Reduce Electric Power Cost Project
#16,005
Final Report
PAG Import, Inc.
Reduce Electricity Consumption Project Team
January 7, 2004
Team Members
Project Champion
Mr. M. Fukui (AS/ Quality Development)
Project Members
Ms. K. Uemura (Process Owner – HR/ Administration)
Ms. Akiyama (AS/Quality Development)
Mr. I. Takayama (FIN/ Information Technology)
Mr. S. Ohno (FIN/Finance, Business Plan)
Mr. N. Adachi (BB – AS/ Quality Development)
HR: HR & Administration division／AS: Aftersales Business Division／FIN: Finance
Division
Problem Statement
Much electric power is consumed at the PAGI/FJL
office. The electric power cost is significantly
influencing the profitability of the company. The electric
power cost of the Kamiyacho office (Headquarter) was
$218,700 in 2002. This corresponds to $50 per
employee per month.
Purpose of the Project
Reduction of Electricity Cost
Project Scope
Activities of the project will be limited to the whole
Kamiyacho office (7/8/10/11F).
Air-conditioning cost of the business hours (8:00-18:00/
weekdays) is included in the office rental cost (fixed). As we
cannot reduce or control this cost, we do not include this in the
scope of the project. The scope should be normal electric
power cost and the air-conditioning cost of overtime hours.
Basic Strategy of the Project
The activities of the project should be coordinated
with the activities for Volvo ISO14001. The project
should be linked with the activities to protect
environment and should not be perceived as mere
cost reduction.
The negative influence on the business
environment of employees must be minimized.
Measurement
Electricity Cost (Yen) / Electric Power Consumption
(kwh) and Air-conditioning Cost for Extra Hours
Monthly consumption of electric power is notified through the
monthly bill from the Shuwa Building office.
Electric power consumption and its cost are in a linear
relationship. Air-conditioning cost for extra hours is
determined based on the length of operation (hours) and
areas (# of blocks).
Target
Reduce the electricity cost of the
Kamiyacho office by 20% from $218,700 to
$162,100. Reduce A/C cost for extra hours
by 20% also.
VOC - How to Promote Employees’ Efforts in
Saving Electricity Consumption – Points to be
considered when taking measures
Always remind people of saving/ Avoid being forgotten.
Failure of saving could cause some disadvantage on each employee.
Make everything simple/ Avoid complicated processes.
Saving could bring convenience/ improved efficiency (No inconvenience/
no problems in efficiency).
Saving activity looks cool / trendy.
Give the impression that everyone participates.
Efforts of each person is essential for the total success.
Saving is along with the corporate policy.
Each person can contribute a lot/ Problems could occur if I do not save.
Saving effects can be confirmed clearly / We can feel the effects.
Business achievements are evaluated by the saving effect.
Not a temporary project. Activities should be continued.
The Process of Using (Wasting) Electricity
C&E Diagram
C&E Matrix
Lowering wattage (without influencing on the business
environment) is important in this project. Each process input
(areas of electricity consumption and appliances, such as lighting
of office and PCs) should be evaluated according to this.
Points mentioned in the VOC can be basically fulfilled equally by all
the process inputs.
A rough estimate of power consumption shows that lighting of
office and PCs are two major process inputs in lowering wattage.
As the most important process inputs are clear, we did not develop
a C&E matrix for this project.
MSA - Converting the Monthly Data (1)
Monthly power consumption data is reported by the
maintenance company of the building after reading the
wattmeters. It has turned out that there are some problems
for this data as mentioned below. We must adjust the data
to compare the consumption of power of each month
correctly. Tenants cannot check the meters by themselves
at this building.
The days of checking meters differ from month to month. The data
of the dates of checking can be followed to the past.
The numbers of business days differ much from month to month.
The majority of appliances are not used on weekends though some
are used just like main frame computers. Therefore, consumption of
power significantly differs between business days and holidays.
