Available online at www.sciencedirect.

com

ScienceDirect
Available online at www.sciencedirect.com
Energy Procedia 00 (2017) 000–000
www.elsevier.com/locate/procedia
ScienceDirect
Energy Procedia 128 (2017) 10–16
www.elsevier.com/locate/procedia

International Scientific Conference “Environmental and Climate Technologies”, CONECT 2017,

10–12 May 2017, Riga, Latvia

Energy policy for energy intensive manufacturing companies and its

impact on energy efficiency improvements. System dynamics approach
Kristaps Locmelis*, Andra Blumberga, Uldis Bariss, Dagnija Blumberga
Institute of Energy Systems and Environment, Riga Technical University, Azenes iela 12/1, Riga, LV–1048, Latvia

Abstract

In this paper we evaluate Latvian Energy policy for energy intensive manufacturing companies and its impact on energy efficiency
improvements using system dynamic approach. Since mid-2015, energy policy developments in Latvia are targeted to reduce
energy costs for energy-intensive companies while imposing an obligation to implement an energy management system, which
would increase the overall efficiency of the manufacturing industries. Simulation of theoretical energy-intensive manufacturing
company behaviour showed interest in energy efficiency measures until its energy intensity closes to the benchmark stated in the
energy policy, and then following corrective actions are taken to maintain the energy intensity above the benchmark. This results
in lost energy efficiency savings. To minimize energy efficiency effect on the energy intensity calculation, the amendments to the
energy policy is offered and its impacts are simulated using system dynamic modelling.
© 2017 The Authors. Published by Elsevier Ltd.
© 2017 The Authors. Published by Elsevier Ltd.
Peer-review under responsibility of the scientific committee of the International Scientific Conference “Environmental and
Peer review statement - Peer-review under responsibility of the scientific committee of the International Scientific Conference
Climate Technologies.
“Environmental and Climate Technologies”.
Keywords: energy intensive industry; policy evaluation; system dynamics

1. Introduction

Currently there are little doubts that energy costs are a major factor for the competitiveness of manufacturing
industry. Study [1] had shown than different levies related to support schemes are currently raising the price of

* Corresponding author. Tel.: +371-67728371.

E-mail address: kristaps.locmelis@edu.rtu.lv

1876-6102 © 2017 The Authors. Published by Elsevier Ltd.

Peer-review under responsibility of the scientific committee of the International Scientific Conference “Environmental and Climate
Technologies.

1876-6102 © 2017 The Authors. Published by Elsevier Ltd.

Peer review statement - Peer-review under responsibility of the scientific committee of the International Scientific Conference
“Environmental and Climate Technologies".
10.1016/j.egypro.2017.09.005
 Kristaps Locmelis et al. / Energy Procedia 128 (2017) 10–16 11
2 Author name / Energy Procedia 00 (2017) 000–000

electricity and subsequently the energy costs of industries. The most affected are considered to be energy-intensive
industries. Nowadays costs of European Union energy transition as well as new investments in conventional power
generation are passed to the consumers mostly via energy prices. To limit the burden, the European Commission has
adopted harmonized rules on how Member States can relieve energy-intensive companies that are particularly exposed
to international competition from charges levied for the energy support schemes [2]. The description of European
Union harmonized rules and Latvian energy policy on energy intensive industries can be found in the previous article
by the authors [3].
Latvian energy policy for energy intensive industries stimulates energy-intensive companies to implement energy
efficiency measures, yet maintaining rather high level of energy intensity, which are mutually contradictory.
Implementation of an energy management system is one of the key aspects for removing barriers for energy efficiency
measures at the company’s level. Studies [4–10] indicate that energy management systems facilitate new energy
metering and monitoring methods, which play a significant role to achieve energy savings. At company’s level
decisions of implementation of energy efficiency measures would be based on the business cases of measures with
time lags of decision making and implementations of measures, effectively achieving energy efficiency saving
compared to the baseline scenario and gaining financial benefits from reduced consumption. On the other hand, energy
efficiency measures will reduce the company's power consumption and energy costs, subsequently reducing its energy
intensity. If in this case the energy intensity goes below the policy threshold and the increase of electricity costs due
to loss of renewable energy levy repayments outweighs the efficiency gains, energy intensive companies most likely
would sacrifice efficiency measures just to keep intensity above the threshold or even deliberately increase electricity
consumption in order to maintain energy intensity above the benchmark. From the energy policy point of view this is
undesirable consequence that should be evaluated and addressed.
Evidently the undesirable consequences of energy policy occur, where the energy intensity of a company is close
to the benchmark set in the policy and, as a result of changes of various external or internal parameters, energy
intensity would cross the benchmark. If energy intensity of a company is well above or below the benchmark, the
undesirable policy consequences most likely would not appear. Taking into account this consideration modelling
is focused on case, where there is evidence of potential energy policy failure, namely on case where initial energy
intensity of a company is close to the benchmark and relevant policy improvements are evaluated taking into
consideration this aspect.

