World Sweeper Logo

Environmental Information for Sweeping Professionals


City of San Jose Comparison of Sweepers


Email Woodward-Clyde Consultants for more information, or visit their website.

This report was prepared by Woodward-Clyde Consultants for use only by the City of San Jose (however the company provided it as a courtesy to be posted on the WorldSweeper.com website). Nothing contained in this report shall be considered as an endorsement or a lack of endorsement of any of the equipment mentioned in the report.

The City and Woodward-Clyde Consultants specifically disapprove any use of the information contained in this report for any promotional or advertising of any of the equipment mentioned in the report.


TABLE OF CONTENTS

EXECUTIVE SUMMARY

1.0 INTRODUCTION

1.1 OBJECTIVES

1.2 REPORT ORGANIZATION

2.0 STUDY METHODOLOGY

2.1 STUDY DESIGN

2.2 TEST ROUTE LOCATIONS

2.3 SAMPLING PROCEDURES

2.4 CHEMICAL ANALYSIS

2.5 CALCULATIONS

2.6 STATISTICAL APPROACH

3.0 STUDY RESULTS

3.1 ROUTE DIFFERENCES

3.2 SWEEPER DIFFERENCES

3.2.1 PARAMETERS EXAMINED

3.2.2 RESULTS

3.3 QUANTIFICATION OF IMPROVEMENT IN COPPER COLLECTION

3.4 EFFECTIVENESS OF PICKING UP DIFFERENT PARTICLE SIZES

4.0 CONCLUSIONS


EXECUTIVE SUMMARY

The City of San Jose currently uses Mobil broom sweepers to sweep the arterial-commercial streets and bikeways in the city. Some of these sweepers are more than ten years old. Because of the age of the City's existing street sweeper fleet, the City has budgeted money to replace these old sweepers with new sweepers.

The City's purchase of new street sweepers presents an opportunity to increase the effectiveness of street sweeping over that achieved with the existing Mobil broom sweepers. If a 'better' sweeper was purchased, the additional pollutants removed could help achieve the San Francisco Bay Regional Water Quality Control Board's goal of reducing storm water pollution discharges to the Bay.

Five sweeper types were tested, Tymco regenerative air, Elgin regenerative air, Elgin broom, new Mobil broom and an old Mobil broom. On eight occasions, each sweeper swept a 4 to 5 mile route randomly selected from a set of 20 routes. Samples of sediment collected by the sweepers were analyzed for metals, petroleum hydrocarbons and selected organic compounds. The total mass of sediment collected by each sweeper was estimated by weighing each sweeper before and after it swept a route.

It was assumed that most of the pollutants were associated with the sediment contained in the hopper rather than the trash contained in the hopper. To determine the fraction of the hopper mass composed of debris, a subsample of the hopper sediment was sieved through a one-centimeter (cm) sieve. Both the sieved and the unsieved portions were weighed and the ratio of unsieved weight to total weight was the fraction of hopper materials composed of debris.

Differences in the mass of dirt and pollutants between sweeping routes were estimated based on sediment samples collected with an industrial hand-vacuum on each route before it was swept. A total of 40 hand-vacuum samples were collected, two from each route. The samples were analyzed for copper, lead, zinc and particle size. The copper data were compared to determine if any routes were different from the average route.

Only three routes, Blue 3, Red 1 and Yellow 3 appeared to be significantly dirtier than the average route. Route Blue 3 is along a commercial section of Monterey Highway with an Average Daily Traffic (ADT) volume 50% greater than the average ADT of all the routes in this study. Red 1 was located along an industrial section of Lundy and Brokaw roads. It had an ADT of 29,641 cars per day, slightly higher than the average ADT. Yellow 3 is located along a commercial section of Tully Road and had the highest ADT of any of the routes sampled, 52,000 cars per day.

The three measures of effectiveness used in the study were:
  • total mass of sediment collected per curb mile,
  • mass of copper collected per curb mile,
  • concentration of copper in the collected sediment.
Copper was selected as the indicator pollutant because it is commonly found in the street sediments and is a pollutant of concern in the Bay.

Based on a statistical analysis of the data, the conclusions from the study include:

  • Tymco regenerative air, Elgin regenerative air and the new Mobil broom sweepers pick up a greater mass of sediment than the old Mobil broom or the Elgin broom.
  • Tymco and Elgin regenerative air sweepers pick up sediments with higher copper concentrations than do the broom sweepers tested.
  • Tymco and Elgin regenerative air and the new Mobil broom pick up more copper mass per curb mile than the old Mobil broom or the Elgin broom.

No significant differences were seen between the regenerative air sweepers and the new Mobil broom sweeper. Based on these results, use of either of the regenerative air sweepers or the new Mobil broom sweeper would increase the copper collected by the street sweeping program.

The ability of sweepers to remove a range of particle sizes, and especially smaller particles may be important because smaller particles have been shown to have higher metals concentrations and are more easily mobilized during a storm event. In order to investigate this effect sweeper sediments were separated into five size fractions between 75 /m and 1,000/m. Each fraction was analyzed for total mass and concentration of pollutants.

