Improving Steam Turbine Performance with Industrial Internet

(1)

Topias Leppänen

IMPROVING STEAM TURBINE PERFORMANCE WITH INDUSTRIAL INTERNET

Master’s Thesis

Faculty of Engineering and Natural Sciences

Asko Ellman

Matti Vilkko

February 2021

(2)

ABSTRACT

Topias Leppänen: Improving Steam Turbine Performance with Industrial Internet Master’s Thesis

Tampere University

Master’s degree program in Mechanical Engineering February 2021

Most of the global electricity production is done in thermal power plants with steam turbine powered generators. The steam turbine is one of the most critical components in a thermal power plant, and non-optimal performance affects the overall profitability of the plant significantly. Indus- trial Internet utilizes collective measurement data to help in decision making and optimization of different business processes. It is a rapidly developing technology and solutions based on this technology are gaining a stronger foothold in different, varying markets.

The thesis aims to investigate how Industrial Internet could in the best way be utilized to improve the profit-making performance of steam turbines, with the solution being able to be produc- tized as a profitable product into the product lineup of a turbine automation company. The thesis follows a typical product concept development process. First, the need of the market is re- searched through literary study and with qualitative interviews done to both thermal power plants and turbine automation system sales representatives. Based on the market need, acknowledged capabilities of Industrial Internet, and the existing product offering of the steam turbine automation company, a product concept is developed and suggested through iterative action research.

Two main methods of increasing the profit-making performance of steam turbines are found during the market need study, optimizing the usage of the turbine such that the most profitable amounts of process steam, district heat, and electricity are continuously produced, and optimizing the maintenance of the turbine such that maintenance, failure, and condition degradation costs are minimized. Condition monitoring is required for any maintenance optimization to be done, and the process and quality of steam turbine condition monitoring can be improved with better utilization and analysis of data that is available.

A product concept automating and adding quality to steam turbine condition monitoring is developed. The concept aims to monitor the behavior of steam turbines by performing cloud based data analysis on both process and vibration data. By utilization of machine learning, digital twin models are formed which allow indication of anomalies and long-term changes in the behavior of chosen measurements or parameters. Tracking of proper process values also allows indication of conditions that are leading to condition degradation of the turbine. The concept works as a tool for a condition monitoring expert to diagnose emerging faults and current turbine condition, based on which they can provide the plant with recommendations of future actions. A product based on the proposed concept would fit well into the product lineup of the turbine automation company.

Small changes in the process conditions of a steam turbine can affect the monitored measurement values and parameters significantly. The training data used by machine learning to form the digital twin models should contain the typical process conditions that are faced for the anomaly detection to function properly, and the capability of the system to learn as it is used is essential for successful automation of the condition monitoring process. Further study on how successful a functional concept based on the proposition is in automating and improving the quality of steam turbine condition monitoring should be done to prove the added value coming from the concept.

This can be done with a live implementation of the concept that is running preferably at a combined heat and power plant and working as a part of the steam turbine condition monitoring process.

Keywords: steam turbine, Industrial Internet, Internet of Things, condition monitoring, artificial intelligence, digital twin

The originality of this thesis has been checked using the Turnitin OriginalityCheck service.

(3)

TIIVISTELMÄ

Topias Leppänen: Höyryturbiinin suorituskyvyn parantaminen teollisen internetin avulla Diplomityö

Tampereen yliopisto

Konetekniikan diplomi-insinöörin tutkinto-ohjelma Helmikuu 2021

Suurin osa maailman sähköntuotannosta tapahtuu lämpövoimaloissa höyryturbiinien pyörittämissä generaattoreissa. Höyryturbiini on yksi voimalaitoksen kriittisimpiä komponentteja, ja epäoptimaalinen suorituskyky vaikuttaa merkittävästi laitoksen kannattavuuteen. Teollinen internet hyödyntää kollektiivista mittausdataa päätöksenteon avustamisessa ja erilaisten liiketoimintaprosessien tehostamisessa. Teollinen internet on nopeasti kehittyvä teknologia ja siihen perustuvat ratkaisut ovat saamassa parempaa jalansijaa erilaisissa markkinoissa.

