Guide for Utilization Measurement and Management of Fleet Equipment NCHRP 13-05 Ali Hajbabaie, Ph.D. Department of Civil and Environmental Engineering
Objectives Develop a guide for utilization measurement and management of fleet equipment to be used by State highway agencies Prediction Models Develop statistical models to predict utilization for each equipment asset Formulate utilization cost functions and constraints to satisfy demand for each asset Validate models Management Framework Develop optimization models to: - Minimize total costs - Determine optimal utilization values - Determine optimal fleet size - Determine optimal relocations 2
3
Initial Questions to be Answered What are utilization measurement metrics? Which factors influence equipment utilization? Which equipment to consider? What kind of data to collect? How much? How many years? 4
Literature review and Survey Published technical reports (NCHRP, state DOTs, etc.) Journal and conference (TRR, TRB, etc.) White papers Unpublished work Survey to all state DOTs 32 responded 5
Candidate Asset Utilization Metrics Annual mileage Annual engine hours Usage over the last 12 months Frequency of use 6
Candidate Factors Influencing Asset Utilization 33 factors were mentioned Downtime duration Asset age Cumulative utilization Normal equipment life Unskilled/skilled operators/mechanics Agency management policies/practices Expected workload Supervision level needed Design complexity Environmental impacts Cost 7
Equipment Types Dump Truck Pickup Truck Automobile Van Sport Utility Trailer Front Loader Grader Mechanical Street Sweeper Truck Air Street Sweeper Truck Sweepers / Scrubbers Riding Mower Truck Tractor Snow Removal Attachment Roller Drill Asphalt Distributor Attachments Man Lift Large Truck with Special Body 8
Data Availability and Willingness to Share Does the provided equipment list represent equipment classes that your agency use? 89.5% Yes, 10.5% No Data availability More than 50% said not available Rent reason, Annual rent hours, Rent costs, Annual idle hours, Annual interest charge, Annual taxes, Annual licensing fee, Annual storage cost, Predicted annual demand Less than 30% said available Asset capacity, Annual downtime hours, Seasonality, The frequency of usage Willingness to share: 19 states 9
Data Collection Sent the request to 19+ DOTs, received data from 8 California Florida Idaho Minnesota Montana North Carolina Utah Washington 10
Data Processing Data aggregated at the county level Outliers were removed Three data sets could be used We had enough data for eight equipment types Data allowed fitting models for annual mileage only 11
Fitting Predictive Utilization Measurement Models Model structures considered: linear, quadratic, power, logarithmic, and nonlinear Logarithmic regression yielded the best fit and the most intuitive prediction models The backward elimination approach used to fit the models Pearson correlation test for multicollinearity between independent variables Performed other tests to ensure all assumptions of regression analysis are met 12
Example: Dump Truck Utilization Prediction Model YY: Log (Annual Mileage) Estimate P-value Intercept 8.941 0.000 Annual Fuel Cost 0.0001 0.000 Log (Annual Downtime Hours) 0.073 0.001 In Service Age -0.047 0.000 Class 5-1.894 0.000 Class 7-3.221 0.000 Low Usage -0.408 0.000 Adjusted R-Square 0.98 Sample size 164 YY = 8.941 + 0.0001 xx FFFF + 0.073 log xx DH 0.047 xx ISA 1.894 xx C5 3.221 xx CC6 0.408 xx LU 13
Example: Dump Truck Utilization Prediction Model 40 Count 20 0-0.5 0.0 0.5 1.0 Residuals : predicted mileage observed mileage 14
Example: Dump Truck Utilization Prediction Model Predicted Annual Milea y 0 0.981x 10 R 2 0.98 9 8 7 6 6 7 8 9 10 Observed Annual Mileage 15
Equipment Type Utilization Prediction Model Dump Truck 8.941 + 0.0001 xx FFFF + 0.073 log xx DH 0.047 xx ISA 1.894 xx C5 3.221 xx CC6 0.408 xx LU 0.98 RR 22 Pickup Truck 8.117 + 0.00002 xx PPPP + 0.0002 xx FFFF + 0.047 log(xx DDDD ) + 0.00001xx FFFF 2 0.126 xx CC2 0.614 xx CC3 0.726 xx CC4 1.841 xx LLLL 0.98 Automobile 5.583 + 0.0003 xx SSSSSS + 0.561 log(xx FFFF ) 0.041 xx IIIIII 0.0003 xx FFFF 2.218 xx LLLL 0.98 Van Sport Utility Vehicle Grader Mechanical st. sweeper Large trucks 2.249 + 0.00001 xx PPPP + 0.707 log xx FFFF + 0.242 log xx DDDD + 0.024 xx IIIIII 0.061 log xx FFFF 0.350 xx CC2 0.862 xx LLLL 0.84 4.122 + 0.0003 xx SSSSSS 0.0001 xx UUUUUU + 0.468 log xx FFFF + 0.289 log xx DDDD + 0.066 log xx FFFF 0.883 xx LLLL 0.93 0.271 0.0001 xx SSSSSS + 0.692 log(xx FFFF ) 0.001 xx IIIIII 2 + 0.048 log (xx FFFF ) 2.309 xx LLLL 0.96 4.588 + 0.000003 xx PPPP + 0.00001 (xx SSSSSS + xx UUUUUU ) + 0.432 log(xx FFFF ) 0.253 log(xx IIIIII ) + 0.035 xx FFFF 2.320 xx CC3 0.65 5.279 0.00001 xx PPPP + 0.0001 xx SSSSSS + 0.00002 xx UUUUUU + 0.645 xx FFFF 16 0.0001 xx DDDD 0.090 xx IIIIII + 0.001 xx FFFF 0.575 xx CC11 0.263 xx CC5 2.377 xx 0.98 LLLL
Utilization Management 17
Utilization Management Objective: minimize total cost (fixed and operational) Control Decisions County/region/division level in a year fleet size of a specific asset number of assets to be purchased number of assets to be salvaged number of assets to be relocated within a state average utilization level 18
19
20
21
22
23
Relocation Example 24
Guide for Utilization Measurement and Management of Fleet Equipment NCHRP 13-05 Ali Hajbabaie, Ph.D. Department of Civil and Environmental Engineering
Research Team Ali Hajbabaie Leila Hajibabai Wei (David) Fan Mehrdad Tajalli Amir Mirheli WSU SBU UNCC WSU SBU Xianming Shi Mike Moser Shaowei Wang Miao Yu WSU MyFleetDept ITSNode UNCC 26