Вы находитесь на странице: 1из 16

Meaningful Energy Efficiency Performance Metrics for the Process Industries

Jimmy D Kumana, MS ChE Noshir R Sidhwa


CEO Business Development Manager
Kumana & Associates, Houston, Texas Kumana & Associates, Houston, Texas
jkumana@aol.com nsidhwa@sbcglobal.net

ABSTRACT

An effective energy performance benchmarking should include a consideration of production


rate, product specifications, feedstock mix, and process type, in addition to thermodynamics and
economics. Unfortunately, there is no accepted industry standard for developing Energy
Efficiency (EE) performance metrics for the chemical process industries, and published literature
on the subject is extremely sparse.

This paper will present a comprehensive system of EPIs as applied in a complex multi-product
multi-plant organization in the oil and gas industry. Four categories of EPIs are recommended:
• By equipment
• By process unit
• By product
• By business unit

It will be shown how each type of EPI fulfils a specific business objective in the organization.
Successes and failures are described, and recommendations are provided. The principles and
practices outlined in this paper are generally applicable, and will hopefully lead to a standard
methodology for EE performance reporting.

INTRODUCTION

In recent years, Key Performance Indicators (KPIs) have become a popular tool for monitoring
and managing how well an organization is doing in meeting its stated objectives [1, 2, 9, 10, 11].
Energy Performance Indices, or EPIs, are one such KPI. In this paper, the authors describe their
recommended approach to developing a set of EPIs tailored to the company’s needs, and how
they have been effectively used in a large multi-plant oil & gas company [4].

EPIs can have several different applications, each of which requires a different formulation:
• Report card, for information only
• Benchmarking (historical, competitive, or absolute)
• Economic Dispatch
• Process Improvement, including troubleshooting
• Operations optimization

In general, for a manufacturing company with multiple plants, one would need 4 different types
of EPIs, depending on the application:

Industrial Energy Technology Conference, New Orleans (May 2009) Page 1 of 16


Type/Name Application
Corporate energy KPI Report card for CEO’s dashboard
Product EPIs Dispatching production among similar plants
Process Unit EPIs Benchmarking, process improvement
Equipment EPIs Operations optimization

While simple in concept, the actual implementation of a company-wide performance measure-


ment program is usually fraught with challenges, many of which are often political and cultural
rather than technical [6].

A good KPI should have the following characteristics [7]:


1. Direction of movement is consistent with change in actual performance
2. Magnitude of movement is consistent with of change in actual performance
3. Provides actionable information
4. Performance criterion is related to output (results) rather than input (resources expended)
5. Should be automated for online display and reporting.

A college economics textbook by Paul Samuelson provides an actual case study of what happens
when performance metrics are not formulated correctly:

In the former Soviet Union, steel mill performance was measured by how much steel was
processed in the facility, and the manager was rewarded on the basis of capacity utilization,
measured as “tons processed”. Now it so happened at one remote mill that, there was a chronic
shortage of raw material supply due to transportation problems, and so the mill was operating
well below its capacity. In his performance review, the manager was penalized for falling below
the performance target, although it was due to no fault of his. His solution to improve the KPI
was to re-process some of the mill’s product (output) using the full amount of labor and fuel. His
KPI improved and he was rewarded accordingly, but his personal gain occurred at the expense of
society’s loss.

It is vitally important to ensure that performance targets and KPIs do not create perverse
incentives.

EQUIPMENT EPIs

Common equipment used in process plants can be broadly classified into two groups:

Energy Consumers Energy Converters


Pumps Boilers
Compressors Steam turbines
Process heaters (steam or direct fired) Gas turbines
Dryers – steam or direct fired Internal-combustion reciprocating engines
Distillation columns (in reboilers) Electric motors
Misc rotating machinery Electric generators

Industrial Energy Technology Conference, New Orleans (May 2009) Page 2 of 16


Of the various types of EPIs, Equipment EPIs are the simplest and easiest to formulate. The
proper measure of energy efficiency for energy consumers is:
Useful Energy Absorbed E2 − E1
Efficiency = =
Utility Energy Supplied ES

The proper measure of energy efficiency for energy converters is:


Energy Output E B
Efficiency = =
Energy Input EA

Figure 1. Energy Flows through (a) Consumers and (b) Converters

Alternative formulations of equipments EPIs are as an “intensity” or as an “approach to target”.


