Transcript Advanced Hazardous Waste Inspector

8th Annual
California Unified Program
Conference
7/17/2015
9th Annual California Unified Program Conference
1
What is a valid Waste
Determination?
Part II. Analysis or
Knowledge of Process?
7/17/2015
9th Annual California Unified
Program Conference
2
Most of the Time…

7/17/2015
It’s simple
9th Annual California Unified Program Conference
3
But when it isn't simple,
who makes the waste
determination?
The Generator

The person whose act or process produces
hazardous waste or whose act first causes a
hazardous waste to become subject to regulation.”

A hazardous waste Generator must comply with the
requirements of Title 22 CCR, Division 4.5,
Chapter 12.
7/17/2015
9th Annual California Unified Program Conference
4
§66262.11 Hazardous
Waste Determination

First, the generator must determine if it is a waste.





§66261.2 Definition of a waste
§ 66261.3 Definition of a hazardous waste
§66261.4 Materials which are not waste
§25143.2 Excluded recyclable materials
Next, the generator must determine if it is a hazardous
waste.


Is it listed in article 4 or in Appendix X of Chapter 11?
Or does it exhibit any of the characteristics set forth in article 3
of Chapter 11?
7/17/2015
9th Annual California Unified Program Conference
5
§66262.11 Hazardous
Waste Determination (cont).




The Generator can make a hazardous waste
determination by:
(1) Testing; or
(2) Applying knowledge of the hazard
characteristic of the waste in light of the
materials or the processes used.
This is also called waste analysis.
7/17/2015
9th Annual California Unified Program Conference
6
Waste Analysis

The cornerstone of a hazardous waste program is the
ability of facility personnel to identify properly, through
waste analysis, all the wastes they generate, treat, store,
or dispose of. Waste analysis involves identifying or
verifying the chemical and physical characteristics of a
waste by performing a detailed chemical and physical
analysis of a representative sample of the waste, or
in certain cases, by applying acceptable knowledge of
the waste.
7/17/2015
9th Annual California Unified Program Conference
7
Testing


Accurate analytical data is required to comply
with Chapter 18, LDR requirements.
A written Waste Analysis Plan (WAP) is required
for:



TSDFs,
PBR Treatment and,
Generators treating hazardous to meet LDR standards.
7/17/2015
9th Annual California Unified Program Conference
8
A Waste Analysis Plan




Establishes consistent internal management
mechanism(s) for properly identifying wastes on site.
Ensures that waste analysis participants have identical
information (e.g., a hands-on operating manual),
promoting consistency and decreasing errors.
Ensures that facility personnel changes or absences do
not lead to lost information.
Reduces your liabilities by decreasing the instances of
improper handling or management of wastes.
7/17/2015
9th Annual California Unified Program Conference
9
Waste Analysis Plan?
http://www.epa.gov/epaoswer/hazwaste/ldr/wap330.pdf
7/17/2015
9th Annual California Unified
Program Conference
10
Article 3. Characteristics
of Hazardous Waste
§66261.20.General

(a) A waste, as defined in section 66261.2, which is not
excluded from regulation as a hazardous waste pursuant to
section 66261.4(b), is a hazardous waste if it exhibits any of the
characteristics identified in this article

(c) Sampling and testing pursuant to this article shall
be in accord with the chapter nine of SW-846, the
Department will consider samples obtained using any
of the other applicable sampling methods specified in
Appendix I of this chapter to be representative
samples.
7/17/2015
9th Annual California Unified Program Conference
11
Characteristic Wastes




§66261.21 (a) A waste exhibits the characteristic of ignitability if
representative samples of the waste have any of the following
properties:
§66261.22(a) A waste exhibits the characteristic of corrosivity if
representative samples of the waste have any of the following
properties:
§66261.23(a) A waste exhibits the characteristic of reactivity if
representative samples of the waste have any of the following
properties:
§66261.24 (a) A waste exhibits the characteristic of toxicity if
representative samples of the waste have any of the following
properties:
7/17/2015
9th Annual California Unified Program Conference
12
Representative Sample

§66260.10. Definitions. "Representative sample"
means a sample of a universe or whole (e.g.,
waste pile, lagoon, ground water) which can be
expected to exhibit the average properties of the
universe or whole.
7/17/2015
9th Annual California Unified Program Conference
13
Note:
Enforcement Sample



