Practical Tool for the Design and Cost Optimization of Permeable Pavements in the Planning Stage of Stormwater Management

Abstract

To address issues created by increased stormwater runoff, there is growing interest in sustainable tools such as permeable pavement (PP), which is a green infrastructure (GI) practice. This paper outlines the development and application of a design tool called the “cost optimization tool for permeable pavements (COTPP).” This tool aims to help states, municipalities, and consultants design PPs using reliable hydrological and structural pavement design methods. The COTPP uses a cost optimization algorithm similar to a heuristic search that helps identify which combination of PPs, conventional pavement, and other GIs provides the most economic value. A sensitivity analysis was conducted using the tool, which consisted of designing the three main types of PP using varying design inputs and comparing the COTPP’s results to results from an existing design tool or guidelines. The average percent differences between the COTPP’s results and that of other methods were below 10% for the structural designs and around 15% for the hydrological designs. The differences were caused by the use of a safety factor in the hydrological design and other adjustments to design methods. It was concluded that the tool is reliable and can be confidently used by designers. The tool was then used for a representative case study site (PP parking lot) in which the actual design and cost estimates of on-site components were known. The COTPP was able to identify other design alternatives that helped reduce the actual construction cost of the existing parking lot from $305,918 to as low as $238,790.

Keywords

pervious pavements hydrology and hydraulics and stormwater stormwater facilities stormwater management sustainable and resilient pavements

One of the most common causes of increased runoff volume and pollutant loads is urbanization. Urban stormwater runoff was ranked among the top three factors causing pollution of reservoirs, lakes, and ponds ( 1 ). This led to the development of stormwater management programs by numerous state agencies to replace traditional stormwater best management practices (BMPs) with runoff-reduction practices such as permeable pavements (PPs) and reduce stormwater runoff affecting transportation facilities. PPs are categorized into three main types: pervious concrete (PC), porous asphalt (PA), and permeable interlocking concrete pavers (PICP) (Figure 1). PPs are porous materials that let water permeate the surface and be stored temporarily in aggregate reservoir layers before infiltration into the subgrade (Figure 2) or discharge through an underdrain. PP systems can treat runoff from both their surface as well as small adjacent impervious areas while serving as a structural support for low traffic load conditions ( 2 ). The effectiveness of a PP system is measured by its ability to handle traffic loads, store stormwater, reduce pollutants, and keep an acceptable subgrade permeability for stormwater infiltration ( 3 ).

A survey study in 2017 demonstrated that high cost is the second most important issue delaying the wide implementation of PPs ( 4 ). Fortunately, PPs can be designed in various ways with other green infrastructures (GIs) to reduce costs and improve performance. Some existing tools such as the American Concrete Pavement Association’s PerviousPave software help design one type of PP and other tools are available to calculate costs. However, there is not an available tool that helps design simultaneously the three main types of PP and achieve cost optimization by combining PPs with other GIs or conventional pavement systems. Consequently, there is a need for a practical optimization tool to lower costs and improve the design of PPs, leading to an increase in use.

Research Objectives and Scope

The purpose of this research study is to develop a practical tool to design and optimize the cost of PP systems for stormwater management. To achieve this goal, the research objectives are as follows:

Create an easily accessible, user-friendly tool that supports the design of the three main types of PP (PC, PA, and PICP).

Develop a cost optimization algorithm that helps identify which combination of PPs with conventional pavements, bioretention, and/or infiltration trench can be used to minimize costs.

To obtain successful results, the most recent recommended design methods were chosen for inclusion in the cost optimization tool for permeable pavements (COTPP). The design of underdrains for stormwater transfer between GIs is not considered within the scope of this study, which reflects the capabilities of existing design tools. The components considered in the cost optimization model of the tool are PPs, conventional non-pervious pavements, and two other GI practices: bioretention and infiltration trench. The Microsoft Excel software was selected to host the tool because of its popularity and availability to most design engineers and decision makers. The tool was developed with small pavement project applications in mind, such as car parking and entrance areas, drive aisles, and potentially low-volume roadways.

Methodology

The design of the PP cross sections follows two stages: 1) the hydrological design and 2) the structural pavement design. Excel functions were used to write the PP cost optimization algorithm in the tool for transparency assurance and effortless future addition of more GIs to the tool. Furthermore, the tool’s user experience was improved through visual basic applications (VBA) such as macro color codes assigned to control buttons. The tool’s operation process is presented in the flow chart diagram (Figure 3).