Both are acceptable, depending on which formulation is more useful.

Figure 2. Alternative Equipment EPI Formulations

Taking the example of a process pump, the overall efficiency (for a Newtonian fluid and
assuming the driver is a 3-phase induction motor) is given by:

Q (gpm) x [P2 − P1 ] (psi)


Pump η =
3.982 I.V. cosφ .η M

where Q = process fluid flow, gpm


P1, P2 = suction and discharge pressures, in psia or psig
I = measured current, amperes
V = applied voltage (known)
cos(φ) = power factor, expressed as a decimal, typically 0.85-0.92
ηM = motor efficiency (from data sheet), typically 94-97%

Industrial Energy Technology Conference, New Orleans (May 2009) Page 3 of 16


Normally V is known, but I is not, unless the motor is large enough to be included in the plant’s
power monitoring system. Therefore on-line monitoring of pump efficiency is generally feasible
only for larger sizes.

An alternative off-line diagnostic tool to assess pump performance and mechanical condition is
to plot the actual pump characteristic curve (flow rate versus ∆P) and compare it against the
design curve. If the pump is delivering less TDH (total discharge head) than design at the
measured flow, it means that the performance has fallen off and needs corrective action. The
efficiency loss can be estimated very roughly as the ratio of actual head to design head at that
particular flow rate.

For compressors, use the adiabatic (isentropic) efficiency as a KPI for process efficiency, and the
polytropic efficiency as an indicator of mechanical condition. Also keep in mind that although it
may be an excellent KPI for overall process efficiency, power consumption is not necessarily a
good indicator of compressor equipment efficiency.

As an example, at a gas processing plant, total compression power was reduced drastically (50%)
by making relatively simple piping and controls modifications. The overall process efficiency
improved dramatically, even though some of the individual compressor efficiencies actually
dropped slightly (due to operation at off-design conditions).
Trend line for compressor power consumption
18000

16000
14000

12000

10000
8000

6000

4000

2000

Figure 3. Compressor Station Power Consumption at Constant Process Conditions

For fired heaters, the fuel efficiency is most simply calculated by the heat balance method, also
known as the “direct” method:
W .( H 2 − h1 )
η=
F .HHV

where W = process fluid flow rate in lb/h; h1, H2 = inlet and outlet enthalpies of the process
stream, Btu/lb; F = fuel flow rate; and HHV = higher heating value (in Europe, the lower
heating value is preferred).

While simple in concept, the heat balance method has a major deficiency – it gives accurate
results only when the data quality is extremely good (generally less than 1% error), which is

Industrial Energy Technology Conference, New Orleans (May 2009) Page 4 of 16


hardly ever the case. A more accurate estimate of furnace efficiency can be obtained by the heat
loss method (also known as the “indirect“ method) as follows:

Absorbed duty F .HHV − Losses


η= or alternatively η =
Absorbed duty + Losses F .HHV

Choose the formulation according to which measured value (absorbed duty or fuel input) is
likely to be more accurate.

For energy conversion devices, the formulation is slightly different, even though there could be
several variations, as illustrated by the example of single-exhaust back-pressure turbine.

