稀疏矩阵核处理源清单模型操作与大气排放源清单处理模型

需积分: 18 93 浏览量更新于2024-01-02 收藏 18.92MB PDF 举报

身份认证购VIP最低享 7 折!

领优惠券(最高得80元）

本文介绍了两个与环境污染源清单相关的处理模型，分别是稀疏矩阵核处理源清单模型操作和大气排放源清单处理模型。首先是稀疏矩阵核处理源清单模型操作（Sparse Matrix Operator Kernel Emissions, SMOKE）。SMOKE是由北卡罗来纳大学教堂山分校的环境研究所（The institute for the Environment - The University of North Carolina at Chapel Hill）开发的一个用于处理环境污染源清单的软件模型。该模型的最新版本是v4.7，其用户手册于2019年10月份发布。根据版权声明，读者可以复制和分发这份手册，但不能用于商业目的。同时，版权声明也要求在所有副本中包含以上版权声明和段落。 SMOKE模型的主要功能是处理稀疏矩阵核源清单数据。稀疏矩阵是指具有大量零元素和少量非零元素的矩阵。在环境污染源清单处理过程中，源清单数据往往非常巨大，但其中实际有效数据的比例却相对较低，因此使用稀疏矩阵可以有效地减少数据存储和计算的资源消耗。SMOKE模型通过使用核技术进行矩阵运算，能够高效地处理大规模的密集源清单数据，提高计算速度和处理效率。另一个介绍的模型是大气排放源清单处理模型（Emissions）手册。该手册为SMOKE模型的v4.7版本提供了详细的用户指南。根据版权声明，该手册的复制和分发仅限于非商业目的使用，并要求其他副本中包含版权声明和段落。 SMOKE模型和大气排放源清单处理模型是为了帮助环境保护部门和科研人员更好地处理环境污染源清单数据而开发的。通过这些模型，用户可以将庞大的源清单数据转化为易于处理的稀疏矩阵，并进行相关的核运算和分析。这些模型的使用可以提高数据处理和分析的效率，从而为环境管理和决策提供有力的支持。总之，稀疏矩阵核处理源清单模型操作和大气排放源清单处理模型是为了处理庞大的环境污染源清单数据而开发的两个模型。它们的使用可以提高数据处理和分析的效率，为环境管理和决策提供有力支持。

资源详情

资源推荐

Industrial Classification System (NAICS) codes are optional source attributes; the NAICS code is read

by SMOKE but not otherwise used at this time. Additionally, a “source type” field is available in the

nonpoint inventory to identify major and Clean Air Act (CAA) section 112 area sources. See Sec-

tion 2.3.6, “Standard Industrial Classification codes” [15] for a description of SIC codes, Section 2.3.8,

“Maximum Achievable Control Technology codes” [16] for more about MACT codes, and Section 2.3.9,

“Source types: major and section-112 area sources” [16] for more about source types. We will refer to

the CAA section 112 area sources as simply “section-112 area sources”.

• Data: SMOKE can read emissions data for criteria, particulate, and toxics pollutants for nonpoint/sta-

tionary area and nonroad inventories. The SMOKE system is not constrained with regard to the pollutants

read (although typical examples were given in Section 2.3.1, “Inventory data types” [7]). SMOKE

accepts annual emissions data, average-day emissions data, or both (though not all input formats support

all types). An emission factor value can also be read by SMOKE, but SMOKE does nothing with it.

• Data attributes: Inventories for nonpoint/stationary area and nonroad mobile sources can contain

control efficiency, rule penetration, and rule effectiveness information for each pollutant. SMOKE will

use these data if provided; otherwise it will set default values that indicate that no control-based adjust-

ments have been applied to the inventory pollutant data. The defaults are listed in the file formats in

Chapter 8, SMOKE Input Files [343].

2.3.2.2.2. On-road mobile (SMOKE category: mobile)

• Source characteristics: For on-road mobile inventories, the minimum source characteristics that

identify these sources are country/state/county code and either SCC or road class and vehicle type

codes. When the SCC is provided, it must follow a specific pattern in order to contain the road class

and vehicle type codes (see Section 2.8.4.1, “Special approach for on-road mobile MOVES SCCs” [37]).