Hydrological Design Method

The purpose of the hydrological design is the determination of the reservoir thickness needed for temporary stormwater storage ( 3 ). Since the stormwater ideally infiltrates the underlying soil, the subgrade’s permeability is a crucial element in the hydrological design process. Moreover, the storage capacity of the reservoir layer or base/subbase layer is calculated based on the anticipated runoff treatment volume. Many organizations have recommended hydrological methods for PP, but none has yet been standardized. The industry’s most popular methods are the curve number method, the rational method, the interlocking Concrete Pavement Institute (ICPI) method, and the Los Angeles County method ( 3 ). Fortunately, a single hydrological method can be used to design all types of PP. The ICPI method was selected for the COTPP because it is the most comprehensive method ( 3 ). This method takes into account many important factors as shown in Equation 1:

d_{p} = \frac{{(d_{c} \times R) + P - (\frac{i}{2} \times t_{f})}}{V_{r}}

(1)

where d_p = depth of reservoir layer (ft); d_c = (T_v/A_c) = runoff depth from the contributing impervious area (ft); T_v = treatment volume of water (ft³); R = A_c/A_p = ratio of the impervious area (A_c) to the pervious area (A_p); P = rainfall depth (ft); i = infiltration rate of subgrade (ft/day) with 0.5 in./h as the recommended minimum ( 2 , 5 ); and t_f = time to fill the reservoir layer (days)—typically 2 h or 0.08333 days; and V_r = void ratio of reservoir layer—typically 40% or 0.4. Moreover, a safety factor of 2 is recommended to reduce the infiltration rate ( 2 ).

The steps of the ICPI method are provided in the City of Birmingham manual and the ICPI manual ( 2 , 6 ). The adapted detailed version of the steps used in the COTPP are provided by Biessan ( 7 ).

Structural Design Methods

The structural design method, contrary to the hydrological design, does not have a single accepted method for all types of PP. Two main structural design methods were selected for inclusion in the COTPP.

For the design of PC, the American Concrete Pavement Association (ACPA) method, which relies on a Mechanistic-Empirical approach, was chosen. The ACPA developed a software program called PerviousPave, which designs PC pavements using this method ( 8 ). The ACPA method relies on Westergaard’s model of a plate on a Winkler foundation, which is by origin a conventional rigid pavement model ( 3 ). The applicability of this model on PC pavements was proven by previous research with results from falling weight deflectometer testing on several PC pavements ( 9 , 10 ). The failure criterion of the ACPA method for PC is the fatigue model, which helps establish the number of allowable load repetitions for a given stress ratio as shown in Equation 2 ( 10 , 11 ). The fatigue damage for each axle type is then obtained using Miner’s damage hypothesis equation shown in Equation 3 ( 12 ):

Log N_{f} = {[\frac{- S R^{- 10.24} \log (1 - P)}{0.0112}]}^{0.217}

(2)

FD = \frac{N}{N_{f}}

(3)

where N_f = allowable load applications; SR = stress ratio (%); P = probability of failure (%); FD = fatigue damage per axle type; and N = number of load applications obtained from user traffic data.

Rodden & Smith ( 12 ) provided the steps of this method. Key factors such as traffic data for PPs, formulas for growth factor, and formulas for composite modulus of subgrade/subbase reaction were not provided and had to be found from other sources documented by Biessan ( 7 ).

The design of PA pavements and PICP followed the frequently used American Association of State Highway and Transportation Officials (AASHTO) 1993 method ( 3 ). This method relies on algorithms that resulted from the AASHO road test in the late 1950s and 1960s. Even though the AASHTO 1993 method may have a degree of uncertainty because the AASHO road test conditions are not necessarily representative of current conditions, it is the recommended method by associations such as the National Asphalt Pavement Association (NAPA). For the structural design of PA, the steps and assumptions for this method were provided in the NAPA design guidelines ( 13 ). For the structural design of PICP, recommendations are provided in the guidelines published by the Interlocking Concrete Pavement Institute ( 14 ). The work of Biessan ( 7 ) provides the adapted detailed and complete versions of those steps used in the COTPP.

Cost Optimization Algorithm

The cost optimization algorithm developed in the COTPP to optimize costs of PPs is a search strategy similar to a heuristic search where the expected result (cost) is optimized through iterative designs based on given functions and constraints. The designs and costs of PPs cannot be truly optimized without exploring combinations of PPs with other forms of infrastructure. PPs are typically designed to support specific traffic loadings and treat a specific amount of stormwater. Reducing the reservoir layer depth helps reduce pavement costs, but the remaining amount of stormwater will need be stored and treated in another reservoir such as an infiltration trench. Similarly, reducing the area that a PP occupies can also lower costs since PPs are expensive to construct. However, the cheaper practice, such as conventional hot-mix-asphalt (HMA), used to occupy the remaining area would be impervious and generate runoff. Therefore, the method developed for the COTPP relies on a combination of GIs, combining PPs with other GIs and/or conventional pavements and selecting the most cost-effective combination. The components considered for each category are as follows:

For PPs: PC, PA, and PICP

For conventional pavements: HMA, Portland cement concrete (PCC), and interlocking concrete pavers (ICP).

For other GIs: infiltration trench and bioretention

Three options representing three combinations (Figure 4) were created to achieve the goal of selecting the most economic and appropriate design combination.

An algorithm was developed for each combination option in the COTPP. The algorithm allows the user to input certain constraints based on known parameters of the project site and the components being combined (e.g., main driveway needs to be dense graded asphalt because it is the location of the tractor trailer loading zone). Once the inputs are entered, the dimensions of each component are calculated in the “Optimization” tab of the tool as well as the construction costs to determine the total cost of each combination option ( 7 ). Next, the final construction costs of all three options are displayed, and the user can select the option that is most cost effective. The COTPP only considers construction costs because maintenance is done yearly and is the same for all types of PP. However, it is recommended that the designer assess the maintenance costs of the other practices combined with PPs after running the tool and before making a final decision. The works of Rehan et al. ( 15 ) and Olson et al. ( 16 ) provide maintenance and rehabilitation costs for most components mentioned in this study. The calculation process of the algorithm is further explained below.