HP Steam

WORK
BPST

LP Steam

PROCESS

Figure 4. Schematic of Back-Pressure Steam Turbine

Here again, two kinds of efficiency are applicable. One is the isentropic efficiency of the turbine,
which is an indicator of how well the machine was designed to begin with, and its present
mechanical condition. This “machine” efficiency is calculated as:
H1 − H 2
η=
H1 − H 2 '
where H1 = Enthalpy of HP inlet steam, Btu/lb
H2 = Actual enthalpy of exhaust LP steam, Btu/lb
H2’ = Enthalpy of exhaust LP steam assuming isentropic expansion, Btu/lb

The other kind of efficiency is the overall or “cycle” energy efficiency, which is calculated as a
ratio of the useful energy output as a fraction of the energy input:

3413 x kw + W.H 2
η=
W.H 1

Since enthalpy cannot be measured directly, it must be inferred from pressure and temperature
measurements using an electronic steam properties database. Some plants have “extraction”

Industrial Energy Technology Conference, New Orleans (May 2009) Page 5 of 16


steam turbines with two exhausts – at medium and low pressure. The thermodynamic efficiency
is calculated for each stage, and the overall efficiency is derived from the stage-efficiency
results.

The isentropic efficiency is of more practical use from an operational viewpoint, because it
provides a warning of developing mechanical problems. The cycle energy efficiency is more
useful for design and decision-making purposes, such for calculating the process unit energy
balance when conducting plant energy audits or choosing between project alternatives.

PROCESS EPIs

EPIs for Process Units and Process Areas are intended to measure the energy efficiency of an
entire processing system, rather than single items of equipment. Examples of process units in an
oil refinery are the Crude Distillation, Fluid Catalytic Cracking, Reforming, Hydro-treating, etc.
Examples of process units in gas processing plants are Sweetening, Dehydration, and Sulfur
Recovery. Examples of process units/areas in a Kraft pulp mill are Woodyard, Cooking, Pulp
washing, Bleaching, and Chemical Recovery (could be further broken down into sub-units such
as evaporation, recausticizing, etc).

Some industries have developed standard metrics for their plant performance. A notable example
is the Solomon Energy Intensity Index (EII) for Oil Refining, which builds up the overall plant
energy index from the energy indices for individual process units. All that each refinery has to do
is ensure that the data being used are accurate, and then link them into the EII calculation
spreadsheet for each process unit to automatically display/print trend charts.

Although attempts have been made to formulate a similar Energy Efficiency Index (EEI) for the
gas processing industry, they have not been widely accepted, largely because the simplistic
approaches used do not yield actionable information. For other process industries, there is no
established standard at all for computing usable energy indices. The methodology described in
this paper was developed specifically to fill this glaring unmet need. It is generally applicable to
any type of process plant and for any industrial sector.

In general, each process unit or area will have only two energy indices: a fuel index and a power
index:
FuelConsumption PowerConsumption
Fuel.Index = and Power.Index =
CapacityParameter CapacityParameter

In cases where the process unit uses both steam and direct fuel, it may be desirable to have three
indices for added insight – a process fuel index, a steam index (= boiler fuel index), and a power
index. The energy consumed in the Utilities area of the plant and by common facilities (eg admin
buildings, perimeter lighting) must be properly allocated to the process units. The consumption
indices could also be combined into a single energy cost index.

The “process area” indices are designed to monitor the energy efficiency of groups of process
units that together produce recognizable intermediate products or perform a value-adding
function on the feed – and whose capacities can be measured by the flow rate of those products

Industrial Energy Technology Conference, New Orleans (May 2009) Page 6 of 16


or feeds. A single “capacity parameter” should be chosen to represent the operating rate of each
process area. For a typical gas-processing plant (Figure 5) suitable capacity parameters for the
various units might be: combined feed gas flow, hydrocarbon condensate flow, sweet wet gas,
sour off gas, product gas, etc.