Työn tavoitteena on tutkia, kuinka teollista internetiä voitaisi parhaalla tavalla hyödyntää parantamaan höyryturbiinin tuottavuutta siten, että ratkaisun voisi tuotteistaa kannattavana tuotteena turbiiniautomaatio yhtiön tuotevalikoimaan. Työssä noudatetaan tyypillistä tuotekonseptin kehittämisprosessia. Ensiksi tutkitaan markkinatarve kirjallisuuskatsauksen ja laadullisten, sekä lämpövoimaloihin että turbiiniautomaatio myyjille, tehtyjen haastattelujen kautta. Markkinatarpeen, teollisen internetin kykyjen, sekä olemassa olevan turbiiniautomaatio yhtiön tuotevalikoiman perusteella kehitetään ja ehdotetaan tuotekonsepti käyttäen toimintatutkimusta.

Markkinatarve tutkimuksessa löydetään kaksi eri lähestymistapaa höyryturbiinin kannattavuuden parantamiseksi, optimoimalla turbiinin käyttö siten, että joka hetkellä tuotetaan kannattavin määrä prosessihöyryä, kaukolämpöä, sekä sähköä, ja optimoimalla kunnossapito siten, että huollon, vikaantumisten, ja hyötysuhteen huononemisen kokonaiskustannukset minimoidaan. Kunnonvalvontaa tarvitaan kaikkeen kunnossapidon optimointiin, ja höyryturbiinin kunnonvalvonnan prosessia ja laatua voidaan parantaa laajemman olemassa olevan datan hyödyntämisen ja analysoimisen avulla.

Työssä kehitetään höyryturbiinin kunnonvalvontaa automatisoiva sekä sen laatua parantava tuotekonsepti. Konseptin tavoitteena on seurata höyryturbiinin käyttäytymistä pilvipohjaisella prosessi-, ja värähtely datan analyysillä. Koneoppimisen avulla muodostetaan digitaalinen kaksonen, eli mallit, mitkä mahdollistavat poikkeamien ja pitkän aikavälin muutosten seuraamisen valituissa mittauksissa tai parametreissa. Seuraamalla oikeita prosessiarvoja voidaan myös osoittaa olosuhteet mitkä johtavat turbiinin kunnon heikkenemiseen. Konsepti toimii kunnonvalvonnan asiantuntijan työkaluna alkavien vikojen ja kunnon arvioimiseen, minkä perusteella hän voi antaa laitokselle suosituksia tulevaisuuden toimista. Esiteltyyn konseptiin perustuva tuote sopisi hyvin turbiiniautomaatio yhtiön tuotevalikoimaan.

Pienet muutokset höyryturbiinin prosessiolosuhteissa voivat vaikuttaa seurattaviin mittauksiin ja parametreihin huomattavasti. Koneoppimisen käyttämän opetusdatan digitaalisen kaksonen luomisessa tulisi sisältää kaikki tyypilliset prosessiolosuhteet, jotta poikkeavuuksien seuraaminen toimii, ja järjestelmän kyky oppia käytön aikana on olennaista onnistuneen kunnonvalvonnan automatisoimisen kannalta. Konseptin tuoman lisäarvon osoittamiseksi tulisi jatkotutkimusta tehdä siitä, että miten hyvin toiminnallinen konsepti onnistuu höyryturbiinin kunnonvalvonnan automatisoinnissa sekä sen laadun parantamisessa. Tutkimus voidaan tehdä toteuttamalla toiminnallinen konsepti lämpövoimalaan, mieluiten yhteistuottolaitokseen, joka toimii osana höyryturbiinin kunnonvalvonnan prosessia.

Avainsanat: höyryturbiini, teollinen internet, esineiden internet, kunnonvalvonta, tekoäly, digitaalinen kaksonen

Tämän julkaisun alkuperäisyys on tarkastettu Turnitin OriginalityCheck –ohjelmalla.

(4)

PREFACE

This thesis has been the last task in my studies, and it has prepared me for the new upcoming challenges expected. However, studies and work are never done alone, and I want to thank different people who have supported me during the endeavor of working on this thesis but also the entirety of my life and studies.

I want to thank Valmet Automation and all my colleagues for giving me the opportunity of working with them during the thesis work and the summer trainee periods preceding.

The turbine automation product line deserves a special thanks for the interesting topics of work given to me during my time in the team and the possibility to learn much new.

Etienne Guyon gave me the topic of research for this thesis and has guided me whenever I needed support and direction. He has taught me much about productization and making business. Tom Bäckman was of great help during the customer interviews and Juha Kautto has been of invaluable help during the concepting phase while teaching me the basics of turbine condition monitoring. Teijo Salonpää, Tuukka Harmaala, and Veli-Matti Uski earn a special thanks for helping me and supporting the research from the Industrial Internet perspective.

