Savings Calculator for Centrifugal Pumps

Bachelor’s Thesis 6.10.2014 LUT Energy

Degree Programme in Electrical Engineering

SAVINGS CALCULATOR FOR CENTRIFUGAL PUMPS

Lauri Nygren

ABSTRACT

Lappeenranta University of Technology LUT School of Technology

LUT Institute of Technology, Electrical Engineering Lauri Nygren

Savings calculator for centrifugal pumps 2014

Bachelor’s Thesis.

47 p.

Examiner: D.Sc. Tero Ahonen

Centrifugal pumps are one of the major energy consuming end-devices in developed coun- tries both in industrial and services sectors. According to recent studies, even 30 % of the energy used in pumping systems could be saved by more careful choosing of devices and system design. One of the most efficient and affordable ways to decrease the energy con- sumption of the pumping system is to substitute traditionally used flow control methods, like valve control, with modern variable speed drive (VSD) control.

In this thesis, Microsoft Excel based program, Savings Calculator for Centrifugal Pumps (SCCP), is designed. SCCP calculates the achievable energy and financial savings when throttle control is substituted by VSD control in the pumping system. Compared to the sim- ilar existing programs, the goal is to make SCCP calculations more accurate and require less input information. Also some useful additional features are added to the designed program to make it more user friendly.

The reliability of the calculations of designed program seem to vary depending on case. The results are corresponding accurately to the laboratory measurements, but there occurs high deviations in some cases, when the results are compared to the pump information specified by manufacturer. On the basis of verification in this thesis, SCCP seems to be at least as accurate as similar existing programs and it can be used as help in investment decision whether to have VSD or not.

TIIVISTELMÄ

Lappeenrannan teknillinen yliopisto Teknillinen tiedekunta

LUT Energia, sähkötekniikka Lauri Nygren

Säästölaskuri keskipakopumpuille 2014

Kandidaatintyö.

47 s.

Tarkastaja: TkT Tero Ahonen

Keskipakopumput vastaavat suuresta osasta energian kulutusta kehittyneissä maissa, niin teollisuudessa kuin palvelusektorilla. Viimeaikaisten tutkimusten mukaan, jopa 30 % pump- pausjärjestelmissä käytetystä energiasta voitaisiin säästää paremmilla laitevalinnoilla ja jär- jestelmäsuunnittelulla. Yksi tehokkaimmista ja edullisimmista tavoista vähentää pumppaus- järjestelmän energiankulutusta on korvata perinteisesti käytetyt virtauksen säätömenetelmät, kuten venttiilisäätö, modernilla taajuusmuuttajasäädöllä.

Tässä työssä suunnitellaan Microsoft Excel pohjainen säästölaskuri keskipakopumpuille (SCCP), joka laskee saavutettavat energia- ja taloussäästöt, kun venttiilisäätö korvataan taa- juusmuuttajasäädöllä pumppausjärjestelmässä. Verrattuna samankaltaisiin olemassa oleviin ohjelmiin, tavoitteena on tehdä SCCP:n laskenta tarkemmaksi ja vaatia samalla vähemmän lähtötietoja järjestelmästä. Ohjelmaan lisätään myös muutamia hyödyllisiä lisäominaisuuk- sia käyttäjäystävällisyyden parantamiseksi.

Ohjelman tekemien laskelmien luotettavuus vaihtelee tapauksittain. Laskelmat vastaavat tar- kasti laboratoriomittausten tuloksia, mutta valmistajan julkaisemiin tietoihin verrattuna il- menee niissä joissakin tapauksissa suuriakin poikkeamia. Tässä työssä tehdyn vertailun pe- rusteella SCCP näyttää kuitenkin antavan vähintäänkin yhtä tarkkoja tuloksia, kuin saman- kaltaiset jo olemassa olevat työkalut ja sitä voidaan käyttää apuna taajuusmuuttajan hankin- nan päätöksenteossa.

CONTENTS

Abbreviations and symbols ... 5

1. Introduction ... 7

1.1 Objectives of the study ... 7

2. User interface of the program ... 11

3. Operation of the program and calculations theory ... 18

3.1 Pump ... 18

3.1.1 Throttle controlled system ... 22

3.1.2 VSD controlled system ... 24

3.2 Motor and drive ... 26

3.3 Energy consumption and economics ... 30

4. Evaluation of the results calculated by program ... 32

4.1 Comparison with manufacturer’s performance curves ... 32

4.1.1 Head curves ... 33

4.1.2 Efficiency curves ... 35

4.1.3 Power curves ... 37

4.2 Laboratory measurements... 40

4.2.1 Throttle control ... 41

4.2.2 VSD control ... 42

5. Conclusion ... 45

References ... 46

APPENDICES I Laboratory equipment

II Tables of laboratory measured values III Measurements of the pump nominal values IV Example of pump information and curves

from Sulzer Select

ABBREVIATIONS AND SYMBOLS

IRR Internal Rate of Return NPV Net Present Value

POR Preferred operation region

SCCP Savings Calculator for Centrifugal Pumps VBA Visual Basic for Application

VSD Variable Speed Drive

C Cost

cos(φ) Power factor

e Emission

g Acceleration due to gravity

H Head

i Interest rate

k Coefficient for friction head

l Lifetime

m Number of years

n Rotational speed

nq Specific speed

p Price

P Power

Q Flow rate

r Real interest rate

S Saving

t

T Torque

W Energy

η Efficiency

π Inflation

ρ Density

Subscripts

0 Initial

e Electric

E Energy

fr Friction

i Index

n Nominal

st Static

sys System

tot Total

1. INTRODUCTION

Centrifugal pumps are widely used to transfer fluids in numerous industrial and municipal applications. In European Union, pumping systems account the share of 22 % motor elec- tricity consumption in the industrial sector and 16 % in the services sector (Almeida 2003).