The recommended procedure is as follows:


• Divide the plant into major process areas (as illustrated in Figure 5) which have clearly
defined feeds coming from other process units and clearly defined intermediate “products”
that go to other units. These process areas may include more than one process unit, or may
be only part of a process unit.
• Use feed and product data (flow rates, compositions, temperatures, pressures) and flow
rates + concentrations of key intermediate streams from the plant information (PI) system
to run an overall plant HMB simulation model. Extract the capacity parameters.
• Extract the steam and fuel consumption calculated by the simulation model. Convert steam
usage to proportional fuel usage in the boilers. Add the fuel for steam plus direct fuel use
in the process unit (area) to get the total fuel consumption. Divide this number by the
selected capacity parameter to get the Fuel Index.
• HMB models typically do not currently calculate the plant power consumption. A simple
but approximate method to get around this problem is as follows. (a) Estimate the net total
plant power consumption (= import + generation - export) and call this X. (b) Calculate
the power consumption of the major compressors and pumps from PI data using the
equipment spreadsheet models. Call this Y. (c) Allocate the balance of power usage (X-Y)
megawatts to each process unit based on installed HP of non-major running equipment.
(d) Add the allocated and calculated power consumption for each unit and use the total to
compute the Power Index.

Figure 5. Major Process Areas/Units for a Typical Gas Processing Plant

Industrial Energy Technology Conference, New Orleans (May 2009) Page 7 of 16


Overall process area EPIs following this approach are presented in Table 1. These indices
clearly indicate which units are the major energy consumers, pinpointing specific areas of the
plant that need attention both in terms of operating problems and opportunities for major cost
reduction; in short they provide actionable information.

Table 1. Process Area EPIs for a Gas-Processing Plant

Consider the process EPI trend chart shown for the Amine Gas Treating Unit over a 3-month
period. What does it tell us? During the first two months steam usage is relatively stable,
although a slight uptrend is discernible. Then around the beginning of the third month, it starts to
rise rapidly. The area process engineer came up with three possible explanations:
• Faulty flow meter
• Defective/damaged control valve
• Colder feed to stripper (eg. due to HX fouling)

Upon investigation, it was discovered that there was a small but growing steam leak across one
of the stripper feed pre-heater tubes. Without this process EPI, and the ability to interpret it
correctly, the problem would not have been so readily discovered.

Figure 6. 3-month Trend Line for Steam Usage in the Gas Treating Unit

Process EPIs can also provide immediate feedback on process modifications. The EPI being used
for a pumping station composed of multiple pumps of different capacities and a mix of driver
types (Figure 7) was total pumping cost, in $ per 1000 gal of fluid transferred. Upon
implementing a proposal to change the control strategy, viz. to run the fixed-speed motors in
on/off mode and the turbine-driven pump in variable-speed mode, the Pumping Cost Index
immediately dropped by about 30% (Figure 8), providing quantitative proof of the benefit from
this recommendation.

Industrial Energy Technology Conference, New Orleans (May 2009) Page 8 of 16


Figure 7. Schematic Diagram of Pumping Station in an Oil Refinery

Pumping Cost Index


24

Previous
20 Optimized

16
cents/ Kgal

12

0
11/9 12/29 2/17 4/8 5/28 7/17 9/5 10/25 12/14 2/2

Day of year

Figure 8. Pumping Process EPI Proves Benefit of Proposed New Operating Policy

By the same token, if the index rises, it would indicate a potential problem such as fluid
recirculation, mechanical failure within the pump, or some sort of obstruction in the pipe.

PRODUCT EPIs

The principal purpose of product EPIs is not process improvement, but cost accounting, and as a
planning tool for “economic dispatch” (which means assigning production among multiple plants
on the basis of economic criteria). Incidental benefits include the ability to accurately compare

Industrial Energy Technology Conference, New Orleans (May 2009) Page 9 of 16


production economics at one site versus another (competitive benchmarking), and to track
production cost trends (historical benchmarking).

A common practice is to express the Index as an Energy Intensity, eg. Btu/ton of product. It finds
favor among management for its simplicity and ease of calculation, and allows mid-managers to
satisfy senior management’s demand for KPI reporting, but it does nothing at all to give insight
into the energy efficiency of the company’s production facilities, and is completely worthless for
decision-making purposes.

