1

Management of a compressed air network

Tasso de Figueiredo, J.a,b, Carvalho, R.b, Reis, M.a

a Iberol, S. A., Quinta da Hortinha, Alhandra, 2601-908 - Vila Franca de Xira,, Portugal b Instituto Superior Técnico, Avenida Rovisco Pais, 1, 1049-001, Lisboa, Portugal

ABSTRACT The aim of the present work was the management and optimization of a compressed air network, using the existing case

study in Iberol, S.A., to reduce the excessive costs associated with the production of compressed air. To achieve the

proposed objectives a survey the existing network took place, identifying the critical points in the network and establishing

preventive maintenance methodologies for the compressed air network.

For the development of work we used software tools based on Excel® to support the implementation of the established

methodologies and to evaluate the obtained results in future improvements.

With the execution of the work subsequently demonstrated it was possible to know all the valences of the existing

compressed air network and which points to optimize. It was also possible to implement some optimization and evaluate

the importance of monitoring the existing compressed air network.

KEYWORDS: Compressed Air, optimization, Compressed Air Network, Utilities.

1. INTRODUCTION

Iberol, S.A., founded in 1967, is a Portuguese

company with the main purpose of producing biodiesel

and animal food. This company is one of the national

leaders in the market.

The main sections for the production of these

materials is the preparation and extraction, where is

made the oilseed processing and obtained flour and

crude oil, and the biodiesel production unit (UPB), where

crude oil is neutralized and it is obtained biodiesel.

Iberol, S.A. has an extensive compressed air

network that, at the starting date of the execution of this

thesis, wasn’t fully known. The routes of the compressed

air network and their end-use equipment, weren’t known.

[1]

Compressed air consumption in the different

production processes in Iberol, S.A. was not known.

Consequently, energy consumption for production of

compressed air was also unknown.

To correctly manage and optimize the existing

compressed air network, it’s necessary to identify all the

leaky and obsolete points.

The awareness of the importance of continuous

improvement and autonomous maintenance to the

compressed air users is one of the essential points that

can lead to reductions in compressed air consumption up

to 20%. [2]

The identification of the processes that

generate higher air consumption is also essential to

create a rapid and effective strategy to optimize the

existing compressed air network.

2. METHODS

For the management and optimization of the

compressed air network Iberol, S.A. created a

methodology, according to Figure 1.

Mapping and diagramming

The compressed air network mapping is the first

step for managing the compressed air network.

To properly map all existing network there was

a need to identify all equipment that uses compressed air

and the routes of the pipes that feed them. Also an

analysis was made of the diameters of the pipes and the

type of material that constitutes them. This analysis was

made to allow future optimization. It was also identified

the points where the compressed air supply is still

connected but no longer used.

Identification of critical points and leaks

Next, were analysed the critical points of the

compressed air network. For this analysis it has been

registered all the points that needed intervention and

those have been identified in the previously generated

diagrams.

For the identification of anomalies, so that

maintenance teams could start some repairs, some

concepts related to Kaizen methodology were

introduced. Then, it was placed at the locations in need

of intervention, noncompliance labels. These labels

2

made it easier to identify the repair location when the

repair person visits the site.

With these records it was possible to make an

assessment of which sections needed more intervention

based on the compressed air network density and

quantity of detected anomalies.

Using the graph in Figure 2, it is possible to

identify the sections in which the percentage of

anomalies is higher are the section of the Silos and the

section of Parque de Tanques.

By comparing this analysis to the density of the

existing compressed air network on these sections it can

be concluded that a critical point in terms of compressed

air network is the section of the Silos.

The analysis above shows us that this is a

critical point in terms of poor maintenance. However, it is

also important to define, although the compressed air

supply in the Silos section is dense, who is the greatest

consumer of compressed air in the plant. For this it is

necessary to identify the compressed air consumption of

each section.

Analysis of the main consumers

For the analysis of the main consumers there are

two hypotheses:

Practical tests of the compressed air

consumption;

Model for the estimated compressed air

Consumption

These two methods of identifying consumers are very

different and can complement each other. So, we tried to

carry out both methods.