According to Kaya (2007), 30% of energy used by pumping systems in developed countries could be saved by better system design and careful pump choosing.

Investments in energy efficient devices in pumping systems are profitable in most cases, because energy savings usually mean financial savings. Life-cycle costs of a pumping sys- tem can be roughly divided into investment, operating, maintenance and energy costs. Life- time of a pumping system is usually considered to be from 15 to 20 years, sometimes even longer. Even though purchase and installation costs are predominating at the beginning, the major share of life-cycle costs comes from energy consumption (Ferman 2008). Energy costs can be minimized by using properly sized and efficient pump and motor, and by substituting traditional control methods like valve or bypass control with variable speed control. By using a variable speed drive (VSD) to control the system, the cost of maintenance and lost produc- tion may also be reduced due to better reliability (Europump 2004).

1.1 Objectives of the study

The main goal of this thesis was to create the Microsoft Excel based program, Savings Cal- culator for Centrifugal Pumps (SCCP), which calculates the achievable savings when sub- stituting throttle control with VSD control in the pumping system. Excel was chosen to be the platform for the program because it’s very widely used, graphic user interfaces are easy to create, graph plotting tools are good and programs can be made by combining the ordinary Excel worksheet tools and functions with more traditional code written with Visual Basic for Application (VBA) programming language. VBA is widely used to automate and extend Excel programs, and it can also be used with other Microsoft Office programs (Taanila 2013).

Some programs similar to SCCP are already made, for example Pumping System Optimizing Tool (PSOT) at Lappeenranta University of Technology, PumpSave 5.2 by ABB Automa- tion Group Ltd (ABB 2014), and Vacon Save by Vacon (Vacon 2014).

Pumping System Optimization Tool runs on MATLAB^® and it’s used to find optimized pump-motor-VSD combination for certain process and calculate its energy consumption and life-cycle costs (Mustonen 2013). Compared to PSOT, SCCP concentrates only in savings

that are achievable by the change of the control method in new or existing system. Hence it’s especially useful for the users who don’t want to invest in more efficient pump and mo- tor, but save money and energy by better system control. The program also requires less input information than PSOT. The graphical user interface of PSOT is shown in Fig. 1.1.

Fig. 1.1 Pumping System Optimization Tool (PSOT).

PumpSave 5.2 published by ABB is more similar to the program designed in this thesis. It also runs on Excel and calculates the achievable savings when chancing the control method of the pumping system. The results given by PumpSave are often quite realistic, but the calculations could be adjusted to be more accurate (Taskinen 2008). One of the objectives of this thesis was to find out if the calculations could be made with same or even better accuracy, by using different models for pump, motor, and VSD, and requiring less input information. The other benefits of SCCP are more detailed financial calculations, option to read the process from a text file and save PDF-report with detailed information and graphs.

User interface of ABB PumpSave 5.3 is shown in Fig. 1.2.

Fig. 1.2 ABB PumpSave 5.3.

Vacon Save is also similar tool to SCCP and ABB PumpSave, but it is a stand-alone program and requires installation. Vacon Save requires even more input information than two other mentioned programs, but its financial calculations are comprehensive. User interface of Va- con Save is shown in Fig. 1.3.

Fig. 1.3 Vacon Save.

This thesis is divided into three parts. The first part introduces the user interface and action of designed program. The second part concentrates on introducing equations and mathemat- ical methods behind the calculations. In the third part, the accuracy of the program is verified by comparing its results and models to the laboratory measurements, curves published by manufacturer and results given by ABB PumpSave tool.

2. USER INTERFACE OF THE PROGRAM

The designed program runs on Microsoft Excel and it is used to calculate energy- and finan- cial savings available in a pumping system, when using variable speed control instead of throttle control. The program operation is illustrated generally in Fig. 2.1, and it’s described more detailed in next chapter.

Input information Evaluation of the pump characterictics

Printing the results and plotting the graphs

Calculation of required shaft power

Evaluation of the motor and drive efficiencies Calculation of the required

electric power Calculation of the energy

consumption

Calculation of economic information

Fig. 2.1 Operation of the program in general.

The program makes the calculations on the basis of user given information about the pump- ing system, economic conditions and flow profile. The required input information is shown in Fig. 2.2.

Fig. 2.2 The input page of the program. Pump nameplate and system information, motor and drive efficien- cies and economic conditions are required for the calculations. The used units of each quantity are marked on the right side of the input field. Flow profile can be read from the file or inputted man- ually by clicking buttons on the bottom of the page. The used decimal separator for input values depends on Excel’s language settings.

All required information about pump, motor and drive should be found in the nameplates or datasheets of devices. The density of the liquid used in a process is usually known. Static head is determined by measuring or assessing the vertical difference of supply and destina- tion reservoirs’ surface levels. Information about current economic conditions is also re- quired to calculate the available financial savings. Energy price and related CO2 emission

can be inquired from the electric company. The investment cost includes purchase and in- stallation prices of the variable speed drive. Interest rate and inflation can be determined according to their current values. The lifetime of a pumping system is usually considered to be from 15 to 20 years, depending on system and process.

User can input the flow profile manually or with a text file. In the text file, the information must be divided in two columns and separated by commas. In the first column, there must be timestamps, for example in the form “15.1.2014 00:14:28” or “2014-1-15 00:14:28” or

“1/15/2014 12:14:28 PM”. The correct timestamp notation depends on the used language version of Excel because of the built-in date- and time handling functions used in the de- signed program. Intervals between timestamps do not have to remain constant. In the second column, there must be flow rate values in m³/h and used decimal separator is a dot. An ex- ample of a file with properly arranged data is shown in Fig. 2.3. Reading a flow profile from text file is useful, if user has measured flow rate data from a process. If the text file has header lines, the number of them must be entered on the input page.