Option A: Combination of PP + Conventional Pavement

The goal of this combination is to reduce the PP footprint and replace the remaining area with a conventional pavement. This will cause the reservoir thickness to increase to meet the minimum storage capacity requirement for expected stormwater runoff. The execution of the algorithm for Option A is as follows.

The design inputs required for the calculations in the algorithm are:

Total treatment volume of water (Tv)

The preferred type of PP: PC, PA, or PICP

Type of conventional pavement: HMA, PCC, or ICP

Construction costs of all components (in the case these values are unknown, default values are provided based on historical data ( 15 – 19 ).

The constraints that frame the algorithm calculations are:

Minimum depth of the PP reservoir layer

Range of conventional pavement area (%) desired as a function of the total surface area.

Once the design inputs and constraints are entered in the “User Interface” and “Detailed Inputs” Tabs of the tool, the algorithm calculations are done automatically in the “Optimization” Tab. The calculation process is as follows:

1. 10,000 random numbers are generated within the range of the conventional pavement area that was set as a constraint. Each random number represents an impervious conventional pavement area ( $A_{im p_{i}}$ ). Any other contributing impervious area (A_cont) such as rooftop, hardscape and so forth is added to every $A_{im p_{i}}$ to obtain the total impervious contributing area ( $A_{c_{i}}$ ) as shown in Equation 4:

A_{c_{i}} = A_{im p_{i}} + A_{cont}

(4)

2. For every $A_{im p_{i}}$ generated, a new pervious area for PP is calculated using Equation 5:

A_{p p_{i}} = A_{total} - A_{im p_{i}}

(5)

where $A_{p p_{i}}$ = new pervious area of PP (ft²); and $A_{total}$ = total area of the site (ft²).

3. For each $A_{p p_{i}}$ and $A_{c_{i}}$ , a new reservoir layer is designed for PP following the ICPI hydrological method.

4. The final construction cost for each of the 10,000 trials is calculated using all construction cost data provided for the Option A components.

5. The lowest final cost of the 10,000 trials and its corresponding design are displayed on the “User Interface” Tab as the most economical design for Option A.

Option B: Combination of PP + Other GI

The objective of this combination is to reduce the thickness of the PP reservoir layer by constructing another GI like an infiltration trench or a bioretention basin that may contribute to a total project cost reduction. In this combination, the PP area remains constant and the entire site is used to construct the PP and the other GI. It is important to note that the COTPP only designs and provides the storage volume or media volume of bioretention and infiltration trench. For detailed design of those two other GIs, the design manuals provided by the United States Environmental Protection Agency should be consulted ( 20 ). The term “other GI” will henceforth be used throughout the paper to represent infiltration trench or bioretention basin.

The algorithm for Option B is computed as follows:

The design inputs required for the calculations in the algorithm are:

Total treatment volume of water (Tv)

The preferred type of PP: PC, PA, or PICP

The preferred type of other GI: infiltration trench or bioretention

Construction costs of all components

The constraints that frame the algorithm calculations are:

Maximum area available on site to construct other GI

Range of area desired to use for the other GI (%) as a function of the maximum available area to construct the other GI.

Range of storage capacity for the other GI (%) as a function of the total treatment volume of water (Tv). In other words, how much of the expected runoff should the other GI store and treat?

Minimum depth of the other GI

The design inputs and constraints are entered in the “User Interface” and “Detailed Inputs” tabs of the tool. The algorithm calculations for Option B are computed in the “Optimization” tab. The calculation process is as follows:

1. 10,000 random numbers are generated within the range of storage capacity for other GI set as a constraint. Each random number represents the new GI treatment volume ( $T v_{G I_{i}}$ ).

2. For every $T v_{G I_{i}}$ generated, a new GI storage volume is calculated using Equation 6:

V_{S_{i}} = \frac{T v_{G I_{i}}}{0.4}

(6)

where $V_{S_{i}}$ = storage volume for other GI (ft³); $T v_{G I_{i}}$ = treatment volume for other GI (ft³); and 0.4 represents the typical storage void ratio (40%) of both infiltration trench and bioretention ( 2 ).

3. 10,000 random numbers are generated within the range of area desired for other GI set as a constraint. Each random number represents the new GI surface area ( $A_{G I_{i}}$ ).

4. The depth of the other GI for each $A_{G I_{i}}$ is obtained using Equation 7:

d_{G I_{i}} = \frac{V_{S_{i}}}{A_{G I_{i}}} \times 12

(7)

where $d_{G I_{i}}$ = depth of the other GI (in); $V_{S_{i}}$ = storage volume of the other GI (ft³); $A_{G I_{i}}$ = area of other GI (ft²).