A regulator does not necessarily need a representative
sample to support an enforcement action.
The primary reason is that the data quality objectives
(DQOs) of the enforcement agency often may be
legitimately different from those of a waste handler.
A sample taken for enforcement is used to demonstrate
that the waste exceeds a standard (e.g. STLC).
7/17/2015
9th Annual California Unified Program Conference
14
EPA publication SW-846
Test Methods for Evaluating Solid
Waste, Physical/Chemical Methods
OSW's official compendium of approved
analytical and sampling methods for use in
complying with RCRA regulations.
SW-846 primarily is a guidance document
that sets forth acceptable, although not
required, methods for the regulated and
regulatory communities to use for RCRArelated sampling and analysis requirements.
7/17/2015
9th Annual California Unified
Program Conference
15
SW 846
http://www.epa.gov/sw-846/sw846.htm
7/17/2015
9th Annual California Unified
Program Conference
16
SW 846, Chapter 9,
Sampling Plan

SW 846 assumes that:

The concentration of a contaminant in individual
samples will exhibit a normal distribution.
 Simple random sampling is the most appropriate
sampling strategy.
As more information is accumulated, greater
consideration can be given to different sampling
strategies.
Start with simple random sampling and assume a
normal distribution.


7/17/2015
9th Annual California Unified Program Conference
17
Population
90 bags of candy
10 bags contain 0 pieces of (0 pieces)
20 bags contain 1 piece of (20 pieces)
30 bags contain 2 pieces of (60 pieces)
20 bags contain 3 pieces of (60 pieces)
10 bags contain 4 pieces of (40 pieces)
Population mean is 180/90 = 2







7/17/2015
9th Annual California Unified Program Conference
18
Histograph of candy population
(normal distribution)
Population mean = 2
30
20
10
0 1
7/17/2015
2
3
4
9th Annual California Unified Program Conference
19
Random sample



Four samples from web 14, 37, 40, 81 (90 bags)
Four samples from web 7, 19, 35, 41 (50 bags)
Six samples from web 3, 24, 64, 71, 76 , 90
7/17/2015
9th Annual California Unified Program Conference
20
Histograph of Samples
3
2
1
0 1
7/17/2015
2
3
4
9th Annual California Unified Program Conference
21
Normal Distribution
In a normal distribution a
Samples
bell shaped curve is used to
represent the boundaries
of the population. The
“true” population (under the
blue curve) is never known,
but precise and unbiased
samples will provide an
accurate estimate of the The sample population under the
true population.
magenta curve is an estimate of
the true population.
7/17/2015
9th Annual California Unified Program Conference
22
A Bell Curve has Tails!
I left the tails off most
of the diagrams because
I couldn’t figure out
how to draw them!
The
The
The
got
7/17/2015
X axis is the concentration.
Y axis is the number of samples.
tails are where the people who
100% or 0% on an exam are found.
9th Annual California Unified Program Conference
23
Reliable Waste Analysis




Reliable information concerning the chemical
properties of a solid waste is needed for
comparison with applicable regulatory thresholds.
If chemical information is to be considered reliable,
it must be accurate and sufficiently precise.
Accuracy (no bias) is usually achieved by
incorporating randomness into the sample
selection process.
Sufficient precision is most often obtained by
selecting an appropriate number of samples.
7/17/2015
24
Sample size


7/17/2015
Small samples (A) cause the
constituent of interest to be
under-represented in most
samples and over-represented
in a small proportion of
samples. Larger samples (B)
more closely reflect the parent
population.
Sometimes you sample a large
portion or even the entire
population, so you don’t need
statistics to determine a
confidence interval.
9th Annual California Unified Program Conference
25
Terminology
Precise, Accurate & Biased