Converting the Monthly Data (2)
We applied the multiple regression analysis on monthly
power consumption (Y), the number of business days (X1)
and the number of holidays (X2) in the past data. We
estimated the relationship between power consumption and
Xs – the number of business days and holidays.
We converted the consumption data by using this
relationship to make correct comparison between months
with different number of business days and holidays.
It has turned out that the power consumption was fairly
stable between October 2002 and the start of the project by
the analysis of the converted past data.
Regression Analysis： KWH versus Work days, Holidays
The regression equation is
KWH = 14486 + 2929 Work days + 1627 Holidays
Predictor Coef SE Coef T P
Constant 14486 12417
Work day
Holidays
S = 3649 R-Sq = % R-Sq(adj) = %
Analysis of Variance
Source DF SS MS F P
Regression 2 503844693 251922347
Residual Error 17 226339999 13314118
Total 19 730184693
Source DF Seq SS
Work day 1 236055862
Holidays 1 267788831
Unusual Observations
Obs Work day KWH Fit SE Fit Residual St Resid
5 95248 88682 1692 6566
7 102631 95191 1134 7440
Power Consumption
(Before Starting the Project)
The power consumption of the four floors (Number of days
adjusted) combined is shown below.
Comparison with
the previous
year(%)－Left
column
Consumption(kwh)－
Right column
Basic Concept of Z Value/ DPMO
<PCs and Spaces With Independent Light Switches>
Defects for PC usages are defined as that PCs are ON
after their users left office.
Some PCs, which are kept ON to have data transmission
with abroad are excluded from defects.
Defects for spaces with independent light switches are
defined as that lights of spaces are ON while no one is
using.
Z Value/DPMO (Before Taking Measures)
Z values and DPMOs were measured for spaces with
independent light switches and PCs.
The investigation were made for the spaces with independent switches on
August 22 and 26 (average). That for PCs was made on September 19.
Conclusion From the Results of the
Investigation
The Reasons of High DPMO
Generally speaking, employees do not think they should
turn off switches of PC monitors. Many employees do not
know that PC monitors have switches.
Some are not conscious that lights of spaces with
independent switches such as meeting rooms should be
turned off. Also, light switches are generally so small that
many do not notice their existence.
FMEA
Measures Taken (1)
Checked ON-OFF of PCs before business hours and gave warning to
users of PCs kept ON.
Checked ON-OFF of lights of spaces with independent light switches.
Announced the results and asked employees to improve.
Developed and installed labels that highlight switches of spaces with
independent light switches.
Attached maps of the areas covered by each switch to the switch board.
Conducted a “Save Energy Idea Contest” and implemented some
excellent ideas applied for it (47 ideas applied from 19 employees).
Informed all employees light switches that cover their offices and persons that
are with the same switches (to ease turning off lights).
Lowered the temperature of the hot-water supply (95->40C/ 203->104F).
Stopped supply during weekends.
Installed labels that show the power consumption of areas controlled by each
switch (In money/ weight of CO2/ areas of forests required for absorbing CO2).
Installed caution labels to PCs, which call attention for turning off when users
leave office.
Measures Taken (2)
Developed action guidelines for saving power cost. Delivered them to
employees by an E-mail and uploaded them in the PAGI web-site.
Asked and enlightened employees to save power through various kinds of
E-mails.
The rule of extending the air-conditioning hours that the requests must go
through the Administration department is strongly enforced.
Shortened the length of air-conditioning hours during summer compared
with the previous years (until 20:00->19:00).
Developed and posted posters that call for saving electric power.
Developed and announced the rule of turning on/ off the lights of the
reception of 8th floor (consumes high wattage).
Installed electric bulbs with lower wattage for the reception of 8th floor
(85W->65W).
Removed some of fluorescent lights near windows.
Measures Taken (3)
Outsourced checking and turning off switches of lights,
printers and copy machines on every Saturday morning
(Started on November 15).
Replacing PCs with energy-saving models (300 units).
Adjusted existing PCs to energy saving mode (63 units).