Practical tests of the compressed air

consumption

To perform the practical tests on compressed air

consumption, we followed the method shown in Figure

3. These tests consist in isolating the different

consumers and obtain readings of the respective time

intervals, in the corresponding energetic counter.

Figure 3 - Scheme of energy consumption testing strategy.

To select the correct time intervals in order to

obtain consistent results, it is necessary to evaluate the

sensitivity of the evolution of the energy counters.

The sensitivity of the counters can be evaluated

by recording counters values of similar days and then

checking the variation that occurs. With this variation we

can calculate the time required to perform the tests.

For the evaluation of energy counters resorted

to the history of meter readings. The days selected for

evaluation were days when the production of the various

sections is similar to the productions planned for the days

of testing.

So 12 days were selected in a four-year history.

Then its time evolution was analysed in order to

determine the number of hours required. For the

evolution of the counter to be relevant it is necessary a

Figure 1 - Exemplification diagram of applied methodology.

Figure 2 - Distribution of anomalies identified by section.

3

change of the sixth digit. It was found that the optimal test

time would be about a ten hours period analysis, i.e., t1

= 10 hours.

The determination of the value t2 is a bit more

subjective. Thus, it is important to ensure that the

evolution of the counter is one digit, so there are no

errors due to the losses. Thus, the minimum value of t2

is three hours, that is, t2 = 3 hours.

The conditions for the test include the stop of

the preparation and extraction processes and also

stopping all charges and discharges of materials,

because cutting the compressed air will damage the

equipment. Therefore it is necessary that the plant

remain the same conditions for 39 consecutive hours,

two days. It is necessary to repeat the testing for

confirmation of the values obtained in the first part, so

the required time of the test is 78 hours, four days.

Model compressed air Consumption

Estimation

In this model, the goal is to assign compressed

air consumption according to the production and

consumption of the raw materials and products. For this,

it was possible to assemble data of raw materials

consumption, production and associated energy

consumption. It was possible to build a history since

September 2013.

Outliers are points that are diverted from normal

distribution of data points. To calculate these points we

used the method of the least squares. Thus, it was

calculated for each data set, the mean, the lower quartile

and the upper quartile. [3]

The evaluation points followed a few criteria,

which are thereafter reported.

It was considered the production of neutral oil

as a whole, rather than by specifying the type of oil, as

they are produced based on a certain set composition

according to the client as intended.

As for rapeseed and soybean, for not being

easy and accurate the measurement of the production

process, it was decided to rely on seed consumed in

each case.

After this analysis, it is important to also take

into account the days of extraordinary use of

compressed air due to works or unexpected work. The

way to analyse these points is doing an assessment to

outliers similar to that performed previously, but for the

electricity consumed in the compressed air production

room.

Figure 4 - Points used for model calculations.

The designed model, aims to calculate the

specific compressed air consumption for each process

and also the wasted energy estimation of air leaks. For

this, some assumptions were made:

The value corresponding to the waste

of energy in compressed air leakage is

constant, i.e. independent of the

working process because the pressure

in the pipes and size of holes are

constant;

The efficiency of the compressor is a

constant percentage.

To build the model, since the electricity counter

isn’t only measuring the compressors but also includes

the compressed air dryer, it is necessary to consider

some other parameters.

So the model is represented by the following

equation.

𝐸𝑐𝑜𝑛𝑠𝑢𝑚 = ∑(𝑚𝑖 × 𝑄𝑖) + 𝐸𝑙𝑒𝑎𝑘𝑠 + 88,8 + (∑(𝑚𝑖 × 𝑄

𝑖))

× (1 − 𝜂) [𝑘𝑊ℎ]

The first part of the equation refers to the power

consumption of compressed air at each section, where

𝑚𝑖 refers to the specific consumption.

The value of 88,8 refers to the energy

consumption of the dryer, as indicated by the

manufacturer.

The last set refers to the lost value due to the

inefficiency of used compressors.

So for the used points, it applied the method of

least squares and then used the Solver Excel®

application. The values of the estimated parameters are

shown in Table 1.

Table 1 - Obtained parameters for the model.