Precise means all of the samples are similar; they form
a “tight group” on the graph. Taking more samples or
taking larger samples will increase the sample
precision.
Accurate or unbiased means that you’re taking truly
random samples. Properly planned random samples are
accurate and unbiased samples.
Inaccurate samples are synonymous with biased
samples. They are not representative samples. Poor
tool selection or calibration can cause sample bias.
7/17/2015
9th Annual California Unified Program Conference
26
Biased & Imprecise
Samples
Biased samples do not
represent the true
population. The biases
could result from poor
tool selection or
contamination.
Imprecise samples
have a lot of
variation. More
samples should
decrease variation.
7/17/2015
0
Mean
1012.5
9th Annual California Unified Program Conference
2000
27
Biased & Precise
Samples
A poor sampling plan
could lead to biased or
inaccurate samples.
Poor tool selection,
poor sampling design
or contamination are
some causes.
Biased sampling shifts
the population curve.
0
Sample True
Mean Mean
2000
Who can think of another cause for biased samples?
7/17/2015
9th Annual California Unified Program Conference
28
Unbiased & Imprecise
Samples
Unbiased samples are
Random samples. Random
samples fall inside the bell
curve that represents the
true population.
Take more
samples to
increase the
precision.
7/17/2015
0
Mean
1012.5
9th Annual California Unified Program Conference
2000
29
Unbiased & Precise
Samples
Unbiased samples
are a function of
randomness. Random
sampling requires proper
plan design and tool
selection.
Precise samples
are a function
of the number
of samples.
7/17/2015
0
Mean
1012.5
9th Annual California Unified Program Conference
2000
30
Waste Analysis (Testing)

To evaluate the physical and chemical properties
of a solid waste
The initial -- and perhaps most critical -- element
is the sampling plan.

Analytical studies, with their sophisticated
instrumentation and high cost, are often perceived
as the dominant element.

But analytical data generated by a scientifically
defective sampling plan have limited utility.
7/17/2015
9th Annual California Unified Program Conference
31
SW 846




Waste characterization requires a
representative sample.
At least two samples of a material are required
for any estimate of precision.
SW 846 uses an 80% confidence interval as an
acceptable degree of sampling accuracy and
precision.
Normally data from four representative
samples is the minimum required to achieve
an 80% confidence interval.
7/17/2015
9th Annual California Unified Program Conference
32
How many samples are
enough? An example



A business wants to dispose of a pile of used blast
medium. It has been reused and it is well mixed. It
might have been used to remove paint with lead
pigment.
Is it hazardous?
Testing or knowledge of process?


It might have lead? – Knowledge??
How many samples do need for testing?

Four?
7/17/2015
9th Annual California Unified Program Conference
33
Sampling Plan




Make a 3-D grid of the pile. Number each area
of the grid.
Select four numbers randomly. Random
number generators are on the web, tables or in
textbooks.
Sample from the four areas represented by the
number.
Analyze the samples using TTLC.
7/17/2015
9th Annual California Unified Program Conference
34
Sample Results





The TTLC for lead is 1000 mg/kg.
Sample A contains 1000 mg/kg. Is sample A
hazardous waste?
Is the waste pile hazardous?
Sample B contains 1050 mg/kg, sample C
contains 980 mg/kg and sample D contains 1020
mg/kg.
Is the waste pile hazardous?
7/17/2015
9th Annual California Unified Program Conference
35
Is it hazardous?
A. Yes, 3 of 4 is good enough.
B. No, it’s 100% or nothing.
C. More analysis and maybe more samples are
required.
The answer is C!
7/17/2015
9th Annual California Unified Program Conference
36
More Analysis?

Yes, more analysis.

The samples were pretty close, A contains 1000
mg/kg, B contains 1050, C: 980 & D: 1020.
A range of only 70 mg/kg.

Do we need more samples?

7/17/2015
Yes, well…
9th Annual California Unified Program Conference
37
Guess
how many samples
4
5
15
20?
The answer is 15.31
Where did that number come
from?
7/17/2015
9th Annual California Unified Program Conference
38
A Seven step Statistical
Process is used to determine
number of samples (SW 846 Table 9-1)
1.
2.
3.
4.
5.
6.
7.
7/17/2015
Determine the mean
Determine the variance
Determine the standard deviation
Determine the standard error
Determine the confidence interval
Determine if the variance is > the mean
Determine the appropriate number of samples.
9th Annual California Unified Program Conference
39
Statistics,
the last time…




7/17/2015
I would have gotten a
PHD if I liked math.
Give it a chance!
It’s just addition,
multiplication and
division.
Oh, and square roots, but
you can use a calculator.
9th Annual California Unified Program Conference
40
If you really hate
Numbers
Pretend to listen, it’s the
polite thing to do, and
remember:
 You need at least four (4)
samples.
 More samples may be
required if the waste is:


7/17/2015
Heterogeneous, or
Close to the regulatory
threshold
9th Annual California Unified Program Conference
41
Step #1: The Mean
Samples
A: 1000ppm B: 1050ppm
C: 980ppm D: 1020ppm
The sample mean
is the average
value of the
samples. It’s an
estimate. The
true mean is
never known.
7/17/2015
0
Sample
Mean
1012.5
9th Annual California Unified Program Conference
2000
42
Normal Distribution
Variance
The variance is
the sum of the
differences
between the
sample values and
the mean, squared.
The variance sets
the boundaries of
the distribution.
7/17/2015
variance
0
Mean
9th Annual California Unified Program Conference
2000
43
Standard Deviation
Standard
Deviation
The standard deviation
is the square root
of the variance.
variance
0
Mean
1012.5
7/17/2015
9th Annual California Unified Program Conference
2000
44
Normal
Distribution CI
80%
Confidence
Interval (CI)
If you take 100 samples,
80 should fall inside the
boundaries of the
80% CI.
variance
0
Mean
1012.5
7/17/2015
9th Annual California Unified Program Conference
2000
45
Normally you would
evaluate all four samples


All four randomly selected samples must be
considered in a valid statistical analysis.
In the following example, four sets of two will
also be analyzed to illustrate the effects of:



7/17/2015
Decreasing the variance in concentration in the
samples.
Increasing number of samples.
The relationship of the mean to the Regulatory
Threshold (RT).
9th Annual California Unified Program Conference
46
Step 1. The Mean

Add the results of all samples and divide by
the number of samples
Sample A=1000ppm
Sample C=980ppm
Sample B=1050ppm
Sample D=1020ppm
MEAN

A+B+C+D =(1000+1050+980+1020)/4 = 4050/4 =
1012.5 ppm




7/17/2015
A+B
C+D
B+D
A+C
= 2050/2 = 1025 ppm
= 2000/2 = 1000 ppm
= 2070/2 = 1035 ppm
= 1980/2 = 990 ppm
9th Annual California Unified Program Conference
47
Step 2. The Variance
Variance = (sample A - mean)2 + (sample B - mean)2 +(..)
Number of samples - 1


(1000-1012.5)2+(1050-1012.5) 2 +(980-1012.5) 2 +(1020-1012.5)2
3
(12.5)2 + (37.5)2 + (32.5)2 + (7.5)2 = 2675/3 = 891.67
3
A+B: (1000-1025)2 + (1050-1025) 2 =1250
1
C+D: ( 980 - 1000) 2 + (1020 - 1000) 2 = 800
1
B+D: (1050 - 1035) 2 + (1020 - 1035) 2 = 450
1
A+C: (1000 - 990) 2 + ( 980 - 990) 2 = 200
1
7/17/2015
9th Annual California Unified Program Conference
48
Step 3
Standard Deviation
A=1000 ppm, B=1050 ppm, C= 980 ppm, D=1020 ppm

Standard Deviation = Variance 1/2
The variance of A+B+C+D is 891.67; the square root of
891.67 (standard deviation) = 29.86




A+B: Variance = 1250; standard deviation = 35.35
C+D: Variance = 800; standard deviation = 28.28
B+D: Variance = 450; standard deviation = 21.21
A+C: Variance = 200; standard deviation = 14.14
7/17/2015
9th Annual California Unified Program Conference
49
Step 4
Standard Error
A=1000ppm, B=1050ppm, C= 980ppm, D=1020 ppm
Standard Error = Standard Deviation
(Number of samples) ½

Standard error ABCD = 29.86/(4)1/2 = 14.93




Standard error A + B = 35.35/1.41 = 25.07
Standard error C + D = 28.28/1.41 = 20.06
Standard error B + D = 21.21/1.41 = 15.04
Standard error A + C = 14.14/1.41 = 10.03
7/17/2015
9th Annual California Unified Program Conference
50
Step 5
Confidence Interval
A=1000ppm, B=1050ppm, C= 980ppm, D=1020 ppm
Confidence Interval = Mean ±(student “t”)(standard error)
 A+B+C+D: 1012.5 ± (1.638)(14.93) = 1012.5 ± 25.46
(988 to 1038). 80 of 100 samples should have
concentrations between 988 and 1038 ppm.




A+B: 1025 ± (3.078)(25.07) = 1025 ± 77 (948 to 1102)
C+D: 1000 ± (3.078)(20.06) = 1000 ± 62 (938 to 1062)
B+D: 1035 ± (3.078)(15.04) = 1035 ± 46 (989 to 1081)
A+C: 1010 ± (3.078)(10.03) = 1010 ± 31 (979 to 1041)
7/17/2015
9th Annual California Unified Program Conference
51
Step 6. Is the Variance
>the Mean?