No matter how intriguing, work is what we do for a living, and we must always remember what is most important to us. I feel gratitude for my family that has given me an upbringing and taught me how to live in this world. My friends have been enriching my life throughout times and most recently my studies. The love and caretaking of my dear wife Laura has been endless and giving much joy to each day received from the Lord.

Tampere, 17 February 2021

Topias Leppänen

(5)

LIST OF FIGURES AND TABLES

Figure 1. The scope of the thesis in a typical product concept development

process (adapted from Keinonen & Takala 2006, p. 79). ... 4

Figure 2. An ideal and actual Rankine cycle on the steam h-s diagram (reproduced from Spliethoff 2010, p. 62). ... 6

Figure 3. The process devices essential to plants operating on the Rankine cycle (reproduced from Spliethoff 2010, p. 62). ... 7

Figure 4. A Rankine plant with preheating, reheating, and district heat production. ... 9

Figure 5. Flow through steam turbine stages (Kadambi & Prasad 2015, p. 2). .... 12

Figure 6. The structure of a steam turbine (Leyzerovich 2008, p. 45). ... 13

Table 1. Typical steam turbine failure modes. ... 18

Figure 7. The structure of an Industrial Internet platform (adapted from Gilchrist 2016, p. 77). ... 21

Figure 8. Product offering in the steam turbine performance improvement market. ... 27

Figure 9. How automation aids in improving turbine profitability. ... 40

Figure 10. How automation aids in reaching optimal workflow. ... 40

Table 2. Design drivers leading application specification. ... 56

Figure 11. Data flow from measurements to user. ... 63

Table 3. Assumptions about production value, reliability, and maintenance costs and duration for added value calculations. ... 64

Table 4. Condition monitoring added value for power plant customer steam turbines with average electricity production capacities of 50MW and 100MW. ... 65

(7)

LIST OF SYMBOLS AND ABBREVIATIONS

ABB Asea Brown Boveri

AI Artificial intelligence

CHP Combined heat and power

DCS Distributed control system

DMM DNA Machine Monitoring

GE General Electric

HMI Human-machine-interface

HP High pressure

Hz Hertz

IIoT Industrial Internet of Things

IoT Internet of Things

IP Intermediate pressure

KPI Key performance indicator

LP Low pressure

OEM Original equipment manufacturer

rpm Rounds per minute

UI User interface

VII Valmet Industrial Internet

(8)

1. INTRODUCTION

Steam turbines are one of the key equipment of thermal power plants, and most of the world’s electricity is produced by steam turbine powered generators (Korpela 2011, p.

137). Industrial Internet, typically utilizing measurement data to aid in decision making and provide added value in different business processes, has been in rapid growth in the past years and products based on the technology have found their way into the market providing financial benefits for both customers and suppliers (Ramgir 2020, p. 5). The thesis aims to combine Industrial Internet and the world of steam turbines by seeking for and proposing a product concept improving the profit-making performance of steam turbines with the aid of Industrial Internet.

Even small improvements in steam turbine performance can raise power plant total efficiency significantly resulting in more profitable power generation business, lower electricity prices, and reduced emissions resulting from thermal plants (Speight 2013, p. 469).

If all thermal plants in the world could gain this small performance improvement, the global emission reduction would be notable.

Industrial Internet is becoming more mature as a technology and solutions around this technology are gaining a better and more stable foothold in different global markets. The true benefits and disadvantages of this technology have become more apparent and the solutions are truly helping increase customer profitability while also being profitable business for the solution provider. (Ramgir 2020, p. 5)

Industrial Internet at thermal power plants including steam turbines is not a new idea.

Solutions based on the technology are being used at power plants to monitor power plant equipment condition and to warn about upcoming failure. The benefits being gained are reduced plant downtime, reduced operation costs, and even reduced insurance costs.

(Ramamurthy & Jain 2017; Murty et al. 2018) Solutions improving plant flexibility, which allow the plant to reach more demanding production quantities while reducing the emissions of the plant, have also been implemented (Ramamurthy & Jain 2017). Solutions allowing for production optimization based on the current product value and market situ- ation have been successfully implemented as well (Ramamurthy & Jain 2017; Qu et al.

2018). Also, solutions finding the optimal process parameters of the power plant have

(9)

been successfully implemented to improve the total efficiency of a thermal power plant running a steam turbine (Xu et al. 2015; Kolesnikova et al. 2020).