1.0 INTRODUCTION

The City of San Jose (City) presently uses Mobil broom sweepers to sweep the arterial-commercial streets and bikeways (ACB) in the city. Some of these sweepers are more than ten years old. Because of the age of the city's existing street sweeper fleet, the city has budgeted money to replace these old sweepers with new sweepers.

The city's impending purchase of new street sweepers presents an opportunity to further reduce the loads of pollutants to San Francisco Bay through more efficient street sweeping. The City decided to evaluate whether the purchase of a particular sweeper brand or model would improve the removal of pollutants from city streets over that occurring with the existing Mobil broom sweepers. If a more efficient sweeper were used, the additional amount of copper removed could be applied as a credit towards achieving the San Francisco Bay Regional Water Quality Control Board's ( Regional Board) goal of reducing storm water pollution discharges to San Francisco Bay. The following sweepers were evaluated in the study in addition to the existing Mobil broom sweeper:

1. Tymco regenerative air sweeper
2. Elgin Crosswind regenerative air sweeper
3. Elgin Eagle broom sweeper
4. New Mobil broom sweeper

1.1 OBJECTIVES

The objective of this study is to evaluate the four different sweepers listed above and the existing sweepers used by the city of San Jose and provide recommendations to the City on which sweeper is the most effective at removing copper from the street surface.

1.2 REPORT ORGANIZATION

Section 2.0 of this report describes the procedures used in conducting the study. Section 3.0 describes the results of the study. Section 4.0 provides conclusions and recommendations on which sweepers appear to provide the greatest benefit in copper removal.

2.0 STUDY METHODOLOGY

Three variables that influence the effectiveness of street sweeping are the existing street conditions, (i.e., 'dirtiness' of streets or wet or dry), the frequency of street sweeping and the frequency and amount of rainfall. The following steps were taken to control these effects:

1. Streets were swept only when dry and after a period of at least two weeks with no rain. Dry streets are the normal condition for the San Jose area and a two week interval with no rain allows 'dirt' to accumulate on the streets prior to conducting the tests.

2. Sweepers were randomly assigned to test routes.

3. Sweeping frequency was fixed at two week intervals.

This section describes the selection of routes and the collection of data.

2.1 STUDY DESIGN

The study was designed to statistically test the null hypothesis that the Elgin Eagle broom sweeper, the Elgin Crosswind regenerative air sweeper, the Tymco regenerative air sweeper, a new Mobil broom sweeper and an old Mobil broom sweeper all remove the same amount of copper from the street surface.

On days when the sweepers were tested, each of the five sweepers (four new sweepers plus one old sweeper) were randomly assigned two test routes. One of the test routes was swept during the first half of the sweeping shift and one was swept during the second shift of the day. Thus, a total of ten test routes were swept each day.

2.2 TEST ROUTE LOCATIONS

Twenty test routes were selected for the study. The test routes represented the different land uses and traffic volumes found within the ACB area. Each test route consisted of four to five curb miles of city street. Table 2-1 (available with the complete report from Woodward-Clyde Consultants) lists the characteristics of each test route. In the table, class indicates the type of street (arterial or collector), lane provides an indicator of the size of the street and average daily traffic describes the volume of traffic.

To determine the variability between test routes, samples of street dirt were collected from each test route with a hand-vacuum prior to conducting each street sweeping test. This is in addition to a pilot study to examine route variability that was conducted prior to this study.

2.3 SAMPLING PROCEDURES

To evaluate sweeper effectiveness it was necessary to collect enough data to adequately determine the variability in the sample data. This information was then used to determine the statistical significance of the results obtained. As mentioned earlier, on a sampling day, each sweeper swept two test routes, one during the first shift and one during the second shift. Separate samples were obtained from each sweeper on each shift. Each sweeper was tested for a total of 8 shifts, yielding a total of 8 samples per sweeper and 40 samples for all 5 sweepers.

From the point of view of routes, two samples were collected on most routes, however, due to some problems in sweeper assignments on the first day of the test, a third sampling round was conducted. Therefore, three samples were obtained for some routes.

To evaluate effectiveness, both the mass of sediment in a hopper and the concentrations of selected pollutants in the sediments were measured.

After a sweeper completed its test route, its hopper was emptied. It was assumed that most of the pollutants were associated with the sediment contained in the hopper as opposed to the debris. To determine the fraction of the hopper mass that is debris, an approximately one-wheel-barrow sized sample of the hopper material was sieved through a one-centimeter (cm) sieve. Both the sieved and unsieved portions were weighed and the mass fraction of each portion was determined.

All samples collected for chemical analysis were taken from the less-than-1-cm size portion of the load. A representative sample of the sediment was obtained by compositing seven to eight separate sub-samples.

Vacuum Samples
Each test route was sampled with a hand vacuum before it was swept. Samples were collected by vacuuming an approximately one foot wide strip extending from the curb out about four feet towards the centerline. Ten of these strips were vacuumed in each test route. A sample of the sediment collected in the vacuum was sent to the lab for analysis after completing each route.