Consider this. If a company has two plants in different geographic regions – using different
grades of raw materials, making different grades of the same product using different techno-
logies, and with different energy prices, would it make any sense to compare their energy
efficiencies? The obvious answer is a resounding NO. Benchmarking metrics only make sense if
they are formulated to compare like with like.

Here’s a real life example: A couple of oil refinery managers were resisting the deployment of
product-based EPIs because of their “complexity”. We demonstrated that when the refinery yield
of high-value distilled products was improved (very profitable) at the expense of additional
energy, their energy intensity computed simplistically as Btu/MB actually got worse, which is
directionally inconsistent with the right operational decision. The EPI formulation recommended
by K&A, however, was directionally consistent.

Waste W

Product A

Feed F PROCESS Inter- PROCESS


Byproduct B
UNIT 1 mediate UNIT 2

E1 E2
Energy
input E

Figure 9. Schematic Model for Product EPI Value-Added Allocation Procedure

The basic concept underlying correct formulation of Product EPIs is to properly allocate energy
consumption in each of the process units to the appropriate product. There are many possible
allocation strategies – by volume, by market value, by Btu content, by value-added, etc. All of
them are problematic, in the sense that none of them exactly meet the criteria of directional and
magnitude consistency. However, after testing them all under several real-world scenarios, the

Industrial Energy Technology Conference, New Orleans (May 2009) Page 10 of 16


“value-added” method was found to yield the best overall results in terms of providing insights
and actionable results [3]. Furthermore, it is more consistent and in alignment with most
common formulations financial and other corporate KPIs.

Step one, as always, is to produce simplified Process Flow Diagrams (PFDs) and a validated
Heat and Material Balance simulation model.

The mathematical model of the plant required to calculate the EPI does not necessarily follow
the “process unit” modular structure. Rather, the goal is to set up a sequence of modules that are
punctuated by splits and mixes in various intermediate streams in such a way that they can be
easily attributed to a specific product. In some cases, therefore, we may group multiple process
units together into a module, and in other cases, a process unit may be sub-divided into a number
of sub-systems, or even individual items of equipment.

Step two is to apply the “Value-Added” allocation procedure to calculate the overall product
EPIs using the equations below, according to the procedure described in Reference [7].

⎛ VA ⎞ ⎛ VB ⎞
E A = ⎜⎜ ⎟⎟.E; E B = ⎜⎜ ⎟⎟.E
⎝ VA + VB ⎠ ⎝ VA + VB ⎠

where Vi = gross value added to the business by manufacture of product i, and Ei = amount of
energy allocated to product i. The energy intensity indices are EA/A and EB/B.

The methodology, in its complete form, is generally applicable to more complex scenarios with
multiple feeds, multiple sources of energy, and even money-losing byproducts.

Figure 10. Product EPIs for Two Grades of Liquid Butane

Figure 10 shows the product EPIs for liquid butane product that is produced in two grades. One
is pipeline grade (non-refrigerated) and the other is export grade (refrigerated for export by ship).

Industrial Energy Technology Conference, New Orleans (May 2009) Page 11 of 16


For more than 20 years since plant startup, both grades were priced exactly the same. It was only
when the company went through the exercise of developing EPIs that they realized how much
more expensive it was to produce export-grade product. One recommendation K&A made was to
reduce energy costs by minimizing the residence time of product in the storage tanks through better
inventory management. Another was to adopt differential pricing for the two grades. Such actionable
insights could not have been gained from the feed-based indices that were being used previously.