5. The new required treatment volume of stormwater stored by the PP’s reservoir is calculated using Equation 8:

T v_{P P_{i}} = Tv - T v_{G I_{i}}

(8)

where $T v_{P P_{i}}$ = treatment volume of water for PP (ft³); $T v_{G I_{i}}$ = treatment volume of water for Other GI (ft³); $Tv$ = total treatment volume of water (ft³) where $Tv$ is obtained through Equation 9:

Tv = R_{V} \times (A_{p} + A_{c}) \times P

(9)

where $R_{V}$ = volumetric runoff reduction coefficient ( $R_{V}$ = 0.95 for impervious areas); $A_{p}$ = total pervious area for PP (ft²); $A_{c}$ = total impervious contributing area (ft²); and P = rainfall depth (ft).

6. Using the results from previous steps, a new reservoir thickness for PP is calculated for each of the 10,000 trials using Equation 10:

d_{p_{i}} = \frac{T v_{P P_{i}}}{0.4 \times A_{p}}

(10)

where $d_{p_{i}}$ = depth of PP reservoir (in); $T v_{P P_{i}}$ = water treatment volume for PP (ft³); $A_{p}$ = total pervious area for PP (ft²); and 0.4 representing PP reservoir void ratio (40%) ( 2 ).

7. The final construction cost for each of the 10,000 trials is calculated using all construction cost data entered in the tool for Option B components.

8. The trial with the lowest total cost is selected for Option B and displayed in the “User Interface” tab.

Option C: Combination of PP + Conventional Pavement + Other GI

The purpose of this combination joins the objectives of Option A and Option B together (Option C = Option A + Option B). Therefore, there is no need to generate new random numbers. The generated random numbers and results from Options A and B can be used to design the new PP in Option C. The algorithm for Option C is explained as follows:

1. There are still 10,000 trials. For each trial, the new pervious area of PP ( $A_{p p_{i}}$ ) obtained from Option A and the treatment volume of water for PP ( $T v_{P P_{i}}$ ) obtained from Option B are used to design a new reservoir thickness ( $d_{p}$ ) for PP. Equation 11 is used:

d_{p_{i}} = \frac{T v_{P P_{i}}}{0.4 \times A_{p p_{i}}}

(11)

2. The final construction cost for each trial is then calculated using all construction cost data entered in the tool for Option C components.

3. The lowest cost of the 10,000 trials and its corresponding design are displayed on the “User Interface” Tab as the most economical design for Option C.

At the end of the overall cost optimization algorithm, the final designs and construction costs of the three Options are displayed on the “User Interface” tab to allow the user to compare and select the most economical design option for their project. The user must ensure that the depths of the PP reservoir and other GI respect their local groundwater protection requirements. For instance, in Alabama, a minimum of 2 ft is recommended between the reservoir and the water table ( 5 ). It is important to note that 10,000 was chosen as the total number of trials used in the algorithm based on results from a sensitivity analysis (Figure 5). It was discovered that the average final costs for all three options (particularly option C) were usually consistent using random numbers greater or equal to 10,000.

COTPP—Tool Description and Interface

The following subsections describe the Excel spreadsheet-based tool which includes six tabs or worksheets.

User Interface Tab

The main tab of the tool allows the user to enter all the necessary parameters needed for the design and cost optimization of PPs as well as to view the results. An example of the design output is shown where the dimensions and final construction costs of the three types of PP are displayed (Figure 6).

Another example of output for the cost optimization is shown where the dimensions and final construction costs of each optimization option are shown for the user to decide which one is the most economical and appropriate (Figure 7). The user can change or adjust design constraints in the inputs section of the User Interface tab until they obtain satisfying results for their project.

Detailed Inputs Tab

This is the second tab of the tool where the user can change default input values to adapt to a specific project. It also allows the user to enter construction cost data for the cost optimization process. It is the last tab where changes should be made to design inputs. The remaining tabs are where the calculations are computed and the user should be careful to understand the design and cost optimization algorithm before making any changes.

Design Calculations Tabs

The “Concrete,”“Asphalt,” and “PICP” tabs compute structural and hydrological designs of PC, PA, and PICP, respectively. The last tab is the “Optimization Tab” where the cost optimization algorithm is implemented. The design and cost calculations are computed for each combination option and the results are displayed on the User Interface Tab.

Sensitivity Analysis

The goal of the sensitivity analysis was to evaluate the reliability of the COTPP’s structural and hydrological design results and validate the tool’s effectiveness in the design of PPs. To achieve that goal, comparisons of the COPP’s results to the results from the existing software (PerviousPave) and design examples published in the NAPA and ICPI guidelines were performed.

Design Scenarios

The design scenarios executed in the COTPP for comparison are explained below. The general inputs and assumptions (Table 1) used in the structural design scenarios differ for each PP type because they depend on the parameters used in the existing software and published design examples.

Table 1.