Highlighted door knob and light switches of the smoking
room with noctilucent tapes.
The smoking room does not have windows. The room becomes dark if the lights are
turned off. As the door and the light switch are far apart, it was difficult to find the door
after turning off the light switch.
Cost of The Measures Taken
Outsourcing cost to check and turn off light
switches on Saturday costs $50 every month.
However, the effect of the first trial is estimated
more than $90.
The cost of taking the measures of the project is
only around $20 including those for labels and
noctilucent tapes.
DPMO
The Z value and DPMO for leaving PCs and lights
ON have been greatly improved.
The investigation was made
on November 20.
FMEA – After the Measures Were Taken
Minitab Analysis – Statistic Meaning of the
Improvement (1)
The ratio of PCs left ON dramatically decreased after the
measures were taken. The improvement is statistically
significant.
Test and CI for Two Proportions
Sample X(OK-OFF) N(Total) Sample p
1(Before) 139 367
2(After) 298 366
Estimate for p(1) - p(2):
95% CI for p(1) - p(2): (, )
Test for p(1) - p(2) = 0 (vs not = 0): Z =
P-Value =
Minitab Analysis – Statistic Meaning of the
Improvement (2)
The mix of the spaces with independent light switches left
ON also has been dramatically diminished. The
improvement is statistically significant.
Test and CI for Two Proportions
Sample X OK-OFF) N(Total) Sample p
1(Before) 75 104
2(After) 58 68
Estimate for p(1) - p(2):
95% CI for p(1) - p(2): (, )
Test for p(1) - p(2) = 0 (vs not = 0): Z =
P-Value =
Power Consumption (Adjusted) – All Floors
Combined
The start of
the project
YOY (%)－Left
column
Consumption (KWH)
– right columns
Cost of Extending Air-conditioning Hours
The cost of extending air-conditioning hours has been
dramatically lowered.
YOY (%) – Right
column
Consumption (kwh)－
Left column
Estimate of the Profitability Improvement
Power consumption has been lowered by % from the ’02
data (’03 August – October average).
Extension cost of air-conditioning hours has been reduced by
% (’03 August – October average).
Applying these reduction ratio to the full year cost of ’02, we
would get the following cost improvement.
Power cost： $16,970（Excluding fixed charge/ ’02 total:
$180,600）
Extension cost of air-conditioning：$67,890（’02 total: $133,400）
Total：$84,860 （% of the ’02 total: $314,000）
<$1= Yen/ #682: ’02 – ’05 Average>
+)
Control - I-MR Chart （’02 Jan.～’03 Oct.）
The process will be controlled with I-MR charts for power
consumption (# of days adjusted) and the YOY % of air-
conditioning cost for extra hours.
Power Consumption I-MR Chart YOY % of Extra Air-conditioning Cost I-MR Chart
Control Plan
Control – Process Control in the Future
The follow-up meeting will be held six months after the
project is closed (around ’04 July). Whether the
improvement is continued without problems and
whether there is any potential problems in the process
will be discussed.
Whether the process is in control or not will be judged
every month. If there are any problems, causes will be
immediately investigated and remedies will be taken.
Long Term MSA
Adjusting the number of days in each month will be
done using the conversion method developed in the
Measure phase.
However, the change in the relationship in the
weight of weekdays and weekends could change
because of the measures taken in this project (.
promoting turning off lights on weekdays could raise
the weight of weekends). Checking based on the
new data after the measures were taken will also be
done.
Conclusions
The project has significantly reduced electric power
consumption; however, the reduction is as far as around 10%.
However, as the reduction in the extension cost of air-
conditioning hours is so huge that the combined reduction of
power consumption and extension cost of air-conditioning
hours well surpassed 20%.
Many of the measures taken will fully contribute after
November (data not known yet). More saving effects by the
project is foreseen. This would be confirmed in the follow-up
meeting.
The project has brought excellent outcomes.
Replicating the Project
All the methods taken in this project can be taken also in
other facilities in Japan and other markets (countries).