1.1.2014 00:00:00, 330.50 1.1.2014 11:00:00, 440.33 1.1.2014 12:34:30, 420.48 2.1.2014 01:20:00, 500.00 2.1.2014 15:10:26, 450.11

Timestamp Flow rate in m³/h

Fig. 2.3 The properly arranged process data file. Timestamp and flow rate are separated by comma and dot is used as the decimal separator in flow rate values.

In manual input of the flow profile, information is also entered in two columns. In the first column, utilization time of each flow rate is given in hours and in the second column, flow rate is given in m³/h. Used decimal separator in both values depends on used Excel language version. The manual input is useful especially if user don’t have any measurement data from the process, or the used flow profile in the process is planned to be changed. The manual input page is shown in Fig. 2.4.

Fig. 2.4 The manual input page of the program. The utilization time is given in hours in the first column and the flow rate value is given in m³/h in the second column. After the flow profile is inputted on the field, “Calculate”-button can be clicked to do the calculations. “Clear”-button is used to clear the contents of the table for a new flow profile. Flow profile is automatically filled in the shape of a bell curve from 30 % to 100 % of the pump nominal flow rate. The used decimal separator for input values depends on Excel’s language settings.

The entered flow profile is assumed to be repeated for an entire year, so if there are times when the system is standing still, they should be entered too for more accurate calculations of the annual energy consumption. When all required input data is entered, the program makes the calculations and activates the results page. In the results, information about achievable savings and graphs of pumping system performance and savings potential are shown to the user. The results of the calculations are shown in Fig. 2.5.

Fig. 2.5 The results page of the program. Viability of the variable speed drive investment can be evaluated on the basis of the calculations. The program shows the energy and financial savings achievable by change of the pump control method.

User can draw conclusions about the profitability of the investment in VSD, on the basis of calculated information, which includes energy and financial savings and also a few important economic quantities, which are described in the next chapter.

The program also draws graphs, which describe pump performance and energy savings available with the change of flow control method. These graphs are illustrated in Fig. 2.6.

The pump QH curve illustrates the difference of head losses in the valve and VSD controlled systems. The QP curve compares power consumption of both control methods as a function of flow rate. Histograms are plotted on the basis of given flow profile and QP curves. The upper histogram shows the flow distribution of the process. The lower one shows the achiev- able relative energy savings with the used flow profile.

Fig. 2.6 Graphs of the results page. Pumping system characteristics are illustrated in two graphs on the top, the upper one is QH curve and the lower one is QP curve. Two lower histograms describe the flow rate distribution and savings potential at different flow rates with the given flow profile.

For more detailed information and graphs, user can move to the information page by clicking

“Detailed information”-button on the bottom of the results page. The information page con- tains input information and results of the calculations, QH and QP curves, histograms about flow rate and energy consumption distribution, specific energy and rotational speed curves, total efficiency curves for both control methods, and efficiency curves for all devices sepa- rately in both control methods. It is possible to save the information page by clicking “Save as PDF”-button. Information page also includes the text box for the pump name on the top of the page, and the text box for the user notes on the bottom of the page. The information page is shown in Fig. 2.7.

Fig. 2.7 Part of the information page. Page includes the input information, results of the calculations and different curves and graphs.

3. OPERATION OF THE PROGRAM AND CALCULATIONS THEORY

The calculations of the program are done on the basis of the system information, the plate information of devices, and the flow profile. The nominal values of the devices are used to evaluate the system characteristics so that all system variables can be described as a function of flow rate. In this chapter, all equations and mathematical methods used to get from the initial pump and process information to the calculated energy consumption and achievable savings are introduced.

3.1 Pump

The calculation process starts from the modelling of the pump and system. On the basis of pump nameplate and system information, program creates the characteristics model for the pumping system, so the required shaft powers for the pump with valve and VSD control can be solved at different flow rates. The entire calculation process for the pump and the system is illustrated in Fig. 3.1 and it is described in detail later in this chapter.

Throttle VSD

-Flow profile -Pump nameplate

information -System information

Pump curve based on specific speed

System curve

Rotational speed from pump curve fitting

Efficiency curve based on specific speed

Efficiency by Sârbu&Borza equation nq

n_q ,ηn

ρ η

H_sys n

Q, Hpump

Shaft power with VSD Shaft power with throttle

η Qn ,Hn

Hst

ηn

Q_n ,H_n

Fig. 3.1 Calculation process for the pump. Head, efficiency, rotational speed, and shaft power values for two control methods are calculated as a function of flow rate, based on the pumping system char- acteristics model created from the pump nameplate and the system information.

In almost all cases, the objective of a pump is to transfer liquid. Making the liquid flow requires production of pressure to overcome the losses in the system. These losses are de- scribed by the system head Hsys [m], which is divided into static and friction head. The static head Hst [m] is the height difference between the supply and destination reservoirs’ surface levels and friction head Hfr [m] is the loss caused by the friction of the liquid movement. The resulting system curve is calculated by the equation

𝐻_𝑠𝑦𝑠 = 𝐻_𝑠𝑡+ 𝐻_𝑓𝑟 = 𝐻_𝑠𝑡+ 𝑘𝑄², (3.1)

where k is the coefficient for friction head and Q [m³/h] is the flow rate. The locations of static and friction heads in a pumping system are illustrated in Fig. 3.2. The static head is assumed to remain constant, so the difference of reservoirs’ surface levels isn’t changing during the pumping.

H_st Hfr

Fig. 3.2 Pumping system with static and friction head. The static head Hst is the height difference between the supply and destination reservoirs’ surface levels and the friction head Hfr describes the friction of moving liquid through the pipes, valves, and other equipment in the system. The static head is assumed to remain constant during pumping, but the friction head grows quadratically as a function of the flow rate.

In addition to the system curve, also the pump curve is needed to describe the operating points of the pump in QH plane. Pump characteristic curves are usually published by the manufacturer and they describe the performance of the pump for instance at different rota- tional speeds and impeller sizes.