2. Methodology

Taking into account the complexity of company’s potential behaviour, a system dynamic model is proposed in this
study. System dynamic modelling can simulate the behaviour of energy-intensive companies with different input
parameters and assess appropriateness of Latvian energy policy. System dynamic approach also provides the necessary
tools to simulate different improvements in policy [11]. The modelling period is set for 7 years until 2023, which
includes the period where the energy policy is in place and period without it.
Electricity consumption in manufacturing industry depends on various factors. Causal loop diagram is shown in
Fig. 1(a). The stock correspond the company's electricity consumption, which increases every year as a result of
growth of company’s production output, while the outgoing flow of electricity consumption represents energy
efficiency measures. Normally the model would have two causal loops – reinforcing loop (R1) and balancing loop
(B1). The strength of two loops is determined by the consumption growth factor, which largely depends on company’s
gross value added (GVA) and gross domestic product (GDP) growth, and energy efficiency factor, which depends on
company’s identified energy efficiency potential to be implemented.
Since energy efficiency measures balance the consumption growth, but does not affect growing gross value added,
company's energy intensity, which is calculated as total electricity costs to GVA, are decreasing. The manufacturing
company may not be advantageous to reduce its energy intensity below the benchmark of 0.2, in which case it loses
the energy-intensive status and consequently economic benefits of renewable energy levy repayment. If the company's
savings from energy efficiency (reduce energy costs) is less than the levy refund, then the company should adopt a
rational decision to stop delivering energy efficiency measures and to keep the energy intensity above benchmark.
Consequently the system dynamic model has an additional reinforcement and balancing loops that essentially weakens
the main balancing loop (B1) (see red causal links in Fig. 1a). This system conform to Senge (1990) “fixes that fail”
12 Kristaps Locmelis et al. / Energy Procedia 128 (2017) 10–16
Author name / Energy Procedia 00 (2017) 000–000 3

archetype [12]. In particular, energy-intensive manufacturing companies are initially created an incentive (renewable
energy levy repayments) to implement energy efficiency measures to improve their competitiveness, however, the
solution with delay might lead to a situation where energy efficiency measures are no longer implemented (even if
they are reasonably cost effective), as to be too effective means not to be energy-intensive.

Energy
price
–
–
+ Energy
costs
GDP
+ +
+ Energy +
(R1) Consumption (B1)
GVA savings
+ –
Decision
+
to take Energy
+
(R2) – measures efficiency
– potential
MPC
repayment
gains Risk of
– (B2)
losing MPC
(R3)
repayment

– Energy +
intensity

(a)

Accumulated savings
Energy efficiency
savings

Initial consumption

Delay time MPC fee Network tariff

Consumption
Consumption Energy efficiency
growth
Implementation
Electricity price Retail price
delay

Electricity costs
Corelation between
manufacturing
consumption and GDP
Decision for energy
growth
efficiency measures Energy intensity

EE benefits
MPC costs reduction Energy intensity to
benchmark

Initial GVA
Energy savings

Energy intensity
GDP manufacturing vs benchmark
GDP national

Max MPC fee

reduction GVA
Energy efficiency
GVA growth
potential
GDP growth
manufacturing

GDP growth

(b)

Fig. 1. Energy-intensive manufacturing company’s decisions on energy efficiency (a) causal loop diagram; (b) SD model.