Parameter Value Units

𝑚𝑁𝑒𝑢𝑡𝑟𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛 0,169 𝑘𝑊ℎ 𝑇𝑜𝑛𝑃𝑟𝑜𝑑𝑢𝑧𝑖𝑑𝑎⁄

𝑚𝐵𝑖𝑜𝑑𝑖𝑒𝑠𝑒𝑙 0,067 𝑘𝑊ℎ 𝑇𝑜𝑛𝑃𝑟𝑜𝑑𝑢𝑧𝑖𝑑𝑎⁄

𝑚𝑠𝑡𝑒𝑎𝑚 0,205 𝑘𝑊ℎ 𝑇𝑜𝑛𝑃𝑟𝑜𝑑𝑢𝑧𝑖𝑑𝑎⁄

𝑚𝑆𝑖𝑙𝑜𝑠 0,020 𝑘𝑊ℎ 𝑇𝑜𝑛𝑀𝑜𝑣𝑖𝑚𝑒𝑛𝑡𝑎𝑑𝑎⁄

𝑚𝐴𝑟𝑚𝑎𝑧é𝑛𝑠 0,086 𝑘𝑊ℎ 𝑇𝑜𝑛𝑀𝑜𝑣𝑖𝑚𝑒𝑛𝑡𝑎𝑑𝑎⁄

𝑚𝑠𝑜𝑦𝑏𝑒𝑎𝑛 0,011 𝑘𝑊ℎ 𝑇𝑜𝑛𝐶𝑜𝑛𝑠𝑢𝑚𝑖𝑑𝑎⁄

𝑚𝐹𝑢𝑙𝑙−𝑓𝑎𝑡 0,657 𝑘𝑊ℎ 𝑇𝑜𝑛𝐶𝑜𝑛𝑠𝑢𝑚𝑖𝑑𝑎⁄

𝑚𝑅𝑎𝑝𝑒𝑠𝑠𝑒𝑒𝑑 0,007 𝑘𝑊ℎ 𝑇𝑜𝑛𝐶𝑜𝑛𝑠𝑢𝑚𝑖𝑑𝑎⁄

𝐸𝐿𝑒𝑎𝑘𝑠 400,0 𝑘𝑊ℎ

𝜂 22,4 %

Making a brief analysis of the obtained

parameters, you can make some conditions for the

verification of the model:

The efficiency obtained through the

model is low, taking into account the

values commonly obtained for this

02

/20

14

05

/20

14

08

/20

14

12

/20

14

03

/20

15

06

/20

15

10

/20

15

0

200

400

600

800

100 0

120 0

kWh

4

type of compressor, between 65 and

75%;

for the specific energy consumption for

the production of full-fat is high,

because the quantities produced are

low;

The specific for the flour and soy oil

production process should be a little

higher than the process of producing

flour and rapeseed oil, as it

contemplates a stage that does not

exist on the rapeseed process, and as

such it will increase the compressed

air used in automation;

The value obtained for energy

consumption wasted on air leaks

along the network corresponds to an

average of about 48% of compressed

air consumption. In compressed air

systems this value can vary between

15 to 50% depending on the type and

frequency of maintenance that is

applied.

To verify that the model fits the reality of the

plant, the data for the year 2016 were used.

Then it was obtained an average error of 14%

and a production distribution represented in graphic of

Figure 5.

Although, with this model, it is not be possible

to extract the actual consumption of compressed air, it

can be concluded some events:

It is noted that, due to non-existent

maintenance and concern in recent

years with the compressed air network

by Iberol, S.A., the energy leakage

loss is a major part of the overall

consumption;

It is noted that the inefficiency of the

compressor is a point to improve

because even applying the deviation

of the model obtained, the value of

inefficiency is very low and there are

methods to improve this value;

The specifics that have a greater

influence in the model are related to

the production of steam and Full-fat

and product movement in

warehouses.

Implementation of autonomous maintenance

methodologies

In order to optimize the compressed air network

in Iberol, S.A. it was decided first to execute

optimizations that do not involve major investment, and

secondly optimize, if possible, performing actions where

significant investment is needed. [4]

So the actions that do not imply increased

investment are:

Reduce the number of leaks;

Disposal of obsolete Points;

Adjust the quality of the compressed

air.