The operating point of the pump is always at the intersection of the pump- and system curves and two different ways to change it, throttle and variable speed drive control, are discussed in this thesis. With throttle control the pump is driven at fixed speed and the operating point is changed by throttling and opening the valve. When the valve is throttled, the friction head is increased due to greater flow resistance, which causes reduction of flow rate. With variable speed drive, the flow is controlled by changing the rotational speed of the pump so the flow resistance can be kept in the optimal value. The flow reduction with both control methods is illustrated in Fig. 3.3.

THROTTLE CONTROL

VSD CONTROL

Fig. 3.3 Example of flow control with throttle and VSD in QH plane. Produced flow rates in both graphs are 60, 80 and 90 m³/h. The operating point of the pump is always at the intersection of the red pump curve and the blue system curve. The power produced by the pump at each operating point, and pump rotational speed at each pump curve are shown in graphs.

0 20 40 60 80 100 120

0 2 4 6 8 10 12 14 16 18 20

Q (m³/h)

H (m)

P=3.5 kW P=2.9 kW

Valve is throttled

P=3.7 kW

n=1450 rpm

0 20 40 60 80 100 120

0 2 4 6 8 10 12 14 16 18 20

Q (m³/h)

H (m)

P=3.7 kW

P=2.6 kW

P=1.2 kW

n=1300 rpm

n=1000 rpm n=1450 rpm

As Fig. 3.3 illustrates, reducing the flow with throttle control will always cause greater head losses than with variable speed drive, because of increased flow resistance in the system.

Therefore, when system is controlled with VSD, same flow rate can be usually produced with less power than with throttle control. This is the main reason for VSD being more effi- cient control method than throttling.

3.1.1 Throttle controlled system

In a throttle controlled system, rotational speed is kept constant so the operating point is changed by moving the system curve as shown in Fig. 3.3. Therefore the head values are following the pump curve as a function of flow rate. The actual shape of the pump head and efficiency curves can be approximated on the basis of the pump specific speed. Specific speed nq is a dimensionless quantity that is used to describe the centrifugal pump character- istics regardless of pump size, and it’s defined by the equation

𝑛

= 𝑛

𝐻_𝑛^3⁄4, (3.2)

where n [rpm] is the rotational speed, Qn [m³/s] is the nominal flow rate of the pump, and Hn

[m] is the nominal head of the pump.

The program creates nq based head and efficiency curves for pump on the basis of the digit- ized relative curves that are represented in J.F. Gülich’s Centrifugal Pumps (2008). The curves are relative to the pumps’ nominal head, efficiency, and flow rate values and they are digitized for specific speeds 20, 60, 100 and 250. Curves are digitized with the program Engauge Digitizer 4.1, which can be used to convert a graph to CSV file containing the values of the curve. The average resolution of digitized values is 2.52 % in the range 0 % to 120 % of relative flow rate (Q/Qn). The curves are shown in Fig. 3.4.

Fig. 3.4 Relative head and efficiency curves based on pump specific speed nq. Curves are digitized from J.F.

Gülich’s Centrifugal Pumps (2008). Curves are covering the range of specific speeds from 20 to 250 and they are relative to the pumps’ nominal head, efficiency and flow rate values.

Even though there are digitized curves for only four specific speed values, head and effi- ciency curves can be approximated for all specific speeds ranging from 20 to 250 by linear interpolation. If specific speed of the pump is for example 35, head- and efficiency curves for the pump are interpolated from the digitized curves, in this instance from curves nq=20 and nq=60. This is illustrated in Fig. 3.5.

0 20 40 60 80 100 120 140

0 50 100 150 200 250 300

Q/Qn (%) H/Hn (%)

n_q=20 nq=60 nq=100 nq=250

0 20 40 60 80 100 120 140

0 10 20 30 40 50 60 70 80 90 100 110

Q/Qn (%)

/n (%)

nq=20 nq=60 nq=100 nq=250

Fig. 3.5 Solving the head curve for certain specific speed. In the figure above, head curve for the pump with specific speed nq=35 is created from the digitized curves for specific speeds nq=20 and nq=60 by linear interpolation. The efficiency curve is created the same way.

With throttle controlled system, pump head and efficiency values at each flow rate are solved from created curves. Values are relative so they are multiplied by the nominal values.

3.1.2 VSD controlled system

When the pump is controlled with VSD, the flow rate is regulated by changing the rotational speed so the flow resistance can be kept at the optimal value. Therefore the operating points are following the system curve, while pump curve is varied like shown in Fig. 3.3. Head values at different flow rates are solved from the system curve for VSD controlled system.

With VSD, pump rotational speed at each flow rate is necessary to calculate, because it has effect on the operation of pump, motor and drive. Rotational speed is dependent on head and flow rate. A low specific speed pump’s head curve has typically a parabolic shape so it can be fitted to the form (Leonow 2013)

𝐻₀(𝑄₀) = 𝐴𝑄₀²+ 𝐵𝑄₀+ 𝐶, (3.3)

where Q [m³/h] is the flow rate and H [m] is the head. The fitting is created by solving the coefficients A, B and C by using the least squares fitting method. With the high specific speed axial flow pumps, this fitting is not as accurate as with the low specific speed pumps,

0 20 40 60 80 100 120 140

0 50 100 150 200 250 300

Q/Qn (%) H/H n (%)

nq=20 nq=35 nq=60

but it is still used in the program for all pumps for simplicity. Rotational speed can be taken into account by extending the created fitting with the affinity laws

𝐻 = 𝐻₀(_𝑛^𝑛

0)², (3.4)

𝑄 = 𝑄₀(_𝑛^𝑛

0) . (3.5)

When pump efficiency is assumed to remain constant, extension leads to the equation (Le- onow 2013)

𝐻 = 𝐴𝑄²+ 𝐵𝑄 (_𝑛^𝑛

𝑜) + 𝐶 (_𝑛^𝑛

𝑜)². (3.6) Head values are already calculated from the system curve with flow rate values, so the rota- tional speed n can be solved with the quadratic formula

𝑛 =

2𝐶

𝑛

Because of rotational speed of the pump is varied, the digitized curves used to approximate the efficiency values with throttle control cannot be used with VSD. However, there exist several equations that can be used to estimate the pump efficiency as a function of rotational speed. In designed program, the efficiency is calculated by the equation (Sârbu 1998)

𝜂 = 1 − (1 − 𝜂_𝑛) (^𝑛_𝑛^𝑛)^0.1, (3.8)

where η [%] is the efficiency of the pump.