When road class and vehicle type codes are provided to SMOKE directly, SMOKE translates these to

SCC values.

• Optional source characteristics: A link code may also identify on-road sources. This code must be

unique within each county and SCC (or road class/vehicle type combination).

• Source attributes: The inventory year is associated with all sources in the inventory input files. In

addition, SMOKE assigns a time zone (see Section 2.9.14, “Assign country codes, years, and time

zones” [57]) and an approach for normalization of temporal profiles (see Section 2.9.13, “Set the

weekday averaging approach” [57]). For sources with link codes, SMOKE will use the starting and

ending coordinates of the link, using either latitude-longitude (lat-lon) values or coordinates in the

Universal Transverse Mercator (UTM) coordinate system.

• Data: Emissions data for criteria, particulate, and toxics pollutants can be read for on-road mobile in-

ventories. SMOKE is not constrained with regard to the pollutants read (although typical examples

were given in Section 2.3.1, “Inventory data types” [7]). SMOKE accepts annual emissions data, av-

erage-day emissions data, or both (though not all input formats support all types).

Additionally, on-road mobile inventories can contain VMT and average speed activity data, which are

needed when users would like SMOKE to run MOVES to compute emissions. A combination of pre-

computed emissions and VMT data is also acceptable for input to SMOKE, but you are responsible

for preventing duplication of emissions. Duplication could occur if you input precomputed emissions

for the same sources that you use SMOKE to compute the emissions on the fly, by multiplying the on-

road emissions factors from MOVES by hourly VMT, and the off-network emission factors from

MOVES by annual vehicle populations.

• Data attributes: No data attributes are associated with on-road mobile sources.

2.3.2. Inventory source categories

2.3.2.2.3. Point sources (SMOKE category: point)

• Source characteristics: The source characteristics for point sources depend on the inventory input

format. The SMOKE one-record-per-line (ORL) and Flat File 10 Format (FF10) formats identify

sources by country/state/county code, plant code, point code, stack code, segment code, and SCC. 95.

• Optional source characteristics: SMOKE can support up to five location identifiers within a plant,

although the most used in any currently implemented input file format is four.

• Source attributes: As with other source categories, inventory year is associated with all sources in

the inventory input files. SMOKE also assigns a time zone (see Section 2.9.14, “Assign country codes,

years, and time zones” [57]) and an approach for normalization of temporal profiles (see Section 2.9.13,

“Set the weekday averaging approach” [57]). In addition, point sources have the following required

source attributes not associated with other source categories: latitude, longitude, stack height, stack

diameter (at the exit location), flue gas exit velocity, and flue gas exit temperature. Finally, the following

optional source attributes are also used by SMOKE: SIC codes, MACT codes, plant descriptions,

emissions release type point (e.g., horizontal stack, fugitive), source type (major or section-112 area),

Office of Regulatory Information Systems (ORIS) identification codes, and boiler identification codes.

Also, the MACT code and source types are supported only by the ORL format. See Section 2.3.6,

“Standard Industrial Classification codes” [15], Section 2.3.8, “Maximum Achievable Control Techno-

logy codes” [16], and Section 2.3.9, “Source types: major and section-112 area sources” [16] for more

information.

• Data: Emissions data for criteria, particulate, and toxics pollutants can be read for point inventories.

SMOKE is not constrained with regard to the pollutants read (although typical examples were given

in Section 2.3.1, “Inventory data types” [7]). SMOKE accepts annual emissions data, average-day

emissions data, or both.

Optionally, point-source emissions data can be provided using day-specific or hour-specific records.

The formats for these data are described in Section 8.2.6, “PTDAY: Point source day-specific emis-

sions” [365] and Section 8.2.7, “PTHOUR: Point source hour-specific emissions” [367].