If the variance is not greater than the mean, go to
step 7.
A + B + C + D: 891.67 is not > 1012.5
If the variance is greater than the mean , you have
to transform the data. An example follows for
samples A & B.




7/17/2015
A+B:
C+D:
B+D:
A+C:
1250 is > 1025
800 is not > 1000
450 is not > 1035
200 is not > 990
9th Annual California Unified Program Conference
52
Is the Variance > the
Mean?
Mean
0
Variance
0
If variance is > mean then part of the population is less than zero, i.e.
with samples A & B the population is between -225 and 2275. You can’t
have a concentration of less than zero so you have to transform the data.
7/17/2015
9th Annual California Unified Program Conference
53
Not more math!



7/17/2015
OK, we won’t transform
the data, here
But in your handout the
next four slides take the
square root and go
through the steps 1 to 6
and square the data to
return to real numbers.
Go to step 7.
9th Annual California Unified Program Conference
54
Transform the data if the
variance is >the mean
Usually data is transformed into a smaller number
by taking either the log or the square root of the
value.
 Step 1a. Transforming the mean
 10001/2 = 31.62
1/2 = 32.40
 1050
Total
64.02/2 = 32.01

7/17/2015
9th Annual California Unified Program Conference
55
Transforming the Variance &
Standard Deviation

Step 2a. Transforming the Variance

Variance = (sample A - mean)2 + (sample B - mean)2


Number of samples - 1
A+B: (31.62 – 32.01)2 + (32.4 - 32.01) 2 = 0.304
1
Step 3a. Transforming the Standard Deviation
Standard Deviation = Variance 1/2
A+B = (0.304)1/2 = 0.5515




7/17/2015
9th Annual California Unified Program Conference
56
Transforming the Standard
Error and Confidence Interval


Step 4a. Transforming the Standard Error
Standard Error = Standard Deviation

(Number of samples) 1/2
A+B:
0.5515/1.41 = 0.39
Step 5a. Transforming the Confidence Interval
 Mean ±(student “t”)(standard error)
 A+B: 32.01 ± (3.078)(.39) = 32.01 ± 1.20

7/17/2015
9th Annual California Unified Program Conference
57
Step 6a
Variance > Mean

A+B: The transformed variance (0.304) is not
greater than the transformed mean (32.01).

Now go to the last step #7.
7/17/2015
9th Annual California Unified Program Conference
58
Step 7. Determine the
number of samples (n)

The Regulatory Threshold (RT) using TTLC for
lead is 1000 ppm.
 n = (student t)2(variance)
(RT – mean)2d

Use the square root of the RT for lead (36.62)
for transformed data.
7/17/2015
9th Annual California Unified Program Conference
59
n = (student t)2(variance)
(RT – mean)2

A + B + C + D: (1.638)2 (892) = 15.31
(1000 – 1012.5)2

A+B:
(3.078)2 (0.304) = 7.73 samples


C+D:
(3.078)2 (800) =


B+D:
4.27
(1000 – 1035)2
(3.078)2 (200) = 18.94
A+C:

7/17/2015

(1000 – 1000)2
(3.078)2 (450) =


(32.62 – 32.01)2
(1000 – 990)2
9th Annual California Unified Program Conference
60
So, fewer samples are
required if,
The waste is essentially homogenous
or
Well above or below the threshold
7/17/2015
9th Annual California Unified
Program Conference
61
Other Types of
Sampling
7/17/2015

Stratified random sampling

Systematic random sampling

Authoritative sampling
9th Annual California Unified Program Conference
62
Stratified random
sampling




Stratified random sampling is appropriate if a batch of
waste is known to be non-randomly heterogeneous.
An example is a pile of blast media. One layer is from
blasting lead paint, the next layer is from blasting new
aluminum parts prior to painting. Another example is a
stripping tank that is used to clean different parts and is
periodically changed. The waste could vary from batch to
batch.
Stratification may occur over space (locations or points in
a batch of waste) and/or time (each batch of waste).
The units in each stratum are numerically identified, and a
simple random sample is taken from each stratum.
7/17/2015
9th Annual California Unified Program Conference
63
Systematic random
sampling