When flow rate, head and pump efficiency values are known, required shaft power from motor Pshaft [W] with two control methods can be calculated by the equation

𝑃

=

𝜂 , (3.9)

where ρ [kg/m³] is the liquid density and g is the acceleration due to gravity (~9.81 m/s²).

3.2 Motor and drive

Pumps are usually driven by an induction motor which can be connected to a frequency converter or directly to the grid. In motor and drive, there occurs some power losses so they have an effect on the required electric power. They still have usually much better efficiency factors than pumps. In throttle controlled system, the motor is connected directly to the grid, so only motor efficiency is taken into account. With VSD control, both motor and frequency converter are affecting the total efficiency. The calculation process of the required electric power at different operating points, on the basis of motor and drive information, and pump shaft power and rotational speed, is illustrated in Fig. 3.6 and described in detail in this chapter.

Throttle VSD

-Motor information -Drive information

-Shaft power -Rotational speed

Calculation of torque

n, P_shaft

Electric power with throttle

Motor efficiency from relative effiency map

Electric power with VSD Drive effiency by the equations (3.11)-(3.13)

T

n, Pmotor(nom), Pshaft

T

ηdrive

ηmotor

P_shaft

Pshaft

ηmotor

Fig. 3.6 Calculation process for motor and drive. The consumed electric power and the efficiencies of the motor and drive are calculated on the basis of rotational speed, shaft power and motor and drive information.

Efficiencies of the motor and drive in each operating point are calculated on the basis of nominal efficiency of device, rotational speed and torque. Torque T [Nm] is calculated by the equation

𝑇 =

2𝜋₆₀^𝑛 . (3.10)

Efficiency of the motor is calculated on the basis of an efficiency map. The program uses only one motor efficiency map for the sake of simplicity, so the compromise had to be made, whether the map is more suitable for small or large motors. The map was chosen on the basis of the study made by VTT (2008). According to the study, 87 % of induction motors at Kaukas (Lappeenranta, Finland) and Kaukopää (Imatra, Finland) factories had a power rat- ing of less than 37 kW. Therefore the motor, which efficiency map is used in the program, was decided to have power rating of 37 kW. The total number of induction motors at these two factories is about 15 000, so the chosen efficiency map should correspond to the average motor quite well and be also a good compromise between small and large motor. Information about motor used for map, ABB M4BP 225SMA 4-pole, is shown in Table 3.1.

Table 3.1 Motor catalogue information of ABB M4BP 225SMA 4-pole induction motor.

ABB M4BP 225SMA 4-poles

IEC frame size 225

Nominal input power 37 kW

Nominal current 65 A

Nominal voltage 400 V

Nominal rotational speed 1482 rpm Synchronous rotational speed 1500 rpm

Power factor 0.87

Frequency 50 Hz

IE energy efficiency class IE3

The efficiency map for the motor is made by approximating the equivalent circuit of the motor on the basis of the motor catalogue information shown in Table 3.1. The equivalent circuit is then used to create the efficiency map for an optimal flux control scheme, but with- out taking into account the losses caused by non-sinusoidal supply voltage (Tamminen 2013). In the program, rotational speed, torque and efficiency values of the map are trans- formed into relative form, as in Fig. 3.7.

Motor efficiency ηmotor [%] can be determined from the efficiency map by relative torque and rotational speed values. In the program, relative efficiency is calculated by linear inter- polation from the two-dimensional table that consists the values of the map. In the motor efficiency table, the average interval between rotational speed values is 10.5 % in the range

of 0 % to 300 % of nominal rotational speed and interval between torque values is 6.9 % in the range of 0 % to 200 % of nominal torque.

Fig. 3.7 Relative motor efficiency map used in the program. Relative efficiency is calculated by linear in- terpolation based on rotational speed and torque relative to nominal values.

In the case of the speed controlled system, also frequency converter efficiency has to be taken into account. Rough approximation of efficiency of the drive at different operating points can be made by the equations published by ABB Automation Group Ltd (2002). Pdrive nomloss [W] is the power loss at the frequency converter’s nominal point and it’s defined by the equation

𝑃𝑑𝑟𝑖𝑣𝑒 𝑛𝑜𝑚𝑙𝑜𝑠𝑠 = 𝑃𝑑𝑟𝑖𝑣𝑒 𝑜𝑢𝑡 (𝑛𝑜𝑚)(^1−𝜂𝑑𝑟𝑖𝑣𝑒 (𝑛𝑜𝑚)

𝜂𝑑𝑟𝑖𝑣𝑒 (𝑛𝑜𝑚) ), (3.11)

where Pdrive out (nom) [W] is the drives nominal output power and ηdrive (nom) is the nominal efficiency of the drive. Power losses of the drive Pdrive loss [W] are calculated by the equation

𝑃_{𝑑𝑟𝑖𝑣𝑒 𝑙𝑜𝑠𝑠} = (0.35 + 0.1 ∗ (_𝑛^𝑛

𝑛) + 0.55 ∗ (_𝑇^𝑇

𝑛)) ∗ 𝑃𝑑𝑟𝑖𝑣𝑒 𝑛𝑜𝑚𝑙𝑜𝑠𝑠. (3.12)

0.50.5 0.70.7

0.7

0.8 0.8

0.8

0.88 0.88

0.88 0.88

0.9 0.9

0.9 0.9

0.92

0.9

2 0.92

0.92 0.9

2

0.93

0.93 0.93

0.93 0.9

3

0.94

0.94

0.94 0.94

0.95

0.95

0.95 0.95

0.96

0.9 6

0.96 0.96

0.97

0.97

0.9 7 0.97

0.98

0.98

0.98

0.9

9

0.9 9

0.99 0.99

ABB M4BP 225SMA 4-poles

n/n_n (%) T /Tn (%)

0.5 1 1.5 2 2.5 3

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

In equation (3.12), the relative frequency in equation published by ABB is substituted with the relative rotational speed. This is done for simplicity, and the substitution is valid if a slip of an induction motor is assumed to remain constant. Even if slip varies, it has no significant effect on power consumption.