Mass of Material in Hopper
The mass of material collected by each sweeper was determined by weighing the truck at a public scale before sweeping the test route and again after completion of the test route. The difference between these weights represents the mass of material collected by the sweeper. The sweeper's gas and water tanks were filled to the same level before it was weighed.

Equipment Blanks
Four equipment blank samples for the vacuum were collected to estimate the level of contamination, if any, contributed by the vacuum. These samples consisted of clean commercially purchased Monterey sand. The sand was poured onto a plastic sheet and vacuumed with the vacuum using the same filter bags used in the study. A sample of the sand from the vacuum and a blank sample of the sand were submitted to the laboratory for analysis.

2.4 CHEMICAL ANALYSIS

Samples were analyzed for the constituents listed in Table 2-2 (available with the complete report from Woodward-Clyde Consultants). Analysis methods and detection limits are also shown in the table. Consistent with the scope of work, this report describes the results for copper only.

2.5 CALCULATIONS

Analysis of the data collected during the study were conducted for total copper concentration, mass of material collected and mass of copper collected per mile swept. The copper concentration and mass of material collected were obtained directly. The mass of copper per curb mile swept was calculated as:

Mcm = Ch *Mh/L/1E06
where:
Mcm = Mass of copper per curb mile [lbs/mile]
Ch = Copper concentration in hopper sediments [mg/kg]
Mh = Mass of material in hopper [pounds]
L = Length of route [miles]
1E06 = unit conversion from mg to kg [mg/kg]

2.6 STATISTICAL APPROACH

The objective in the statistical analysis is to select and apply a method that can distinguish between the effects of sweeper performance and other factors (specifically routes) that also could affect the amount and chemical concentration of material collected by the sweepers. The following describes the procedures used in this study.

There are several different statistical methods which can be used to determine differences between groups of data. Most of these methods are based on comparisons between estimated means of the groups. Associated with each estimated mean is a range, called the confidence interval, within which the actual mean is expected to fall with a certain probability. Typically, 95% confidence intervals are used, meaning that the actual mean will fall within the confidence interval 95% of the time. Statistical methods calculate and compare the confidence intervals of different groups (e.g. sweeper types) to determine if and how much they overlap. If the overlap between the confidence intervals is much less than that expected by chance alone, the two means are concluded to be statistically different. A typical statistical criteria is 95% confidence or 5% error (P < 0.05). At this level of confidence the test result is predicted to be correct 95% of the time and incorrect 5% of the time.

Two major classes of statistical techniques are 'parametric' and 'nonparametric'. Parametric techniques compare the means of the actual data values (i.e., copper mass/curb mile) while nonparametric techniques compare the ranks of the data. To generate ranks, the data are ordered (i.e., ranked) from low to high and the lowest value is assigned a rank of 1, the next lowest is assigned a rank of 2, etc. up to the highest value which is assigned the highest rank. The ranks are then treated as if they were actual values (i.e. mean ranks and confidence intervals calculated and compared). The choice of whether to use nonparametric or parametric methods is usually determined based on the distribution of the data (i.e., normal) and the number and size of any outliers.

Analysis of variance (ANOVA) is an advanced statistical method which looks at the variability of all the data and at the variability within each class of data at the same time (a class would be one type of sweeper). The method constructs a mean and confidence interval for each sweeper type based on the data values and number of observations. When there is only one class or effect being examined (sweeper type only) the tests are called one-way ANOVA. If more than one effect (i.e., sweeper and route) is being examined the test is called a two-way ANOVA.

It should be stressed that in this study each sweeper only swept 7 or 8 of the 20 routes used in the study. This is sufficient data to compare different sweeper types but characterization of the route effects is less reliable then if each sweeper swept all routes.

There are two assumptions which must be satisfied in order for the parametric ANOVA to be considered valid:
  • 1. Homogeneity in variance between groups, i.e., the variance for each sweeper should be similar.
  • 2. The residuals (the error in the statistical estimate of each value) should be normally distributed.

If these assumptions are violated parametric ANOVA can not be used. In these cases, nonparametric tests can be used because they treat each data point equally and are not influenced by outliers. However, the disadvantage of the nonparametric technique is that such tests can only determine if two data sets are different. The tests cannot determine the amount by which the data sets differ.

Nonparametric ANOVA procedures were used because both assumptions of the parametric ANOVA were not satisfied due to a few extreme data values generally associated with a few routes, notably Red 1, Yellow 2, and Blue 3. Therefore, nonparametric ANOVA procedures were used to determine if there were significant differences between sweepers. Parametric procedures, excluding the extreme concentrations were used to estimate the size of the differences found.


Return to the Table of Contents

Continue with Part 2 of the Study



This article came from: Woodward-Clyde Consultants
World Sweeper Logo

© 2005 - 2016 World Sweeper
All rights reserved.

Back to Environmental Information
Site Map / Table of Contents