• Data attributes: EGU and non-EGU point sources can contain control efficiency and rule effectiveness

information for each pollutant. SMOKE will use these data if provided; otherwise it will set default

values that indicate that no control-based adjustments have been applied to the inventory pollutant

data. The defaults are listed in the file formats in Chapter 8, SMOKE Input Files [343].

2.3.2.2.4. Wildfire with precomputed plume rise (SMOKE category: point)

• Source characteristics:Wildfires with precomputed plumes are identified by the country/state/county

code and the fire name.

• Optional source characteristics: There are no optional source characteristics for wildfire sources.

• Source attributes: Like other source categories, inventory year is associated with all sources in the

inventory input files. SMOKE also assigns a time zone (see Section 2.9.14, “Assign country codes,

years, and time zones” [57]) and an approach for normalization of temporal profiles (see Section 2.9.13,

“Set the weekday averaging approach” [57]). In addition, wildfire sources require the latitude and

longitude source attributes. Finally, additional hour-specific source attributes for wildfire sources must

be provided for the fraction of emissions in the surface layer, the height of the bottom of the plume,

and the height of the top of the plume. These hour-specific attributes are provided to SMOKE using

the point source hour-specific formats described in Section 8.2.7, “PTHOUR: Point source hour-specific

emissions” [367].

2.3.2. Inventory source categories

• Data: Wildfire source inventories can contain criteria and particulate pollutants. SMOKE is not con-

strained with regard to the pollutants read (although typical examples were given in Section 2.3.1,

“Inventory data types” [7]). These data must be provided as day-specific or hour-specific emissions

values using point source formats specified in Section 8.2.6, “PTDAY: Point source day-specific

emissions” [365] and Section 8.2.7, “PTHOUR: Point source hour-specific emissions” [367].

• Data attributes: No data attributes are associated with wildfire sources.

2.3.2.2.5. Wildfires with internal plume rise calculation (SMOKE category: point)

• Source characteristics: Wildfires with internal plume rise calculation are identified by the coun-

try/state/county code, fire identification, fire name, location identification, and SCC.

• Optional source characteristics: There are optional source characteristics for fire sources, such as

material burned, vegetation types, size of area burned, fuel loading, and fire start/end hour. The size

of area burned and fuel loading are used for computing the fire-specific plume rise. Fire starting and

ending hours are needed to adjust the hourly temporal profiles for the emissions.

• Source attributes: Like other source categories, inventory year is associated with all sources in the

inventory input files. SMOKE also assigns a time zone (see Section 2.9.14, “Assign country codes,

years, and time zones” [57]) and will re-normalize temporal profiles based on the starting and ending

hours of the fire. In addition, wildfire sources require the latitude and longitude source attributes to

locate the fire. Note that all emissions for a fire will be assumed to come from the single grid cell that

contains the latititude and longitude of the fire. Finally, additional day-specific source attributes listed

above for fire sources must be provided for calculating the heat flux of each fire, which is used to es-

timate the fraction of emissions in the surface layer, the height of the bottom of the plume, and the

height of the top of the plume. See more information about how to process at Section 4.4.17, “Plume

Rise Calculation for Fires” [161]

• Data: Fire source inventories can contain criteria and particulate pollutants. SMOKE is not constrained

with regard to the pollutants read (although typical examples were given in Section 8.2.8.3, “ORL

FIRE Format” [381]). These data must be provided as day-specific emissions values using point source

formats specified in Section 8.2.6.2, “ORL FIREEMIS Emission Format” [366].

• Data attributes: No data attributes are associated with wildfire sources.

2.3.2.2.6. Biogenic land use (SMOKE category: biogenic)

• Source characteristics: Biogenic emission data does not fit as neatly into the source-characteristic

paradigm as the previously described source types. Emissions for biogenic sources are estimated

starting with land use data, which are available for both BELD2 and BELD3 processing. The BELD2

data are available either by U.S. state/county and BELD2 land use category or by grid cell and BELD2

land use category. BELD3 land use data are available by 1-km grid cell over North and Central

America and by BELD3 land use category.