Systematic random sampling, in which the first unit to be
collected from a population is randomly selected but all
subsequent units are taken at fixed space or time intervals.
An example of systematic random sampling is the sampling
along a pipeline at 20 feet intervals.
The advantages of systematic random sampling are the
ease with which samples are identified and collected and,
sometimes, an increase in precision.
The disadvantages of systematic random sampling are the
poor accuracy and precision that can occur when
unrecognized trends or cycles occur in the population.
7/17/2015
9th Annual California Unified Program Conference
64
Authoritative Sampling



Authoritative Sampling - Sufficient information is available to
accurately assess the chemical and physical properties of a
waste, authoritative sampling (AKA judgment sampling) can be
used to obtain valid samples.
This type of sampling involves the selection of sample locations
based on knowledge of waste distribution and waste properties
(e.g., homogeneous process streams). The rationale for the
selection of sampling locations is critical and should be well
documented.
An example is an inspector taking one sample a discarded liquid
that appears to be gasoline (color & odor) to verify that it is
gasoline and has a flash point below 140 °F.
7/17/2015
9th Annual California Unified Program Conference
65
Enforcement Sampling
RCRA Waste Sampling Draft Technical Guidance,
EPA530-D-02-002 (draft), August 2002, page 10 & 11,
{RCRA online # 50940}


2.2.4 Enforcement Sampling and Analysis
The sampling and analysis conducted by a waste handler during
the normal course of operating a waste management operation
might be quite different than the sampling and analysis conducted
by an enforcement agency. The primary reason is that the data
quality objectives (DQOs) of the enforcement agency often may be
legitimately different from those of a waste handler. Consider an
example to illustrate this potential difference in approach: Many of
RCRA’s standards were developed as concentrations that should
not be exceeded (or equaled) or as characteristics that should not
be exhibited for the waste or environmental media to comply with
the standard. In the case of such a standard, the waste handler and
enforcement officials might have very different objectives.
7/17/2015
9th Annual California Unified Program Conference
66
Enforcement Sampling


An enforcement official, when conducting a compliance
sampling inspection to evaluate a waste handler’s
compliance with a “do not exceed” standard, take only
one sample. Such a sample may be purposively selected
based on professional judgment. This is because all the
enforcement official needs to observe – for example to
determine that a waste is hazardous – is a single
exceedance of the standard.
EPA530-D-02-002 (draft), August 2002, Page 11 {RCRA
online # 50940}
7/17/2015
9th Annual California Unified Program Conference
67
Enough on Sampling?

7/17/2015
What about
knowledge of
process?
9th Annual California Unified Program Conference
68
Quick Break

Take 5

Next Speaker John Misleh
7/17/2015
9th Annual California Unified Program Conference
69
Waste Determination
7/17/2015
9th Annual California Unified
Program Conference
70
Waste Determination
by:
Process Knowledge
Knowledge of Process
(KOP)
Generator Knowledge
7/17/2015
9th Annual California Unified Program Conference
71
Waste Determination
CCR 66262.11



(b) the generator may determine that the waste is
not a hazardous waste by either:
(1) testing…; or
(2) applying knowledge of the hazard
characteristic of the waste in light of the materials
or the processes used and the characteristics set
forth in article 3 of chapter 11 of this division.
7/17/2015
9th Annual California Unified Program Conference
72
Waste Determination


CCR 66262.11
Two options:
Process Knowledge
Analysis
7/17/2015
9th Annual California Unified Program Conference
73
Knowledge of Process

Why Use Knowledge
Listed Waste is a function of
how the waste is generated
(knowledge)
 Know that it is Hazardous