When the drive losses are known, the efficiency of the drive can be calculated by the equa- tion

𝜂

=

𝑑𝑟𝑖𝑣𝑒 𝑜𝑢𝑡+𝑃_{𝑑𝑟𝑖𝑣𝑒 𝑙𝑜𝑠𝑠}. (3.13) Consumed electric power Pe [W] can be calculated by dividing shaft power by efficiencies of devices between grid and the pump. With throttle control, there is only a motor between them, but with speed control, there is also a frequency converter. Electric power consump- tions with the two control methods are calculated by the following equations.

𝑃

=

𝑚𝑜𝑡𝑜𝑟 , (3.14)

𝑃

=

𝑚𝑜𝑡𝑜𝑟𝜂_{𝑑𝑟𝑖𝑣𝑒} , (3.15)

where Pe,throttle [W] is the electric power consumption with throttle control and Pe,VSD [W]

with VSD control.

3.3 Energy consumption and economics

Energy consumption and economic information are the end results of the calculations shown to user. User can evaluate the viability of the investment in VSD on the basis of these calcu- lations.

Annual energy consumption Wannual [kWh] is calculated for both control methods by the equation (3.16). Whatever is the duration of the given flow profile, it is assumed to be re- peated during the entire year.

𝑊

=

𝑡_𝑡𝑜𝑡

∗ ∑ 𝑃

𝑡

where ttot [h] is the total duration of given flow information, and Pe,i [kW] and ti [h] are the electric power consumption and the duration of the certain operating point.

Annual savings Sannual [€/a] achieved from energy savings are calculated by multiplying the difference of annual energy consumptions by price of the energy pE [€/kWh], as in following equation

𝑆_{𝑎𝑛𝑛𝑢𝑎𝑙} = 𝑝_𝐸∗ (𝑊𝑎𝑛𝑛𝑢𝑎𝑙,𝑡ℎ𝑟𝑜𝑡𝑡𝑙𝑒− 𝑊_{𝑎𝑛𝑛𝑢𝑎𝑙,𝑉𝑆𝐷}) . (3.17)

When calculating the lifetime savings available, current economic conditions are taken into account by calculating the net present value (NPV) of savings. This requires knowing the nominal interest rate, which takes into consideration the effects of the real interest rate and the inflation. The nominal interest rate i can be defined from the Fisher equation (Fisher 1896)

𝑖 = (1 + 𝑟)(1 + 𝜋) − 1, (3.18)

where r is the real interest rate and π is the inflation. Both of these are assumed to remain constant during the lifetime of the pumping system. The lifetime savings can be then calcu- lated by the equation (3.19), where the investment cost Cinvestment [€] is subtracted from NPV of savings.

𝑆_{𝑙𝑖𝑓𝑒𝑡𝑖𝑚𝑒} = ∑^𝑙 (^{𝑆𝑎𝑛𝑛𝑢𝑎𝑙}_(1+𝑖)𝑙 )

𝑘=1 − 𝐶_{𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡} , (3.19) where l [a] is the lifetime of the pumping system.

Internal rate of return (IRR) is a commonly used rate quantity that describes the profitability of investments. In this case, IRR is the discount rate that makes the NPV of all savings achieved from the investment equal to zero. IRR does not take into account external factors like inflation and interest rate. IRR is solved by iterating it from the equation

∑^𝑙 (_{(1+𝐼𝑅𝑅)}^{𝑆𝑎𝑛𝑛𝑢𝑎𝑙}_𝑙)

𝑘=1 − 𝐶_{𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡}= 0 . (3.20)

The discounted payback period of the investment tpayback [a] is the time, when the investment has paid for itself. To calculate the discounted payback period, number of years with nega- tive cumulative savings m [a] is needed to iterate from the equation

∑^𝑚 (^𝑆_(1+𝑖)^{𝑎𝑛𝑛𝑢𝑎𝑙}_𝑚)

𝑘=1 − 𝐶_{𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡} < 0 . (3.21) The payback period can be then calculated by the equation

𝑡

= 𝑚 +

𝑎𝑛𝑛𝑢𝑎𝑙⁄(1+𝑖)^𝑚+1 , (3.22)

where Sm [€] is the cumulative savings at the end of the period of m years.

Energy saving is also beneficial for the environment by emission reduction. Annual CO2

reduction is calculated by the equation

𝑆_𝐶𝑂2 = 𝑒_𝐶𝑂2∗ (𝑊𝑎𝑛𝑛𝑢𝑎𝑙,𝑡ℎ𝑟𝑜𝑡𝑡𝑙𝑒− 𝑊_{𝑎𝑛𝑛𝑢𝑎𝑙,𝑉𝑆𝐷}) , (3.23)

where eCO2 [kg/kWh] is the CO2 emissions per kWh.

4. EVALUATION OF THE RESULTS CALCULATED BY PROGRAM

In this chapter, the models and results calculated by the designed program SCCP are com- pared with the manufacturers’ pump performance curves, ABB PumpSave’s results and la- boratory measurements. It should be noted that the comparison is done only for three man- ufacturer’s curves and one real pumping system, and the accuracy of the calculations may vary depending on case.