• Optional source characteristics: Biogenic land use data do not include optional source characteristics.

The data are either by state/county or by grid cell.

• Source attributes: There are no source attributes for biogenic land use data.

• Data: The biogenic land use data consist of fractions associated with each land use type within a county

or grid cell.

• Data attributes: There are no data attributes for biogenic land use data.

2.3.2. Inventory source categories

2.3.3. Inventory file formats

SMOKE supports a variety of inventory formats for criteria, particulate, toxics, and activity data inventories,

which are described in detail in Section 8.2, “Inventory Files” [349]. Here, we provide a brief introduction

to these formats, which will be helpful as you read more about SMOKE in the remainder of this chapter

and the chapters before Chapter 8, SMOKE Input Files [343]. All formats described here are text files. To

convert your data to these formats, the best approach is to use a database or spreadsheet program to reformat

and output the data in the requested format. There is not a standard format-conversion method that comes

with SMOKE.

In the following paragraphs, we describe the formats available for nonpoint/stationary area, nonroad mobile,

on-road mobile, point, and point-wildfire sources.

• Nonpoint/stationary area sources: SMOKE supports two formats for nonpoint/stationary area sources.

The ORL and FF10 (Flat File 10) format are list directed (comma or semicolon delimited) and these

file formats may be used to represent many different sources. The header of the file indicates what

source data are in the file.

• Nonroad mobile sources: There are three available inventory formats for nonroad mobile sources.

The FF10 (Flat File 10) format is list directed (comma or semicolon delimeted) and the header of the

file is used to indicate the nonroad mobile source data is within the file.

• On-road mobile sources: The Flat File 10 (FF10) format is list directed (comma or semicolon delimited)

and contains activity inventory such as VMT, speed, and vehicle population data. This format requires

VMT, SPEED, and VPOP inventory data.

• Point sources: SMOKE has formats for annual or average-day inventories, for day-specific inventories,

and for hour-specific inventories. For annual or average-day inventories, the ORL and FF10 formats

can be used for criteria, particulate, and toxics inventories. Finally, the CEM data format can be used

for day-specific or hour-specific data : SMOKE uses the ORIS codes and boiler codes in the annual

inventory files to match sources from the CEM data files.

• Wildfire sources: There are two approches available that you can provide wildfire data that are being

treated as point sources to SMOKE using the ORL and FF10 point-source formats.

• Precomputed plume rise approach: Certain fields must be left blank (such as stack parameters)

because they do not apply to wildfire sources. When using wildfire data provided as point sources,

you must also provide day-specific or hour-specific wildfire emissions and hour-specific precomputed

plume rise using the FF10 day-specific and hour-specific formats.

• Internal plume rise calculation approach: Requires two separate inventory files that are provided

in a modified ORL format: (1) a list of fires with fire-specific characteristics including

county/state/county, fire identification, location coordinate, fire name, SCC and others, as described

in Section 8.2.8.3, “ORL FIRE Format” [381], and (2) a day-specific fire data including size of area

burned, fuel loading, and star/end hour of fire (Section 8.2.6.2, “ORL FIREEMIS Emission

Format” [366]). Unlike the approach listed above, this approach internally estimates the plume rise

using the size of the area burned and fuel loading, and it adjusts temporal profiles using the start

and end hours of the fire. See detail at Section 4.4.17, “Plume Rise Calculation for Fires” [161]

2.3.4. Country, state, and county codes

SMOKE uses a 6-digit integer code to identify a country, state (or province), and county (or other region)

for a particular source. Most U.S. inventories input to SMOKE have the 5-digit U.S. Federal Implementation

Planning Standards (FIPS) state and county codes. All inventory input formats have been adapted to include

a special header record with which you can specify the country, effectively allowing the inventories to be

provided with the 6-digit code that SMOKE uses. The 6-digit system was designed for use in the United

2.3.3. Inventory file formats

States with states and counties, as well as Canada and Mexico, but it can be adapted for other uses. The

format used by SMOKE for the codes is:

The SMOKE installation is set up to use U.S.-centered codes as defined in the COSTCY or the GEO-

CODE_LEVEL4 (if USE_EXP_GEOCODES Y) file, which contains the codes and their associated names

and time zone settings. In this file, the U.S. country code is zero, which allows the U.S. country/state/county

codes to be the same as the FIPS state/county codes that appear in U.S. inventories. See Section 8.10.2,

“COSTCY: Country, state, and county names and data” [448] for more information on the COSTCY file

format.