7/17/2015
9th Annual California Unified Program Conference
74
Knowledge of Process
OSWER 9938.4-03
RCRA Online 50010
7/17/2015
9th Annual California Unified Program Conference
75
Knowledge of Process
OSWER 9938.4-03
• Process Knowledge • What goes in + contaminants introduced = what
comes out.
• Waste Analysis Data from other facilities
• Old Analytical Data
• A lot of the information that is acceptable to
demonstrate knowledge of process (KOP),
looks a lot like Analytical Data
7/17/2015
9th Annual California Unified Program Conference
76
Knowledge of Process
OSWER 9938.4-03
Process Knowledge
•
•
•
•
•
Material Balances
Engineering Production Data
Material Safety Data Sheets (1% = 10,000 ppm)
Process Kinetic Information and Process Rates
Other Engineering Calculations
7/17/2015
9th Annual California Unified Program Conference
77
Knowledge of Process
OSWER 9938.4-03
Analytical Data From Other Facilities
• Another plant that conducts the same process
and is managing the same waste and has
Analytical Data.
• A TSD that relies on waste analysis Analytical
Data from offsite generators.
7/17/2015
9th Annual California Unified Program Conference
78
Knowledge of Process
OSWER 9938.4-03
Old Analytical Data
• Process and materials must be the same.
• Detection limits and equipment have improved.
7/17/2015
9th Annual California Unified Program Conference
79
Knowledge of Process
OSWER 9938.4-03
Situations where using KOP may be appropriate:
• Constituents are well documented such as for F
or K listed waste
• Wastes are discarded unused chemicals (P & U
listed)
• Health & safety issues in sampling (too
dangerous to sample)
• Physical nature of waste (construction debris)
makes sampling impractical
7/17/2015
9th Annual California Unified Program Conference
80
Knowledge of Process
OSWER 9938.4-03
Conclusion (EPA Guidance Doc)
• Although EPA recognizes that sampling and
analysis are not as economical or convenient as
using acceptable knowledge, they do usually
provide advantages. Because accurate waste
identification is such a critical factor for
demonstrating compliance with RCRA,
misidentification can render your facility liable
for enforcement actions.
7/17/2015
9th Annual California Unified Program Conference
81
Knowledge of Process
Faxback 11918
Conservative Classification
The regulations allow a generator to characterize
its waste based on process knowledge, and it is
understood that generators may at times
characterize their wastes as hazardous
conservatively, rather than incur the costs of
testing every batch or stream.
7/17/2015
9th Annual California Unified Program Conference
82
Knowledge of Process
Faxback 11608
Analyzing Munitions
“not specifically required to test their waste”
“The determination may be made by either applying
knowledge of the waste, the raw materials, and the
process used in its generation or by testing”
“if they think that the above munitions items would fail
the TCLP-extract analysis for lead or dinitrotoluene,
then these wastes could be declared as hazardous,
and no testing would be necessary.”
7/17/2015
9th Annual California Unified Program Conference
83
Knowledge of Process
Faxback 11592, 11579
Limited analytical
Labs being unable to determine
conclusively that the waste is or is not
hazardous . . . It would probably be prudent
for the generator to manage those wastes as
hazardous waste.
7/17/2015
9th Annual California Unified Program Conference
84
Seeking Concurrence with
DTSC? CCR §66260.200
(m) A person seeking Department concurrence with a nonhazardous
determination or approval to classify and manage as nonhazardous a
waste which would otherwise be a non-RCRA hazardous waste shall
supply the following information to the Department:
(5) laboratory results including results from all tests required by
chapter 11 of this division and a listing of the waste's
constituents. Results shall include analyses from a minimum of
four representative samples as specified in chapter 9 of "Test
Methods for Evaluating Solid Waste, Physical/Chemical
Methods," SW-846, 3rd Edition, U.S. Environmental Protection
Agency, 1986 (incorporated by reference in section 66260.11 of
this chapter);
7/17/2015
9th Annual California Unified Program Conference
85
Is ONE sample good
for anything?
• Faxback 11907 - Representative sampling
(Fluorescent Tubes)
“it appears that you tested one spent fluorescent tube to
conclude that all of your spent fluorescent tubes are non
hazardous. . . . Based on one tube, we have no way to
assess the variability between fluorescent lamps. . . A
representative selection of lamps randomly chosen should
be analyzed to make this determination.”
7/17/2015
9th Annual California Unified Program Conference
86
KOP Documentation
OSWER 9938.4-03

“…EPA looks for documentation that
clearly demonstrates that the information
relied upon is is sufficient to identify the
waste accurately and completely.”
7/17/2015
9th Annual California Unified Program Conference
87
KOP Documentation
•
The generator is very familiar with the waste generation
process and the California and Federal hazardous waste
laws and regulations.
•
Detailed chemical information for all the chemicals and
materials utilized in the process is available.
•
A detailed review of the generating process has been
completed and the point of generation has been properly
been identified.
•
All documentation utilized to make the determination is
included in the operating record associated with the waste
stream.
•
The generator has evaluated the information gathered and
made a written determination.
7/17/2015
9th Annual California Unified Program Conference
88
Ten
minute
Break
7/17/2015
89