Reduced the leakage number

To eliminate air leaks it was decided to integrate

this task in the Kaizen methodologies. The methodology

is called the autonomous maintenance and has as main

objective to motivate and centralize the equipment

maintenance in the people who work directly with it. In

short, the operators do the survey, detection, simple

repair and communication with maintenance operators if

necessary.

For this, and with the survey done earlier, it was

analysed the necessary training for the operators to be

able to perform some simple tasks.

It was also necessary to assess whether it was

required to perform some maintenance to restore the

equipment initial conditions.

47,33%

32,56%

4,88%

2,02%

8,38%

4,25%

6,05%

1,13%

5,25%

0,60%

32,56%

Perdas Processos Neutralização Transterificação Central de Vapor Silos Armazéns Soja Full-Fat Colza

Figure 5 - Distribution compressed air consumption by various processes.

5

As this would imply a high investment in parts

and working hours of maintenance operators, we chose

to perform only the tasks that would be fundamental. The

remaining repairs were carried out by operators in their

respective sections. So it may be faster to implement and

solve problems briefly.

To assist the organization and schedule of the

routes for the future, and to be possible to organize the

air routes and integrate with the Autonomous

Maintenance procedural equipment of various sections it

was developed a tool in Excel®.

At the beginning of each month, is generated in

each section a timetable for completion of routes where

operators consult the days when you need to perform the

routes.

This tool allows the autonomous maintenance

management of all sections. In this tool you can add and

remove Autonomous Maintenance routes, change the

frequency and day of the week they occur. These

changes are updated in the different sections at the

beginning of each month, before being generated the

timetable, Figure 6.

At the end of each month, to evaluate the

performance of the autonomous maintenance is made

an update of the autonomous maintenance management

tool where updated graphics and performances are

compared to each section and also the type of

anomalies, among other analyses of interest with the

tool.

Throughout the month, to ensure that there are

no issues to point the operators and to ensure that the

autonomous maintenance is made with the highest

quality, it is possible to carry out audits of the various

sections. For this, the management tool of the

autonomous maintenance has the possibility of the

auditor to select the days and sections you want to audit.

After this selection, the tool returns a token that must

accompany the auditor to where valuations are

registered, the execution quality of the route and some

observations that may be necessary. At the end of the

audit, the auditor, from the tool in Excel, you can conduct

an audit report that will be sent to the corresponding

section heads, so that they become aware of the

improvements to be made and the observations of the

auditor. This auditor in the case of Iberol, SA, is

responsible for the mechanical section which specializes

in all maintenance tasks and is able to assist operators

with all the doubts and questions that may arise during

the implementation of the autonomous maintenance

routes.

At the same time, and to improve the

organization of the autonomous maintenance in specific

sections, each section was created in one autonomous

maintenance framework. This framework summarizes all

routes for the current month and some useful information

that is required. This table is an advantage, because

when working in shifts, sometimes operators have

difficulty knowing if the route has already been made or

if there was no opportunity in the previous round and thus

being attached to your query is easier.

Disposal of obsolete Points

The next step in optimizing the existing

compressed air supply is the elimination of redundant

dots, i.e., all points in the former network that have been

used, but are now no longer necessary and, therefore,

are points with leakage potential that can be eliminated.

To do this it was analysed all the compressed

air network and identified numerous points.

After identifying all the points, were made the

improvements that were possible without the intervention

of the maintenance team or involving degassing

equipment.

Figure 6 - Autonomous Maintenance management tool.

6

Quality setting compressed air

The recommended compressed air quality in

Iberol S.A., analysing the benchmark values, are the type

(2,3,2). [5]

This class is attributed to the existence of

instrumentation along the compressed air network and

these devices require a more limited quality of

compressed air.

Thus, analysing the existing compressed air

network, it’s found that filters and dryers that are in the

compressed air generation room does not reach the

desired values.

The air drying is not sufficient and therefore is

compromising the operation and longevity of the existing

instrumentation equipment.

To evaluate the result of insufficient drying air,

measurements were made during a time interval, to

check the quality of compressed air at the dryer outlet.