4.1 Comparison with manufacturer’s performance curves

The accuracy of the pump model used in the program was verified by comparing it to the pump performance curves published by manufacturer. Because the pump modelling in SCCP is based on the pump nominal values and specific speed, the pumps for comparison were selected from a wide range of specific speeds. Selected pumps were Sulzer APP41-300 C with 335 mm closed impeller, APP51-300 with 430 mm closed impeller, and APP55-200 C with 625 mm closed impeller. The nominal values of these pumps are shown in Table 4.1.

Table 4.1 Nominal values of the pumps selected for comparison. Values are available on Sulzer Select online tool.

Values at best efficiency point APP41-300 C APP51-300 C APP55-200 C

Flow rate (m³/h) 1046.8 1058.2 1033.4

Head (m) 19.04 19.22 125.9

Rotational speed (rpm) 1483 980 1489

Efficiency (%) 84.26 86.36 78.59

Power (kW) 63.31 64.04 450

Specific Speed 85.73 57.88 21.23

The comparison was made for head, efficiency and power curves. The performance curves for selected pumps were digitized with Engauge Digitizer 4.1 from curves that are available on Sulzer Select, which is an online tool for pump selecting. Example of pump information and curves is shown in APPENDIX IV.

To verify the accuracy of the curves created by SCCP and PumpSave, maximum deviations in head, efficiency and power values were studied in the entire flow rate range and in pre- ferred operation region (POR). Preferred operation region for centrifugal pumps is between 70% and 120% of the nominal flow rate. When pump is used in the POR, adverse hydraulic loads, vibration and impeller flow recirculation are not significantly effecting on the lifetime

of the pump (Ferman 2008). Because digitized manufacturer’s curves and curves created by programs have different sets of points for curves, the maximum deviations were solved from fittings created from curves. This means deviation values are not perfectly accurate, but ra- ther indicative.

The manufacturer’s pump performance curves used in comparison are measured to fulfill the requirements of grade 2B accuracy criteria in ISO 9906 standard (ISO, 2012). It deter- mines the allowable tolerances by which the actual operating point can deviate from the values of the published curves. The allowed deviations from the guaranteed values are shown in Table 4.2.

Table 4.2 Allowed deviations from the guaranteed values in grade 2B accuracy criteria in ISO 9906 standard (2012). The tolerances are relative deviations from the actual operating point.

ISO 9906:2012 Grade 2B

ΔQ ± 8 %

ΔH ± 5 %

ΔP + 8 %

Δη - 5 %

4.1.1 Head curves

In creation of head curves, SCCP uses relative specific speed based curves, and PumpSave uses quadratically decreasing curve from zero flow rate and maximal head to nominal oper- ating point. SCCP requires only nominal operating point information to create the curve, but PumpSave also requires the maximal head value at zero flow rate (ABB 2012).

As Fig. 4.1 illustrates, pump head curves generated by SCCP differs from manufacturer’s curve especially at low flow rates and it seems to be more suitable for low specific speed pumps. The curves generated by ABB PumpSave are almost perfectly accurate, but on the other hand it requires maximal head value to create the curve.

Fig. 4.1 Comparison of the pump head curves by manufacturer and programs. ABB PumpSave’s head curves seem to be more accurate than SCCP’s.

The maximum deviation from the manufacturer’s head curves are shown in Table 4.3. The head curves created by SCCP are not very accurate, except near the nominal operating point.

The reason for this is that the profiles of manufacturer’s curves look like specific for pumps with much lower specific speeds, than calculated from pump nominal values. This may in- dicate that relative specific speed based curves used in program are not valid for modern centrifugal pumps. Another possible reason is that manufacturer’s curves are too optimistic in head losses at low flow rates.

0 10 20 30 40

0 200 400 600 800 1000 1200 1400 1600

H (m)

Sulzer APP41-300 C (n_q=87.73)

Sulzer SCCP ABB Nominal operating point

0 5 10 15 20 25 30 35

0 200 400 600 800 1000 1200 1400 1600 1800

H (m)

Sulzer APP51-300 C (n_q=57.88)

0 20 40 60 80 100 120 140 160 180

0 200 400 600 800 1000 1200 1400 1600 1800

H (m)

Q(m³/h)

Sulzer APP55-200 C (n_q=21.23)

Table 4.3 The maximum deviation of head curves generated by programs from manufacturer’s curves. The maximum deviation is defined in entire flow rate range and in preferred operation region. Next to absolute value in parentheses is shown its share of the nominal head value.

Head SCCP PumpSave

Pump Entire range POR Entire range POR

APP41-300 C 6.3m (33.1%) 3.1m (16.3 %) 1.0m (5.3 %) 1.0m (5.3%) APP51-300 C 5.8m (30.1%) 1.6m (8.3%) 0.6m (3.1 %) 0.6m (3.1 %) APP55-200 C 16.5m (13.1 %) 16.5m (13.1 %) 6.3m (5.0 %) 6.3m (5.0 %)

4.1.2 Efficiency curves

The designed program uses relative specific speed based curves also in generation of pump efficiency curves. PumpSave uses experience based equation to create the correcting factor for nominal efficiency value (ABB 2012). Basically, the idea is the same in both programs, but PumpSave uses only one efficiency curve for all pumps, while SCCP creates the curve by interpolating it from four digitized curves. As shown in Fig. 4.2, pump efficiency curves created by SCCP are corresponding to the manufacturer’s curves quite well. Especially with high specific speed pumps, SCCP’s pump efficiency model seems to be better than PumpSave’s model.

Fig. 4.2 Comparison of the pump efficiency curves by manufacturer and programs. SCCP’s efficiency curves seem to be more accurate than ABB PumpSave’s.