To change the meaning of the country, state, or county codes in SMOKE, the COSTCY or the GEO-

CODE_LEVEL4 (if USE_EXP_GEOCODES Y) file must be modified to use different names associated

with each country, state, county or tribal number. All SMOKE input files must also use this new numbering

scheme, including inventory files and cross-reference files.

Acceptable values in SMOKE for the country code are 0 through 9. Acceptable values of the state code

are 1 through 99. Acceptable values of the county code are 1 through 999. No alphabetic codes are accepted,

since SMOKE stores these values as integers.

2.3.5. Source Classification Codes

EPA uses Source Classification Codes (SCCs) and area and mobile source (AMS) codes to classify different

types of anthropogenic emission activities. SCCs have 8 digits for point sources, while AMS codes have

10 digits, and sometimes include a leading “A” as an eleventh character. In SMOKE, we refer to both kinds

of codes as “SCCs”, and we ignore the leading “A” in the area and mobile codes. Additionally, SMOKE

permits the nonpoint and point toxics inventories to use both 8-digit and 10-digit SCCs in the same inventory

input file, because both 8- and 10-digit codes are contained in the nonpoint and point inventories in the

1999 NEI for HAPs. The maximum field width in SMOKE and its input files for SCCs is 20 characters as

of SMOKE v4.0. The 8 or 10 digit SCC are still supported, but if the SCC is greater than 10 digits the SCC

hierarchial approach will not be supported.

For SCC's of size 10 characters or less, the codes use a hierarchical system in which the definition of the

code gets increasingly more specific as you move from left to right. (NOTE: if the SCC is greater than 10

characters in length the hierarchial system is not used). For SCC's of 10 characters or less, it is important

to understand the hierarchy of the codes, because you can take advantage of the hierarchy in building cross-

reference files for assigning emissions processing factors to inventory emission sources. In the diagrams

below, level 1 is the least specific and level 4 is the most specific.

The code structure for the 8-digit point-source codes is:

An example point-source activity and corresponding SCC can be taken directly from SMOKE’s SCC de-

scription file (SCCDESC): “External Combustion Boilers; Electric Generation; Lignite; Spreader Stoker”

2.3.5. Source Classification Codes

剩余517页未读，继续阅读

ccz007

粉丝: 9
资源: 37

会员权益专享

稀疏矩阵核处理源清单模型操作与大气排放源清单处理模型

稀疏矩阵的操作

稀疏矩阵操作

稀疏矩阵的相关操作

稀疏矩阵与稀疏矩阵相乘

SciPy中的稀疏矩阵处理

华为od机试 稀疏矩阵

利用稀疏矩阵的链式存储实现稀疏矩阵的加，减，乘，转置等简单运算

torch稀疏矩阵索引

matlab稀疏矩阵存储

1. 构建稀疏矩阵M。 2. 求稀疏矩阵M的转置矩阵T。 3. 输出稀疏矩阵M和稀疏矩阵T。

matlab 稀疏矩阵 求逆

python稀疏矩阵密度收缩

写一篇关于稀疏矩阵的论文

matlab 稀疏矩阵连线

在python中，稀疏矩阵是奇异的矩阵，应该如何求这类稀疏矩阵的逆

torch 稀疏矩阵表达

指针稀疏矩阵行列互换

python 稀疏矩阵乘法

使用mkl迭代求解稀疏矩阵

CSR稀疏矩阵是树结构吗

会员权益专享

最新资源

华为od机试稀疏矩阵

matlab 稀疏矩阵求逆