General inputs for structural design scenarios

General inputs	PC	PA	PICP
In user interface tab
pervious area (ft²)	16,000	na	na
contributing impervious area (ft²)	32,000	na	na
projected application	Category B	na	na
design life (years)	20	na	na
design reliability (%)	80	75	80
annual truck traffic growth (%)	2	na	na
subbase layer elastic modulus (psi)	15,000	20,000	15,000
In detailed inputs tab
flexural strength of concrete (psi)	400	na	na
axle load edge support?	No	na	na
PC minimum reservoir depth (in.)	6	na	na
standard deviation	na	0.45	0.45
initial serviceability	na	5	4.2
terminal serviceability	na	2.5	2.5
PA layer coefficient	na	0.40	na
subbase layer coefficient	na	0.10	0.06
applied tire pressure (psi)	na	100	na
diameter of circular load (in.)	na	5.35	na
PICP layer coefficient	na	na	0.3
base layer coefficient	na	na	0.09

Note: PC = pervious concrete; PA = porous asphalt; PICP = permeable interlocking concrete pavers; psi = pounds per square inch; na = not applicable.

Structural Design Scenario for PC

A structural design scenario was simulated in both the COTPP and the software PerviousPave for comparison. This scenario’s objective was to calculate the thickness of the PC surface layer for different traffic levels and soil moduli. The average daily truck traffic (ADTT) and the California bearing ratio (CBR) of the subgrade were varied from 2 to 10 and 4% to 15% (increments of 1), respectively.

Structural Design Scenarios for PA

The NAPA guidelines provided result tables for PA design examples using the AASHTO 1993 method ( 13 ). The results are thicknesses of PA surface layer as a function of different subgrade modulus values and reservoir thicknesses. Each table corresponded to a scenario with a specific traffic level and many scenarios were provided. However, only three scenarios with the lowest design traffic levels (ESALs) were chosen to repeat in the COTPP for results’ comparisons because PPs are typically designed for low traffic areas (<1 million ESALs [ 6 ]). The traffic levels chosen for the design scenarios are 27,000 ESALs, 110,000 ESALs, and 820,000 ESALs. For each scenario, the resilient modulus of subgrade and the depth of reservoir layer were varied from 2,000 pounds per square inch (psi) to 10,000 psi and from 6 in. to 30 in., respectively.

Structural Design Scenario for PICP

The ICPI guidelines provided a table of recommended minimum PICP base and subbase thicknesses as a function of different traffic levels (ESALs) and subgrade strengths (CBR) ( 14 ). The traffic level and CBR were varied from 50,000 ESALs to 1 million ESALs and 4 to 10, respectively. That same design scenario was simulated in the COTPP to compare with ICPI guidelines results.

Hydrological Design Scenario for PC, PA, and PICP

All three PP types have the same hydrological design method, which is also used in PerviousPave. Therefore, a design scenario was simulated in both the COTPP and the software PerviousPave to compare the results. The scenario’s objective was to determine the depth of the PP reservoir layer at different soil infiltration rates and design storm precipitations. Thus, the subgrade infiltration rate and the rainfall depth were varied from 0.5 in./h to 2 in./h (increment of 0.25) and 1 in. to 6 in. (increment of 0.25), respectively.

Sensitivity Analysis Results and Discussion

The comparisons between results were done using linear inequalities graphs. The COTPP’s results were kept on the y-axis while the x-axis was either the results from PerviousPave, NAPA guide or ICPI guide

Structural Design Scenario for PC

The thickness of the PC layer was determined at different traffic levels (ADTT) and soil strengths (CBR). The graph (Figure 8) shows that the PC layer thicknesses obtained from the COTPP and PerviousPave are nearly the same. Most points are located on the left region; therefore, it can be said that the PC structural design results from the COTPP are more conservative than PerviousPave. The percent difference between the results ranged from 0 to 11.8% with a median of 3.8% and an average of 4.4%. This difference is attributed to the formulas used in the COTPP to determine the number of load applications (N) and the composite modulus of subgrade/subbase reaction (k), which were obtained from literature because they were not provided by the ACPA. In view of the results, it can be stated that the COTPP’s structural design results for PC are reliable.

Structural Design Scenarios for PA

The thickness of the PA layer was determined at three different traffic levels (ESALs), reservoir depths, and subgrade resilient modulus values. The graph (Figure 9) reveals points on the line of equality and others slightly to the left region of the line, showing differences between the PA layer thicknesses from the COTPP and the NAPA guide are minimal. The minor differences between the results can be explained by Burmister’s deflection factors (F₂) needed for the design were obtained manually from the F₂ graph to create a database and curve-fitting functions in the COTPP. This manual work is certainly a potential source of variability that affects the F₂ factors and inevitably the design results. For the 27,000 ESALs design scenario, the percent differences between the results ranged from 0 to 28.6% with a median of 8% and an average of 9.8%. For the 110,000 ESALs design scenario, the percent differences between the results ranged from 0 to 22.2% with a median of 5.4% and an average of 6.4%. For the 820,000 ESALs design scenario, the percent differences between the results ranged from 0 to 18.2% with a median of 4.1% and an average of 5.3%. These differences between the median and average for each scenario show that the COTPP is effective in the structural design of PA.