The maximum deviation from manufacturer’s efficiency curves are shown in Table 4.4. Ef- ficiency model in SCCP seems to be better than head model. SCCP calculates the efficiency more accurate than PumpSave with higher specific speed pumps. With low specific speed pump APP55-200 C, PumpSave is a little more accurate.

0 20 40 60 80 100

0 200 400 600 800 1000 1200 1400 1600

η(%)

Sulzer APP41-300 C (n_q=87.73)

Sulzer SCCP ABB Nominal operating point

0 20 40 60 80 100

0 200 400 600 800 1000 1200 1400 1600 1800

η(%)

Sulzer APP51-300 C (n_q=57.88)

0 20 40 60 80 100

0 200 400 600 800 1000 1200 1400 1600 1800

η(%)

Q(m³/h)

Sulzer APP55-200 C (n_q=21.23)

Table 4.4 The maximum deviation of the efficiency curves generated by programs from manufacturer’s curves. The maximum deviation is defined in entire flow rate range and in preferred operation re- gion. Next to absolute value in parentheses is shown its share of the nominal efficiency value.

Efficiency SCCP PumpSave

Pump Entire range POR Entire range POR

APP41-300 C 7.9 % point (9.4 %) 7.5 % point (8.9 %) 13.6 % point (16.1 %) 8.6 % point (10.2 %) APP51-300 C 5.1 % point (5.9 %) 2.3 % point (2.7 %) 10.3 % point (11.9 %) 10.3 % point (11.9 %) APP55-200 C 5.8 % point (7.4 %) 1.8 % point (2.3 %) 5.7 % point (7.3 %) 1.2 % point (1.5 %)

4.1.3 Power curves

Pump power curve can be formed on the basis of the pump head and efficiency curves by the equation (3.9). The pumped liquid is water in manufacturer’s power curves, so the liquid density used in calculations is 998 kg/m³. The power curves created by SCCP seem to be significantly more accurate with high specific speed pumps than PumpSave’s curves. With low specific speed pump APP55-200 C, PumpSave’s power curve is a little more accurate.

Power curves are shown in Fig. 4.3.

Fig. 4.3 Comparison of the pump power curves by manufacturer and programs. SCCP’s power curves seem to be generally more accurate than ABB PumpSave’s.

The maximum deviation from manufacturer’s power curves are shown in Table 4.5.

0 10 20 30 40 50 60 70

0 200 400 600 800 1000 1200 1400 1600

P (kW)

Sulzer APP41-300 C (n_q=87.73)

Sulzer SCCP ABB Nominal operating point

0 20 40 60 80

0 200 400 600 800 1000 1200 1400 1600 1800

P (kW)

Sulzer APP51-300 C (n_q=57.88)

0 100 200 300 400 500 600 700

0 200 400 600 800 1000 1200 1400 1600 1800

P (kW)

Q(m³/h)

Sulzer APP55-200 C (n_q=21.23)

Table 4.5 The maximum deviation of power curves generated by programs from manufacturer’s curves. The maximum deviation is defined in entire flow rate range and in preferred operation region. Next to absolute value in parentheses is shown its share of the nominal power value.

Power SCCP PumpSave

Pump Entire range POR Entire range POR

APP41-300 C 12.1 kW (19.1 %) 2.4 kW (3.8 %) 25.2 kW (39.8 %) 9.4 kW (14.8 %) APP51-300 C 4.2 kW (6.6 %) 4.2 kW (6.6 %) 9.7 kW (15.1 %) 6.0 kW (9.3 %) APP55-200 C 86.3 kW (19.2 %) 23.3 kW (5.2 %) 38.1 kW (8.5 %) 9.7 kW (2.2 %)

On the basis of the comparison with manufacturer’s curves, PumpSave seems to have better head curve model and SCCP better efficiency model. SCCP also seems to give more realistic shaft power consumption results for wide range of specific speeds, while PumpSave is work- ing only with low specific speed pumps. It should be noted that comparison is done only with pumps of one manufacturer and the accuracy of published pump performance curves is only verified to be within ISO 9906:2012 standard Grade 2B.

4.2 Laboratory measurements

In addition to the comparison made with manufacturer’s curves, the curves created by pro- gram were also verified by comparing them to a laboratory measured results. The objective of the laboratory measurements was to monitor head and power curves of throttle and VSD controlled systems.

The pumping system used for measurements in the LUT pump laboratory includes Sulzer APP 22-80 centrifugal pump with 255 mm open impeller, ABB M3BP160M4 induction mo- tor and ABB ACS880-01-032A-3+L503 frequency converter. Piping and reservoir of the pumping system are equipped with multiple pressure, flow and temperature sensors. Pump shaft is equipped with torque and rotational speed sensors. Consumed electric power was measured with Fluke-1735 Power Logger. The used equipment in laboratory is described more detailed in APPENDIX I.

Measurements were made for open and closed loop systems and water was used as pumped liquid. The open loop system includes static head that was measured to be about 5.8 meters.

The closed loop system doesn’t have static head at all.

The nominal values for the pump were defined by measurements and they differed from values specified by the manufacturer. Measurements are described in APPENDIX II. The nominal values were used for input information to the programs and they are shown in Table 4.6.

Table 4.6 Pumping system information inputted in programs. Nominal values of the pump were defined by laboratory measurements. For motor and drive efficiencies, estimated values were used.

Pumping system information Nominal flow rate 90 m³/h

Nominal head 16.5 m

Maximal head 22.5 m

Pump efficiency 66 %

Liquid density 998 kg/m³

Static head 0 m/ 5.8 m

Rotational speed 1450 rpm

Specific speed 28.0

Nominal motor power 11 kW Nominal motor efficiency 92 % Nominal drive efficiency 98 %

4.2.1 Throttle control

Throttle controlled system has identical head and power curves in open and closed loop sys- tem, so the measurements were done only for open loop system. The curves created by SCCP and PumpSave are almost identical and they correspond to the measured curves quite well as Fig. 4.4 illustrates.