterioration Index () 9/23/25 Average Deterioration Curve Stage: Average Deterioration Behavior of similar instances (e.g., Aluminum windows) Field inspection data of similar instances =33 =33 D D2 3 4 W= W2=3 D3 2 W3=4 5 D4 W4=2 Now (year ) Future (year 2) (year 5) Dynamic model Lack of data Other parameters Interaction 6 D D2 3 4 W= W2=3 D3 2 W3=4 5 D4 W4=2 Now (year ) Future (year 2) (year 5) Dynamic model Lack of data Other parameters Interaction 6 Min. error Stage2: Customization for a specific instance (e.g., Aluminum window ) Customized Curve for the selected instance Future Prediction =33 =33 =39. D 3 W= D2 4 W2=3 D3 2 W3=4 5 D4 W4=2 Now (year ) Future (year 2) (year 5) Dynamic model Lack of data Other parameters Interaction 6 D 3 W= D2 4 W2=3 D3 2 W3=4 5 D4 W4=2 Now (year ) Future (year 2) (year 5) Dynamic model Lack of data Other parameters Interaction 6 3

14 Deterioration Index () Deterioration Index () 9/23/25 Min. error Average Deterioration Markov Predicted Curve s 3++ 3/( ) Stage: Stage2: Stage3: Average Customization Future defects 4++ Deterioration 4/( ) for a the specific Behavior Average instance of similar s Incremental increase for the original instances (e.g., (i.e., Aluminum (e.g., Field Customized inspection Aluminum 2++ Curve data 2/( ) for of severities window using their ) windows) relative values similar the selected instances 5++ instance 5/( ) Future Prediction Sum of Errors Objective Function Min. error Average Deterioration Curve 3++ 3/( ) Stage: Stage2: Stage3: Average Customization Future 4++ Deterioration 4/( ) a Behavior Incremental defects increase for the for specific the original instance of similar instances (e.g., (i.e., Aluminum (e.g., Field Customized inspection Aluminum 2++ Curve data 2/( ) for of severities window using their ) windows) relative values similar the selected instances 5++ instance 5/( ) Future Prediction =33 =39. Abs. Error =33 =39. D D2 3 4 W= W2=3 D3 2 W3=4 5 D4 W4=2 Now (year ) Average Markov Future (year 2) (year 5) Dynamic model Lack of data Other parameters Interaction D D2 3 4 W= W2=3 D3 2 W3=4 5 D4 W4=2 Now (year ) Future (year 2) (year 5) Dynamic model Lack of data Other parameters Interaction Stage3: Future defects for the specific instance (i.e., Aluminum window ) Incremental increase for the original severities using their relative values 3++ 3/( ) 4++ 4/( ) 2++ 2/( ) 5++ 5/( ) =33 =39. =33 =39. D 3 W= D2 4 W2=3 D3 2 W3=4 5 D4 W4=2 Now (year ) Future (year 2) (year 5) Dynamic model Lack of data Other parameters Interaction 6 D 3 W= D2 4 W2=3 D3 2 W3=4 5 D4 W4=2 Now (year ) Future (year 2) (year 5) Dynamic model Lack of data Other parameters Interaction 6 4

15 ANN Prediction Model ANN (training process) INPUTS Hidden Nodes OUTPUT AGE L.O.S Floor CI function orientation ANN(prediction of deteriorated CI) The age is increased by year. 3 Year + Repair Representation Current & Future Severities Year D D2 D3 D W = W2 =3 W3 =4 W4 =2 Best Repair Scenario for each year in the planning horizon 5

16 Repair Scenarios Repair Scenarios Defects Repair scenario Severities at year k (e.g., k=3) Severities (After repair) Yearly Analysis D D 2 D 3 D W = W 2 =3 W 3 =4 W 4 =2 D D 2 D 3 D Current & Future Severities Year D D2 D3 D W = W2 =3 W3 =4 W4 =2 Cost AR3 = 25% 3.7 Best Repair Scenario for each year in the Repair Action Improvement planning horizon Cost BR3 = 44.9 AR3 = 3.7 Repair Action Improvement Cost Repair Scenarios Repair Scenarios Defects Repair scenario Defects Repair scenario Severities at year k (e.g., k=3) Repair Cost Severities at year k (e.g., k=3) 2 2 d D D 2 D 3 D W = W 2 =3 W 3 =4 W 4 =2 D D 2 D 3 D 4.25*2 D D 2 D 3 D W = W 2 =3 W 3 =4 W 4 =2 D D 2 D 3 D 4 Which Scenario BR3 = 44.9 Cost = 25% BR3 = 44.9 Repair Action Improvement Cost Repair Action Improvement Cost 6

17 Repair Model Best Repair Scenario Repair Model Best Repair Scenario Which Scenario 2 d Cost After-Repair Condition Variables New severities af ter applying the RS Least Acceptable Least Constraint % RC d i.25w i RS i ( k ) AR d i W S i ik ( RS ) Best scenario if repair in year... Sequential Optimizations Best scenario if repair in year 5 i Objective Function (minimize) SOAP Sequential Optimization Analysis SOAP Sequential Optimization Analysis If No-Repair Cost: $ Repair: [ ] 5 s: Shown If repair in year Cost: $26k Repair: [ ] 6 s: Shown If repair in year 2 Cost: $3k Repair: [ ] 6 s: Shown If repair in year 3 Cost: $39k Repair: [ ] 6 s: Shown 7

18 SOAP Sequential Optimization Analysis Network-Level Optimization If repair in year 4 If repair in year 5 Cost: $45k Repair: [ ] 6 s: Shown Cost: $6k Repair: [ ] 6 s: Shown 6 Optimized Repair Options: No repair.... Repair at year.. Repair at year 5 Repair year for each instance in the network 2 Network-Level Optimization Network-Level Optimization Impact of Repair Impact of Repair k= k= k= Now i= i= i= Time No- repair option Repair in year 5 Repair in year 3 EP () = 6 EP (5) = 5.67 EP (3) = 35 (worst) (best) Least-Cost Repair Scenario Planning horizon Cost = $ Cost = $5K Cost = $3K Expected Performance (the average s) 8

19 Network-Level Optimization Network-Level Optimization All instances (Network) 2 3 j Variables Possible repair year Cost Im plication 5 S $ S $5 Year budget Year 5 budget Condition Implication Expected Performance EP EP 2 EP 3 EP j x Relative Importance RIF RIF2 RIF3 RIFj Overall Network Condition Objective Function Min ( $ C ( k ) RIFj EPj j N ) Min j jk RIFj ( k ) k j $ j $ IRC j B k if j is the Network ( N) repaired at year k Constraints Network-Level Optimization Objective Function Constraints Thank you Decision to repair at year 3 Binary Variables 9