The structure of system dynamic model is given in Fig. 1(b). The mains stocks in the system dynamic model are
amount of energy-intensive manufacturing company’s electricity consumption, energy-intensive manufacturing
company’s gross value added (GVA) and accumulated energy efficiency savings. Changes in stocks are functions of
the input and output flows, and these flows are dependent on various factors.
 Kristaps Locmelis et al. / Energy Procedia 128 (2017) 10–16 13
4 Author name / Energy Procedia 00 (2017) 000–000

The electricity consumption in a given year (stock) depends on consumption growth factors and energy efficiency
factors; see Eq. (1):

t  2023
E  (E
t  2017
gr  E EE )  dt  Eini (1)

where
E electricity consumption in a given year, MWh;
Egr electricity consumption increase, MWh/year;
EEE energy efficiency measures the company has implemented, MWh/year;
Eini initial electricity consumption, MWh.

Although energy-intensive manufacturing company’s energy consumption in a given year could be used as factor
to evaluate the energy policy outcomes, it is more appropriate to use accumulated new energy efficiency savings as
an energy policy evaluation measure. The volume of accumulated new energy efficiency savings are derived from the
energy efficiency measures the company has implemented, see Eq. (2):

t  2023
E AS  E
t  2017
EE  dt , MWh (2)

2.1. Consumption growth factors

Electricity consumption growth factor correlates with Latvian GDP growth of manufacturing industries and Latvian
national GDP growth. To determine the correlation, Latvian national real GDP (at constant prices), real GDP of
manufacturing industries and electricity consumption of manufacturing industries from 2010 are compared and linear
regression lines are constructed. The difference in the slope of the linear regression lines show the factors that
represents ratio between manufacturing GDP and national GDP and ratio between manufacturing electricity
consumption and manufacturing GDP. The data were taken from publicly available sources, published by Latvian
Central Statistical Bureau (CSB) and summarized in Table 1.

Table 1. Correlation of national real GDP, manufacturing industries real GDP and electricity consumption of manufacturing industries.
Manufacturing*
Manufacturing real Manufacturing*
National real GDP electricity
GDP real GDP
(2010=100) consumption
(2010=100) (2010=100)
(2010=100)
2010 100.00 100.00 100.00 100.00
2011 106.36 105.12 105.55 104.04
2012 109.81 109.67 111.49 106.51
2013 112.04 107.53 111.85 119.54
2014 113.97 107.96 112.30 117.81
2015 116.82 112.59 117.11 118.26
2016 118.41 118.86
Pearson correlation coefficient 0.912 0.850
Slope of linear regression line 2.869 2.494 3.033 4.162
Factor (manufacturing GDP to national GDP) 0.869
Factor (manufacturing consumption to
1.372
manufacturing GDP)
* Except production of basic metals.
14 Kristaps Locmelis et al. / Energy Procedia 128 (2017) 10–16
Author name / Energy Procedia 00 (2017) 000–000 5

The only independent variable in model that influences the consumption growth is national real GDP forecast,
provided by the Ministry of Finance of Republic of Latvia until 2020 [13]. Therefore electricity consumption increase
could be calculated according to Eq. (3):

E gr  E  GDPgr  factorGDPind / GDPnat  factorcons / GDPind (3)

where
GDPgr national real GDP growth, year-1;
factorGDPind/GDPnat factor that represents ratio between manufacturing GDP and national GDP, according to Table 1;
factorcons/GDPind factor that represents ratio between manufacturing electricity consumption and
manufacturing GDP, according to Table 1.

2.2. Energy efficiency factors

The objective of energy management systems is to promote sustainable energy performance enhancement by
introducing a structured approach to the company's operations and procedures. Evaluation of Swedish program for
improving energy efficiency in energy intensive industries [8] indicated that net impact form energy efficiency
measures taken by energy intensive industries involved in the Swedish programme had resulted in energy savings that
reaches 3.4 % of consumption. Taking into account the similarities from the Swedish energy efficiency programme
and obligations arising from energy management systems, the system dynamic model provides a potential of annual
energy efficiency savings at a 3.4 % rate. To simulate the time lag of the management decision making process and
full effect of implemented efficiency measures, a delay functionality is introduced in system dynamic model.