Structural Design Scenario for PICP

Concerning PICP, the thicknesses of the surface layer, the bedding layer, and the base layer are usually fixed values. Consequently, the scenario’s objective was to calculate only the structural subbase thickness at different traffic levels (ESALs) and subgrade strength (CBR). The graph (Figure 10) shows that all points are extremely close to the line of equality. The subbase layer thicknesses from the COTPP are found to be negligibly lower than the ones from the ICPI guide. These minor differences are because some assumed inputs such as the allowable serviceability decrease (ΔPSI) and the standard deviation (S₀) were not provided by the ICPI guide. Therefore, the recommended AASHTO values were used for the design and this can be a source of variability in the comparison. The percent difference between the results ranged from 0 to 11.8% with a median of 4.9% and a mean of 4.8%. Based on this comparison, the COTPP provides adequate PICP structural design results.

Hydrological Design Scenario for PC, PA, and PICP

In this scenario, the thickness of the reservoir layer had to be found based on varying soil infiltration rates and design storm precipitations. The graph (Figure 11) shows the summary of the comparison.

The percent differences between the results ranged from 0 to 22.9% with a median of 16.3% and an average of 15.1%. However, the trend observed in the graph shows that the reservoir layer thicknesses from both the COTPP and PerviousPave are closer when the design storm precipitation ranges from 0 to 2 in. This is a positive sign since PPs are typically designed to store the “first flush” runoff, which varies between 1 and 1.5 in. in Alabama ( 5 ). When the design storm precipitation is between 0 and 2 in., the median and average percent differences are 11% and 11.7%, respectively. Per Figure 11, reservoir depths from the COTPP progressively become larger than the ones from PerviousPave as the storm precipitation increases. This tendency can be explained by, unlike PerviousPave, the COTPP not using the PP surface layer as additional storage and the reservoir thicknesses are calculated using a recommended safety factor (infiltration rate of soil is halved) as recommended in the design guidelines ( 2 ). This practice helps avoid premature damage of PP. Based on all the reasons given to explain the high percent differences, the hydrological design results from the tool are considered acceptable since PPs are typically designed for low design storm precipitations.

In consideration of all the comparisons performed, it can be confirmed that the structural and hydrological design results from the COTPP are reliable and appropriate for PP.

Case Study

The objective of the case study was to investigate the capability of the COTPP to optimize design and costs of PPs. An existing PP site was selected and its bid item construction costs were input into the COTPP to verify whether or not the tool can reduce the total costs. The existing site is a permeable parking lot located in Alabama. It is comprised of PICP for center parking spaces, HMA pavement for driving lanes and outside parking spaces, and a bioretention basin to store runoff from the HMA pavement (Figure 12). This corresponds to the previously described Option C combination of the COTPP. The cross sections of all components are presented (Figures 13 and 14). For design purposes, the general hydrologic soil group (HSG) of the subgrade at the existing site is group B as obtained from the NRCS web soil survey ( 21 ). The construction unit costs for each component of the existing parking lot were obtained from bid documents and converted to appropriate units to include in the COTPP for the case study. The unit costs are 15.38 $/ft² for PICP, 1.61 $/ft² for HMA, 0.64 $/ft³ for bioretention, 0.73 $/ft³ for excavation, no data for geotextile fabric, and 0.805 $/ft² and 1.61 $/ft³ for the crushed aggregate base course underlying the HMA and PICP, respectively. The PICP unit cost is unusually high because it may include costs for other additional materials such as geotextile fabric, and so forth. The overall final construction cost for the combination of PICP, HMA, and bioretention at the existing site was $305,918.

The only known general parameters concerning the existing site were the cross-section dimensions (Figures 13 and 14), the areas, and/or the storage capacities of the components. At the existing site, the PICP area (7,800 ft²) and HMA area (29,250 ft²) occupy 21% and 79% of the total paved area, respectively. The bioretention basin has an area of 5,544 ft² and treats about 79% of total treatment volume. To determine the unknown general inputs for the existing site, design iterations were performed using typical PP design inputs until a close match of the as-constructed site design was obtained. The inputs used in the iteration that produced the design closest to the as-constructed site design were:

Design life = 20 years; Reliability = 80 %; CBR = 5%; ESALs = 500,000; Minimum reservoir depth = 9 in.; Base/subbase elastic modulus = 15,000 psi; Percent voids of reservoir layer = 40%. Subgrade infiltration rate = minimum 0.5 in./hr. for HSG B soil. External contributing impervious area (e.g., roof structures of adjacent buildings, etc.) = 0 ft². Rainfall depth = 11.7 in.

All the obtained inputs were used in the tool to verify that the tool can provide a match to the existing site before conducting the case study. As a result, the obtained design was nearly the same as the as-constructed site (Figure 15) and the final cost was $304,320, which is $1,598 lower than the actual bid cost. This difference is because some dimensions obtained were not fully equal to the as-built dimensions. For instance, the PICP area obtained was 7,781 ft² instead of 7,800 ft². Therefore, the $1,598 difference was added as an adjustment to every cost generated by the COTPP during the case study.