1.1 Research questions and research methods

The thesis aims to answer three main questions. The questions are based both on the requirements of successful product concept development projects (Kahn 2013, p. 214) and the requirements Valmet Automation sets on the thesis. The questions are:

1. By what means can the profit-making performance of steam turbines running in thermal power plants be improved?

2. How can the customer needs fulfilment of Valmet steam turbine automation customers be improved?

3. How can these needs and means be answered and improved with a product concept that fits into the Valmet product lineup while being a profitable product for Valmet that utilizes Valmet Industrial Internet capabilities?

The first two questions are tightly bound together and are set to ensure that the product concept development stays on the right track. Customer needs regarding steam turbine profit-making performance improvement and places where they could be improved are sought after. The answers to the first two questions should be known before going on to answer the third question.

The third question is answered by proposing a product concept which answers one or more of the customer needs identified by answering the first two questions, and thus adds value to the customer business while also being profitable business for Valmet.

This concept development is done in cooperation with other Valmet employees but is coordinated with this master’s thesis.

Quantitative research is based on statistical analysis of large amounts of data. Generally, a hypothesis is required which can be tested and around which the study can be done.

(Silver et al. 2013, p. 58) Proper sampling is essential for the success of quantitative research. If the sampling is done correctly, quantitative research ensures that the results truly apply to the entirety of the targeted customers. (Silver et al. 2013, p. 20) Quantitative research results are generally held to be quite reliable and often preferred over qualitative research which does not have large amounts of data to back up the conclusions (Silver et al. 2013, p. 57).

Qualitative research is not based on the quantity, but the quality and depth of data. It often consists of in-depth research done to only a selected number of customers. This

(10)

method of research often requires immersing into the customers daily operation in some way and observing the habits and decisions made by the customer. (Silver et al. 2013, p. 56) In this way, it is possible to notice matters which the customer themselves may not have noticed due to them being routine or an accepted way of working. Also, this provides insight into the mindset of the customer and how they think and make decisions.

However, qualitative research is often criticized for providing biased data as the studiers themselves can guide the study into the direction they prefer (Silver et al. 2013, p. 60).

Action research is often described as a link in between theory and practice. The research method aims to find out what affect an action has. The basic model of action research includes planning an action, acting, observing the affect, and reflecting on what has hap- pened to learn for the future. (Costello 2003, p. 7) Action research should be done in parallel to some task that is being done to see what the task truly affects. In a product development sense of view, it should be done in parallel with the product or concept development to gain insight about how the product or concept truly behaves in operation or how the customer reacts to the product. Iteration is often used in development processes (Kahn 2013, p. 14) and action research provides a good way of evaluating what to change in the next iteration.

The thesis will use all the different research methods described. In seeking answers to the first two questions, qualitative methods will be used due to the difficulty of using quantitative methods for such exploratory research (Silver et al. 2013, p. 58). Both questions are quite exploratory in nature, and clear hypotheses may not initially be formed.

Literary background research is first done to gain some understanding of the field of steam turbine automation and insight on how steam turbine performance may be improved. Interaction with both Valmet sales who work up front with customers and Valmet turbine automation customers through qualitative methods is used to find answers to the first two questions. Valmet sales are interviewed for customer needs and the fulfilment of Valmet offering of those needs. Customers are interviewed in a conversational and explorative nature to gain understanding of the customer mindset and what needs they have and what challenges they face.

Answering the third question involves action research. Based on the answers to the first two questions, a solution that is beneficial for both Valmet and the customer is developed. The target is to find and propose a product concept utilizing Industrial Internet capabilities that answers the places of improvement found by answering the first two questions. Action research is used during the development process to analyze the com- patibility of the solution to answering the desired questions. In addition, while testing the feasibility of some of the features of the Industrial Internet product concept, quantitative

(11)

methods are used as the performance is tested against large quantities of running data received from customer power plants.

1.2 Structure of thesis

The thesis flows like a typical product concept development process depicted in Figure 1. First, customer needs are determined in a customer needs study, based on which a general concept idea is developed. From the general concept idea, more specific details and requirements are determined based on which a functional product can be developed.

(Keinonen & Takala 2006, p. 79-80) The thesis aims to propose a concept with enough details to move the concept forwards towards a functional product but does not include full pilot testing of the concept.

Figure 1. The scope of the thesis in a typical product concept development process (adapted from Keinonen & Takala 2006, p. 79).