3. Model output

The system dynamic model is created using Powersim simulation tool. The simulation step is 1 year, simulation
period 2017–2023 or 7 years. Since the scope of modelling is energy-intensive company's theoretical behaviour and
evaluation of Latvian energy policy, the initial input data correspond to a manufacturing company that operates in the
segment of non-metallic mineral products (C23 according to NACE Rev. 2 classifier), and with initial energy intensity
of more than 0.2.
To evaluate the output of simulation two indicators are used. First – the accumulated new energy efficiency savings
(Fig. 2(a)), that indicates the outcome of energy efficiency measures, and second – energy intensity that shows the
company’s current state of energy intensity (Fig. 2(b)). Results of the simulation show that, while company’s energy
intensity is above the benchmark of 0.2, company implements energy efficiency measures and accumulated savings
are increasing. When the energy intensity closes to the benchmark, energy efficiency measures are not implemented
and accumulated savings are not increasing. After energy intensity increases above the benchmark, company restarts
to implement energy efficiency measures. The behavior of a company can be explained with short-term economic
benefit company is gaining from renewable energy fee refunding instead of benefit company could get from reduced
energy consumption. With that in mind, there might be even situations, where company is even incentivized to
inefficiently increase consumption just to keep the energy intensity above the benchmark and accumulated savings
would even decrease over time, which is highly undesirable form the policy point of view.
 6 Author name
Kristaps / EnergyetProcedia
Locmelis 00 (2017)
al. / Energy 000–000
Procedia 128 (2017) 10–16 15

15,000 0.23

Energy intensity, MWh

Accumulated saving, MWh

0.22

10,000
0.21

0.20
5,000
0.19

0 0.18
0 1 2 3 4 5 6 0 1 2 3 4 5 6
Years Years
(a) (b)

Fig. 2. Results of simulation (a) accumulated savings (MWh); (b) energy intensity.

4. Modeling the improvement of the policy

Since the energy policy is related to state aid issues in accordance with the Treaty on European Union, Member
State's freedom of action is limited and any improvements in energy policy should comply with the related guidelines
laid down.
The energy intensity benchmark in Latvia is set relatively high. According to the European Commission guidelines,
energy intensity benchmark may also be set at a lower level, such as the 0.16. This apparently seems to be the easiest
way of policy change that would fix the undesirable behaviour of energy-intensive company with energy-intensity
around 0.2. However policy improvement is considered to be one which is least likely to change behaviour in
long-term. It still remains "fixes that fail" archetype character – quick solution that does not address the main problem.
Manufacturing company reaching the next benchmark level of 0.16, would behave the same.

0.23

15,000
0.22
Accumulated saving, MWh

Energy intensity, MWh

0.21
10,000

0.20
5,000
0.19

0 0.18
0 1 2 3 4 5 6 0 1 2 3 4 5 6
Years Years
(a) (b)

Fig. 3. Simulation results after policy improvement (a) accumulated savings (MWh); (b) energy intensity.

Since it is obvious that energy efficiency measures reduces energy consumption and subsequently energy intensity,
it would be suggested, that those avoided energy costs of energy efficiency measures could be added to the actual
energy costs. Thus the energy intensity calculation would reflect the base case scenario without the energy efficiency
measures and company would not be penalized for taking efficiency measures.
Simulation of amended energy policy shows that opportunity to add avoided energy costs of energy efficiency
measures to actual energy costs enables to keep the energy intensity above the benchmark all the simulating period
even if energy efficiency measures are taken (see Fig. 3(b)). Accumulated energy savings (Fig. 3(a)) increases
continuously and the overall achievement at the end of the simulated period is higher that without amendments to
the policy.
16 Kristaps Locmelis et al. / Energy Procedia 128 (2017) 10–16
Author name / Energy Procedia 00 (2017) 000–000 7