OVERALL CORPORATE ENERGY KPI

A well-known international management consulting firm was retained by our client to create a
complete “dashboard” of corporate KPIs, with energy being one of them. They conducted a
world-wide survey of 29 major companies in the oil and gas industry which revealed that almost
all of them express their overall corporate energy index as an Energy Intensity, viz. Btu/MBoe.
Based on this finding, they recommended the following corporate energy KPI:

While this may be convenient and easy to calculate using readily available data, it is a very poor
indicator of whether the company is becoming more or less efficient, and useless for comparing
the company’s energy efficiency against it peers. The reason is that each company has many
different types of operations, each with a different energy intensity. The energy intensity for each
plant is affected by several uncontrollable external factors such as field aging, weather, etc.
Furthermore, the composite corporate index comprises a volume-weighted average of all the
different products/technologies. Over a period of several years, the company’s business shifted
towards more oil refining in proportion to crude oil production, and to heavier (cheaper) grades
of crude oil feedstock that require more energy to refine. Thus their corporate energy index (as
formulated above) rose even though they had a highly effective energy conservation program
(see Figure 11 for individual business lines), and company operations actually become more
profitable overall. In short, this simplistic index failed to meet the “directional consistency” test.

To ensure directional consistency, the energy KPI should be formulated as a ratio, eg:

Energy Efficiency Index (EEI) = 100 * {Actual Energy Cost/Standard Energy Cost}

N
where Actual Energy Cost = ∑ (energy consumption ) (energy price)
i
i i

N
Standard Energy Cost = ∑ (standard consumption) (energy price)
i
i i

and i = type of energy used, viz. gas, oil, power, etc.

Industrial Energy Technology Conference, New Orleans (May 2009) Page 12 of 16


Energy Intensity, Oil & Gas Industry
70
Oil & Gas prod'n
60 Oil Refining
Gas Processing
50
Cents/BOE
40

30

20

10

0
1999 2000 2001 2002 2003 2004 2005 2006

Figure 11. Energy Intensity Trends for Different Business Units

The “standard energy consumption” depends upon a number of variables, principally throughput,
but also production volumes, feedstock composition, product mix, product specs, fuel mix,
technological advances, and ambient temperature. Ideally, it should be calculated from first-
principles using rigorous simulation models, although shortcut approximations may be used on
an interim basis to get started.

The calculated actual corporate EEIs for the cited example are shown as blue squares (years
2000-2006) in the Figure 12. The target EEI trajectory is shown as green triangles. In this
formulation, the lower the EII, the better. The difference between the actual EEI and 100 is a
measure of the % savings achieved by the energy efficiency improvement initiatives.

Figure 12. Correctly Formulated Corporate EEI Gives Actionable Information

Industrial Energy Technology Conference, New Orleans (May 2009) Page 13 of 16


The company’s stated goal was to reach an overall corporate EEI of 50 by 2010, from a baseline
of 100 (by definition) in 2000. The chart indicates that as of end 2006 (we do not have data for
2007-08) the company was ahead of schedule in meeting this goal, but that without continued
sustained effort, the savings rate would level off in 2009, and fall short of the 2010 target.

The beauty of this index is that not only is it an accurate KPI for past performance, but also has
predictive capability, so that timely management action can be taken to make mid-course
corrections.

Finally, it should be noted that even though energy prices vary over time, including them in the
calculation procedure will not have a destabilizing effect on the EEI because prices occur in both
the numerator and denominator. To further reduce the impact of energy price fluctuations, the
recommended strategy is to re-calculate the EEI for all previous years using the new price
structure whenever there is a significant change, so that the numerical value of all indices will
reflect the latest and most representative economic conditions, and year-to-year comparisons can
be made on a common basis.

It suffers from only one drawback – calculating the “standard” energy consumption requires
accurate simulation models, which can expensive to develop and maintain. Unfortunately there is
no other way of accounting for a dynamic business environment. There’s a huge difference
between simple and simplistic. Simple is good; simplistic is bad.

CONCLUSION

The proposed comprehensive methodology has been proven to meet all the principal objectives
important to management, and the formulation principles are generally applicable. It is
recommended for adoption by all the process industries.