In Figure 7, the data is collected for evaluating

the compressed air humidity for the end-use equipment.

The lower horizontal line represents the target value for

the dew point to meet the requirements of the quality of

compressed air. Thus, it is concluded that the

compressed air leaving the air output area has an excess

of moisture.

One way to minimize the amount of water

present in compressed air is the removal of any water

that is condensed along the tubing. To achieve this we

need to make Purges. Existing Purges in Iberol, S.A. are

represented in the diagrams previously executed. These

drains are of the manual type, and as such, the

maintenance of these purges are included in the self-

maintaining routes by operators.

However, existing Purges are not sufficient.

According to the literature examined, purges should be

spaced apart by 20 to 30 meters. As can be seen from

the diagrams, throughout the UPB compressed air

network, which supplies compressed air in

instrumentation there are no purges. This is a sensitive

area of the air humidity and the placement of drains in

the various floors of the UPB is critical. [6]

3. CONCLUSIONS AND FUTURE PROPOSALS

In order to manage and optimize the existing

compressed air network in Iberol, S.A several different

tasks took place.

Initially, with drawing diagrams of compressed

air network it was possible to determine which customers

require compressed air and which routes to run through

the pipes and their derivations.

In parallel with this drawing, points have been

identified on the network that were leaking and in need

of intervention. These points were reduced by about

60%.

The reduction of leakage points was achieved

due to the collaboration of the operators of the various

sections; with the completion of the defined Autonomous

Maintenance routes it was repaired much of the existing

trails. Thus, existing leaks in the plant are currently in its

bulk, a small leak, not audible to the human ear, or leaks

that require a shutdown process.

In the future, so that the leakage is repaired at

the time of appearance and in order to prevent that

leakage of the orifice increases its size, it is important to

obtain leak detection equipment. This equipment would

be useful now, because operators already have the

routine of checking routes and would be an asset to

increase the performance of the Autonomous

Maintenance.

Another initiative taken in this work was to

perform tests to estimate the consumption of

compressed air. These tests did not lead to relevant

conclusions because the testing time available didn’t

allow a sufficient movement of the counters. For this

reason, there was a search for the compressed air

consumption history over the past three years in the plant

in order to estimate a mathematical model, the

compressed air consumption in each process. This study

showed that consumption wasted on leakage is high and

the efficiency of the compressors is also significantly low.

To improve the efficiency of the compressors is

important to study the decrease in temperature of the

compressor room. As a benchmark, it is estimated that

for every 3C reduced to the inlet temperature it will be

reduced up to 1% in energy consumption.

-25,00

-20,00

-15,00

-10,00

-5,00

0,0 0

5,0 0

De

w p

oim

t (

C)

Figure 7 – Dew point compressed air to the dryer outlet.

7

To determine the most significant consumers

and distribution of energy costs for the air consumption,

the execution of a consumption estimate testing when

there is a scheduled plant stop will be crucial.

Finally, in order to reuse the energy dissipated

in the compression process, it is important to evaluate

the opportunity of reusing the waste heat in the

compression process. For this, Atlas Copco®,

manufacturer of the existing plant compressors has a

heat recovery system which allows reuse of about 70%

of the heat. This measure could be profitable in the Iberol

system, S.A., because of the proximity, the reused heat

can be used to heat the water fed to the reverse osmosis

process. This reuse may allow a reduction in the

consumption of the existing heat exchanger for heating.

4. BIBLIOGRAPHY

[1] Iberol, S.A., "Apresentação Institucional

IBEROL," Maio 2011. [Online]. Available:

http://www.iberol.com.pt/download_pdf/1.pdf.

[Accessed 05 Março 2016].

[2] Atlas Copco, Atlas Copco Compressed Air

Manual, 8º ed., Belgica: Atlas Copco Airpower

NV, 2015.

[3] O. Helene, Método dos Mínimos Quadrados, 2º

ed., Livraria da Fisica, 2013.

[4] Kaizen Institute, Manutenção Autónoma 1,

2016.

[5] ISO, ISO 8573-1:2010, 2010.

[6] Parker Hannifin Corporation, Sistema de

tubulação para distribuição de ar comprimido e

vácuo, 2000.