To demonstrate that the COTPP can optimize the original design and costs of the existing PP site, two cost optimization scenarios were simulated. The first scenario consisted of varying the area of PICP to optimize the design of the existing site and produce a new design cost that is lower than the as-constructed cost ($305,918). Since the PICP occupies 21% of the total area, the PICP area was varied from 10.5% (half of 21%) to 26% at 1% increments. This subsequently varied the HMA area from 89.5% to 74% of total area. The second scenario consisted of keeping the area of PICP constant and varying the runoff volume to be treated by the bioretention to optimize the design of the existing site and produce a new design cost that is lower than the as-constructed cost ($305,918). The Bioretention runoff volume was varied from 10% to 90% of the total treatment volume of the site (at 5% increment). This scenario has two parts. First, the PICP area was kept constant at 11% (about half of 21%) of the total parking lot area. Second, the PICP area was maintained at 21% of the total area, which is the actual percentage at the existing site.

Case Study Results and Discussion

In the first scenario, the final construction costs were obtained from the COTPP at varying PICP areas (10.5% to 26% of the total area) to compare with actual bid construction cost ($305,918). The summary of the comparison is shown (Figure 16). The final construction cost from the COTPP increases as the PICP area increases. This is because PICP tend to have expensive materials and labor owing to more challenging construction techniques than PA and PC. The construction costs were higher than the actual cost ($305,918) at PICP areas above 24% of the parking lot. However, the actual cost was minimized when the PICP area was between 10.5% to 24% of the total area. This proves that the COTPP could have helped the designer of the existing parking lot optimize the design and reduce cost even when the PICP occupies 21% of the parking lot. This cost optimization was achieved by the COTPP through iterations to obtain the most economical design for this combination (Option C). It was noticed that the bioretention basin had to treat more than 90% of the total treatment volume to reach this cost optimization. This is the reason scenario 2 was created to vary the bioretention runoff treatment volume and verify that the costs could still be optimized.

In the second scenario, the construction costs were determined from the COTPP at varying bioretention runoff treatment volumes (10% to 90% of total treatment volume) and compared with actual bid construction cost ($305,918). The summary of the comparison is shown (Figure 17). It is observed that all construction costs were lower than the actual bid cost when the PICP area was kept constant at 11% of the parking lot area. Also, when the PICP area was 21% of total area the construction costs were only lower than the actual cost at a bioretention runoff volume equal or higher than 80%. These results show that for this specific existing site the construction costs decrease as the bioretention treatment volume increases and the PICP area decreases. The lowest construction cost reached was $238,790 at PICP area occupying 11% of parking lot area and bioretention treating 90% of total runoff volume.

In view of the results from the case study, it can be concluded that the COTPP was able to optimize the design and costs of the existing PP site. The difference between the actual bid cost ($305,918) and the lowest construction cost ($238,790) reached during the case study was $67,128. This proves that the COTPP can optimize costs of PPs. It is important to note that the controlling factors of the design and cost optimization process are the design constraints and unit costs of all components at the site. However, it might be preferable to keep a large footprint of PICP for aesthetic reasons.

Conclusions

The purpose of this study was to develop a practical tool for the design and cost optimization of PPs. The tool incorporates the various industry standards or recommended methods for the structural design of PC, PA, and PICP. It also uses the ICPI method for the hydrological design of all three types of PPs. Furthermore, it contains a cost optimization algorithm created and included in the tool to optimize design and costs of PPs. The algorithm relies on the combinations of PPs with bioretention or infiltration trench and/or conventional pavements and helps designers select the most economical design for their project. This tool was developed in Microsoft Excel to facilitate its use by designers during their design and decision-making process.

A sensitivity analysis was conducted to evaluate the ability of the COTPP to provide reliable design results. The analysis consisted of comparing the COTPP’s results to design results from the existing software called PerviousPave, NAPA guidelines, and ICPI guidelines to verify that they are similar. The average percent differences between the structural design results were below 10% for PC, PA, and PICP. In addition, the average percent difference between the hydrological design results was 15.1% for all types of PP. This was attributable to the COTPP using a factor of safety that makes the results slightly more conservative. The percent differences were found minimal (below 10%) for structural designs and acceptable (around 15%) for hydrological designs. Therefore, the design results from the COTPP are acceptable and can be a trusted estimate in the design process of PPs.

A case study was also conducted to ensure that the COTPP can practically optimize costs of PPs. For that, an existing permeable parking lot and its materials construction costs were used in the case study to determine whether the COTPP can reduce costs. The components present at the existing site were PICP, a HMA pavement, and a bioretention basin. The results from the case study proved that the actual bid cost of the parking lot ($305,918) could be minimized to as low as $238,790 using the COTPP. Based on the findings from the case study investigations, it was concluded that the COTPP can optimize the design and costs of PPs effectively.

You may access and download the COTPP tool and User's Manual at the following web address: https://eng.auburn.edu/files/centers/hrc/cotpp.zip

Footnotes

Acknowledgements

The authors thank Dan Ballard and Marla Smith from the City of Auburn for assistance with site access and providing case study data and Robson Pachaly for assistance with field work.