The thesis starts with introducing everything that should be known before going on into the customer needs research and concept development endeavor. The aim is to understand the product domain by introducing steam turbines and the power generation industry working with steam turbines. Basic theory behind steam turbines is explained, as well as the steam process surrounding steam turbines. Next, features of Industrial Inter- net are explained to give understanding of the features, strengths, and weaknesses of the technology to be utilized in the concept. At last, the markets steam turbines lie in and the markets that have built up around steam turbines are introduced to aid in business potential evaluation of the concept.

(12)

After general introduction of the product domain, design, implementation, and results of the customer needs study are explained, in which the customer is understood to be a power plant that has steam turbines running in their process. In this section, the first two research questions are answered, and the customer needs and challenges found that work as a basis of concept development are explained.

Based on the specific customer challenges found during the customer needs study, a product concept utilizing Industrial Internet capabilities to improve steam turbine condition monitoring is developed. First, the main idea of the developed concept is introduced, after which more detailed specifications are explained. Customer added value calculations are included and a business model concept is also proposed.

After concept proposal, the reliability of the customer needs study is evaluated. Following this, the suitability of the concept into the Valmet product lineup, the appeal of the concept to customers, and the feasibility of Industrial Internet technology for implementing the concept are evaluated. Last, the recommendations for next steps forward are proposed.

(13)

2. PRODUCT DOMAIN

Understanding the domain where the product concept to be developed is to lie in is essential for successful product concept development. In addition to understanding the customer and the technology the product is to be based on, it must also be understood what the competing global market has to offer, and the effect the launch of a new product such as the concept has on the lineup of other products the company offers. (Keinonen

& Takala 2006, p. 60; Kahn 2013, p. 214)

2.1 Customer technology

The thesis aims to propose a product concept improving steam turbine performance. The target customer of the product concept is any thermal power plant that is operating a steam turbine. Wherever there are steam turbines, there is also a steam cycle in operation. Understanding of the steam cycle and power plants in general is necessary in addition to understanding the behavior and physics behind steam turbines. Some theory behind steam power plants, which are thermal power plants operating on the steam cycle, and steam turbines is introduced which helps in understanding both the results of the customer needs study and features of the proposed concept.

2.1.1 Rankine cycle

Steam power plants operate on the Rankine thermodynamic cycle which operates with a medium of water and water vapor. Through heat exchange and phase changes of the water, mechanical energy can be extracted from a heat source. (De Souza 2012, p. 38)

Figure 2. An ideal and actual Rankine cycle on the steam h-s diagram (reproduced from Spliethoff 2010, p. 62).

The Rankine cycle depicted in the steam enthalpy-entropy diagram in Figure 2 is the simplest configuration of the Rankine cycle. Saturated water is pumped into the steam

(14)

generator pressure which is seen in the enthalpy-entropy diagram as the shift from point 1 to 2. The steam generator heats, vaporizes, and superheats the water into superheated steam at a constant pressure which is seen as the shift from point 2 to 3. The hot, high pressure steam is expanded in the turbine into lower pressure, often slightly moist steam, depicted in the shift from point 3 to 4. In the condenser, the low pressure steam is condensed into saturated water at a constant pressure, seen as the shift from point 4 to 1.

(Spliethoff 2010, p. 62; De Souza 2012, p. 34; Dincer & Zamfirescu 2014) The major differences between the ideal and actual Rankine cycle are seen mostly in the turbine and pump. In the ideal cycle, the rise in enthalpy in the pump and drop in the turbine are isentropic. A true cycle does not have isentropic behavior, but the enthalpy increases decreasing the power available from the turbine and increasing the power consumed by the pump. This behavior is described with the isentropic efficiency. (Sarkar 2015, p. 19) In an actual cycle the pressure also decreases with flow, so the steam generator and condenser are not truly at a constant pressure (Spliethoff 2010, p. 63; Sarkar 2015, p.

20).

As described earlier, a plant operating on an ideal simple Rankine cycle only needs a steam generator, steam turbine, condenser, and pump. (Spliethoff 2010, p. 62; Sarkar 2015, p. 15) Figure 3 shows a process diagram of a simple Rankine cycle depicting the major components necessary for operating on the cycle.

Figure 3. The process devices essential to plants operating on the Rankine cycle (reproduced from Spliethoff 2010, p. 62).