5. Conclusions

Findings of this study shows that Latvian energy policy, which aims to increase the competitiveness of energy
intensive manufacturing industry, has an undesirable aspect when energy efficiency measures that arises from
mandatory implementation of energy management system would not be taken, as otherwise the company would not
be energy-intensive and consequently lose the economic benefit of renewable energy fee repayments. Simulation of
theoretical manufacturing company, whose energy intensity is close to the energy intensity benchmark, showed that
manufacturing company is interested in energy efficiency measures until its energy intensity closes to the benchmark.
The following corrective actions are taken to ensure the maintaining of energy intensity above the benchmark
To solve the problem, authors offer to amend the Latvian energy policy by adding avoided energy costs that arises
from energy efficiency measures to the actual energy costs. In such way the energy intensity would be calculated to
the base case scenario if energy efficiency measures would not be taken regardless of the whether they were introduced
or not. The system dynamic simulation shows that amended energy policy outputs better results in form of
accumulated energy efficiency savings than current policy. It should be noted, that any changes in the Latvian energy
policy for energy-intensive manufacturing industries should be clarified with European Commission.
As the system dynamic model describes theoretical energy-intensive manufacturing company, further studies of
actual energy-intensive companies and its behaviour should be analysed to calibrate the model.

Acknowledgements

The work has been supported by the National Research program “Energy efficient and low-carbon solutions for a
secure, sustainable and climate variability reducing energy supply (LATENERGI)”.

References

[1] Grave K, Hazrat M, Boeve S, von Blucher F, Bourgault C, Breitschopf B, Friedrichsen N, Arens M, Aydemir A, Pudlik M, Duscha V,
Ordonez J, Lutz C, Grosmann, A, Fiaute M. Elelctricity costs of energy intensive industries. An International Comparison. Available:
http://www.ecofys.com/files/files/ecofys-fraunhoferisi-2015-electricity-costs-of-energy-intensive-industries.pdf
[2] Communication from the Commission — Guidelines on State aid for environmental protection and energy 2014-2020. Official Journal of the
European Union 2014;57:1–55.
[3] Locmelis K, Bariss U, Blumberga D. Latvian energy policy on energy intensive industries. . Energy Procedia 2017;113:362–368.
[4] Kursisa A, Gleizde L. Energy consumption and its reduction potential in Latvian industry sectors. Civil Engineering. International Scientific
Conference. Proceedings. 2013:363–371
[5] Dzene I, Polikarpova I, Zogla L, Rosa M. Application of ISO 50001 for implementation of sustainable energy action plans. Energy Procedia
2015;72:111–118.
[6] Thollander P, Ottosson M. An energy-efficient Swedish pulp and paper industry : exploring barriers to and driving forces for cost-effective
energy efficiency investments. Energy Efficiency 2008;1(1):21–34.
[7] Stenqvist C, Nilsson LJ. Process and impact evaluation of PFE – a Swedish tax rebate program for industrial energy efficiency. Paper presented
at 9th eceee summer study Act! Innovate! Deliver! Reducing energy demand sustainably, France, 2009.
[8] Stenqvist C, Nilsson LJ. Energy efficiency in energy-intensive industries – an evaluation of the Swedish voluntary agreement PFE. Energy
Efficiency 2012;5(2):225–241.
[9] Beihmanis K, Rosa M. Energy Management System Implementation in Latvian Municipalities: From Theory to Practice. Energy Procedia
2016;95:66–70.
[10] Blumberga D, Vigants H, Cilinskis E, Vitolins V, Borisova I, Khabdullin A, Khabdullin A, Khabdullina Z, Khabdullina G, Veidenbergs I.
Energy Efficiency and Energy Management Nexus. Energy Procedia 2016;95:71–75.
[11] Sterman J. Business dynamics: systems thinking and modelling for a complex world. McGraw-Hill; 2000.
[12] Senge, Peter M. The Fifth Discipline: The Art and Practice of the Learning Organization. New York: Doubleday; 1990.
[13] Ministry of Finance of Republic of Latvia, Main macroeconomic indicators. Available:
http://www.fm.gov.lv/en/s/macroeconomics/main_macroeconomic_indicators/