Author Contributions

The authors confirm contribution to the paper as follows: study conception and design: Biessan, Bowers, O’Donnell, Vasconcelos, and Ellis; data collection: Biessan, Bowers, O’Donnell, Vasconcelos, and Ellis; analysis and interpretation of results: Biessan, Bowers, O’Donnell, Vasconcelos, and Ellis; draft manuscript preparation: Biessan, Bowers, O’Donnell, Vasconcelos, and Ellis. All authors reviewed the results and approved the final version of the manuscript.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by a USGS 104(b) grant awarded by the Alabama Water Resources Research Institute, grant #G16AP00037-2020AL354B.

ORCID iDs

Benjamin F. Bowers

Frances C. O’Donnell

Jose G. Vasconcelos

Data Accessibility Statement

The data generated during the sensitivity analysis and case study as well as the functions used to develop the tool are available from the corresponding author on reasonable request. The Tool and the user manual are also available on reasonable request.

References

Bean

E. Z.

Hunt

W. F.

Bidelspach

D. A.

Evaluation of Four Permeable Pavement Sites in Eastern North Carolina for Runoff Reduction and Water Quality Impacts. Journal of Irrigation and Drainage Engineering, Vol. 133, No. 6, 2007, 583–592.

Post Construction Storm Water Design Manual. City of Birmingham, AL, 2019.

Weiss

P. T.

Kayhanian

Gulliver

J. S.

Khazanovich

Permeable Pavement in Northern North American Urban Areas: Research Review and Knowledge Gaps. International Journal of Pavement Engineering, Vol. 20, No. 2, 2019, 143–162.

Harvey

Shan

Jones

D. J.

Fully Permeable Pavement for Stormwater Management: Progress and Obstacles to Implementation in California. Airfield and Highway Pavements, Vol. 2017, 2017, 125–136.

Dylewski

K. L.

Brown

J. T. R.

LeBleu

C. M.

Brantley

E. F.

Low Impact Development Handbook for the State of Alabama. Alabama Department of Environmental Management: Montgomery, 2014, pp. 73–86.

Smith

D. R.

Permeable Interlocking Concrete Pavement. 3rd Ed. Interlocking Concrete Pavement Institute (ICPI), Washington, D.C., 2006.

Biessan

D. G. V. V.

Developing a Tool for the Design and Cost Optimization of Permeable Pavements in the Planning Stage of Stormwater Management. MS thesis. Auburn University, AL, 2021.

ACPA (American Concrete Paving Association). PerviousPave. http://www.acpa.org/perviouspave/. Accessed 2020.

Vancura

MacDonald

Khazanovich

Structural Analysis of Pervious Concrete Pavement. Transportation Research Record: Journal of the Transportation Research Board, 2011. 2226: 13–20.

10.

Nassiri

AlShareedah

Preliminary Procedure for Structural Design of Pervious Concrete Pavements. Report No. WA-RD 868.2. Washington State Department of Transportation, Pullman, 2017.

11.

Titus-Glover

Mallela

Darter

M. I.

Voigt

Waalkes

Enhanced Portland Cement Concrete Fatigue Model for StreetPave. Transportation Research Record: Journal of the Transportation Research Board, 2005. 1919: 29–37.

12.

Rodden

Smith

Structural and Hydrological Design of Sustainable Pervious Concrete Pavements. Proc., Annual Conference of the Transportation Association of Canada, Edmonton, Alberta, American Concrete Pavement Association, 2011.

13.

Schwartz

C. W.

Hall

K. D.

Structural Design Guidelines for Porous Asphalt Pavements. National Asphalt Pavement Association (NAPA), Lanham, MD, 2018.

14.

Smith

D. R.

Industry Guidelines for Permeable Interlocking Concrete Pavement in the United States and Canada. Interlocking Concrete Pavement Institute, Herndon, VA, 2012.

15.

Rehan

Werner

Life-Cycle Cost Analysis for Traditional and Permeable Pavements. Proc., Construction Research Congress, New Orleans, Louisiana, Edwardsville, IL, 2018.

16.

Olson

Roesner

Urbonas

MacKenzie

Best Management Practices – Rational Estimation of Actual Likely Costs of Stormwater Treatment. User’s Manual and Documentation Version 2.0. 2017.

17.

US EPA (United States Environmental Protection Agency). Opti-Tool: EPA Region 1’s Stormwater Management Optimization Tool [Overviews and Factsheets]. US EPA, 2016.

18.

Wood and Volkert. Guidance Document for the Integration of Green Infrastructure. City of Auburn, AL, 2019.

19.

Walsh

N. P.

Smallridge

Latest High-Tech Installation Methods Minimize Interlocking Concrete Paver Construction Costs and Speed Project Delivery of Marine Terminal Pavements. In Ports '01: America's Ports: Gateway to the Global Economy ( Collins

T. J.

, ed.), American Society of Civil Engineers, Reston, VA, 2001.

20.

EPA (Environmental Protection Agency). Green Infrastructure Design and Implementation. 2022. https://www.epa.gov/green-infrastructure/green-infrastructure-design-and-implementation#Design%20Manuals. Accessed June 15, 2022.

21.

NRCS (National Resource Conservation Service, United States Department of Agriculture). Urban Hydrology for Small Watersheds, TR 55. Washington, D.C., 1986.