Keeping in mind that the objective of KPIs is to effectively manage energy performance, it is
critical to ensure that the following supporting infrastructure is provided:

1.) All energy consumption should be measured accurately (more/better instrumentation and
data reconciliation software may be needed).
2.) Accurate material and energy balances for all operations.
3.) All energy consumption should be associated with a specific value adding activity.
4.) Primary energy (eg. fuel, purchased power) should be priced at marginal cost; secondary
energy (eg. steam, cooling water, cogenerated power) costs should be traced back to
primary energy costs via the simulation model [3].
5.) A data acquisition and archival system is needed to document accumulated knowledge and
apply it systematically to operational and strategic decision making.
6.) For effective results, a company-wide online Energy Management Software system should
be deployed (Figure 13).

Industrial Energy Technology Conference, New Orleans (May 2009) Page 14 of 16


®
Figure 13. Architecture for OpsKPI Energy Management Software Deployment [8]
NOTE: OpsKPI is a registered trademark of AspenTech Inc, Boston, Mass.

ABBREVIATIONS

Description
BL Black Liquor (Kraft pulp mill)
BOE Barrels of oil, fuel equivalent
Btu British Thermal Unit
CEO Chief Executive Officer
EEI Energy Efficiency Index
EII Energy Intensity Index
EPI Energy Performance Index
HMB Heat & material balance
HP Horsepower (for motors)
HP High pressure (for steam)
K Thousand
KPI Key Performance Indicator
LP Low pressure (for steam)
MB Thousand barrels
MM Million
scf Standard cubic feet (of gas)

Industrial Energy Technology Conference, New Orleans (May 2009) Page 15 of 16


REFERENCES

1. K R Amarnath, J D Kumana and J V Shah, “Benchmarks for Industrial Energy Efficiency”,


presented at Intersociety Energy Conference, Washington, DC (Aug 1996).

2. D Farrell and J K Remes, “How the World Should Invest in Energy Efficiency”, McKinsey
Quarterly (July 2008).

3. J D Kumana and M M Al-Gwaiz, “Pricing Steam and Power from Cogeneration Systems
using a Rational Allocation Procedure”, Proc of 26th Industrial Energy Technology
Conference, Houston, Tx (April 2004).

4. J D Kumana and K D Al-Usail, “Energy Performance Indices as a Process Diagnostic


Tool”, presented at Process Performance Monitoring and Data Analysis Symposium,
Manama, Bahrain (Nov 7–8, 2006).

5. J D Kumana, A H Al-Qahtani, and F H Al-Farsi, “Power Savings via Load Management at


Rabigh Refinery”, presented at 2nd Saudi Arabian Energy Conservation Forum, Dammam,
Saudi Arabia (Nov 28-29, 2006).

6. J D Kumana and A S Aseeri, “Success Factors for a Corporate Energy Program”, presented
at 29th Industrial Energy Technology Conference, New Orleans, La (May 9-10, 2007).
Extended article republished in two parts - Insulation Outlook (Nov 2008 and Dec 2008).

7. Kumana & Associates, “Industrial Energy Performance Measurement and Monitoring” a 3-


day training course (in prep, 2009).

8. E Petela, AspenTech Ltd, Warrington, UK, personal email communication (2007).

9. S Rivoire, M A Shah, P Ranganathan and C Kozyrakis, “JouleSort: A Balanced Energy


Efficiency Benchmark”, presented at SIGMOD conference, Beijing, China (June 2007).

10. US Environmental Protection Agency, “Efficiency Metrics for CHP Systems”,


www.arb.ca.gov/cc/ccei/publications, (2007)

11. World Business Council for Sustainable Development, “Energy Efficiency as a Strategy:
GE Case Study”, www.wbcsd.org, (2008).

Industrial Energy Technology Conference, New Orleans (May 2009) Page 16 of 16

Вам также может понравиться