The high temperature steam is produced in the steam generator which is in most cases a boiler that combusts varying fuels and transfers the thermal energy freed in combustion to the feedwater to raise its temperature, vaporize it, and finally superheat it. The feedwater and steam that is produced flows through various heat exchangers in the boiler

(15)

which transfer the heat in the combustion gases into the water through radiation and convection. (Spliethoff 2010, p. 81; Speight 2013, p. 455)

The turbine is used to extract mechanical energy from the boiler output steam. In the turbine the pressure and temperature of the steam decrease as energy is converted into torque of the turbine axle. The mechanical energy from the turbine is often used to run a generator to produce electricity. (Spliethoff 2010, p. 76; Speight 2013, p. 458)

The turbine exhaust steam is condensed to fluid before pumping it back to feedwater pressure. This is done in the condenser which has a flow of cooling water that extracts the necessary energy from the steam to condense it. Depending on the site, the condenser cooling water may be taken from lakes or oceans, and the heat energy dumped into the environment, or the heat energy taken during condensing may be used as district heat. (De Souza 2012, p. 51; Speight 2013, p. 459)

The overall efficiency of plants operating in the Rankine cycle has been increasing since the cycle was first implemented (Spliethoff 2010, p. 79). Increasing the efficiency means that the products are made using less fuel, which means greater profits during operation in addition to reduced emissions. This efficiency improvement is mainly driven by eco- nomic reasons even if emission restrictions also demand efficiency improvement.

(Speight 2013, p. 470) Currently the best thermal efficiency, the efficiency of a Rankine plant producing only electricity, is around 47% (Speight 2013, p. 466; Breeze 2018, p.

28).

The best method to improve the overall efficiency of the Rankine cycle is to utilize the condenser heat as district heat. Implementing this improvement alone can raise the overall efficiency to around 85%. (Breeze 2018, p. 46) If heat is not needed and electricity is the only necessary product, adding district heat to the products is not viable. This means that the electric generation efficiency must be maximized. Improving the simple cycle electric generation efficiency is done by increasing the steam pressure and temperature, decreasing the condenser pressure, and improving the isentropic efficiency of the turbine (Leyzerovich 2008, p. 390; Oakey 2011, p. 454; Dincer & Zamfirescu 2014). Some plants are operated supercritical, where the steam pressure is so high that there is not a distinct vaporization phase. This increases cycle efficiency but demands more advanced equipment and limits the operation modes of the plant. (Breeze 2018, p. 43)

(16)

Figure 4. A Rankine plant with preheating, reheating, and district heat production.

The thermal efficiency of a simple Rankine cycle cannot reach 47%, but the cycle must be modified to reach high thermal efficiencies. Plants may often add several turbines and reheat the steam after the first turbine to increase the power output and thermal efficiency of the cycle (Sarkar 2015, p. 21). Regeneration, heating the feedwater before feeding it to the boiler, also increases the efficiency of the cycle. This is done by extracting some steam in between turbine stages and feeding it to the feedwater after heat exchangers (Speight 2013, p. 454; Sarkar 2015, p. 23). A Rankine cycle with district heat extraction, reheating and regeneration implemented is depicted in Figure 4. The feedwater tank, not depicted in Figure 3, is needed in a true cycle for removing air from the feedwater before it goes to the pump and boiler.

2.1.2 Steam power plants

Most of the total global power production is done in thermal power plants operating on the Rankine cycle, more specifically steam power plants, even if renewable sources such as solar, wind, geothermal, and hydro power are increasing in market share (Dincer &

Zamfirescu 2014). In 2008, 57% of the global electricity generation was done in thermal power plants (Korpela 2011, p. 137).

Steam power plants typically rely either on combustion or nuclear energy. The operation of such plants is not heavily dependent on the weather. Most renewables such as wind, solar, and hydro power must have proper weather and climate conditions to give electricity and are such quite unpredictable sources of energy (Qu et al. 2018). Steam power plants that give heat and electricity regardless of weather conditions are necessary to complement these renewable sources of power. They enable the power given to the grid

(17)

to be equal to its consumption which is essential for an electric grid to function even if weather conditions change (Qu et al. 2018).

Even if steam power plants are traditionally operated using coal and fossil fuels, biomass and other renewable combustibles are becoming increasingly popular as fuels due to limitations and restrictions both on combustion emissions and usage of fossil fuels (Wu et al. 2015). In fact, any heat source that is hot enough may be used for steam production. This means that even as fossil fuel usage drops globally, steam power plants will stay in operation.

Power generation with main products of electricity and heat is the most typical use of steam power plants (Wu et al. 2015). Plants producing both electricity and heat are called combined heat and power (CHP) plants and are quite typically used if there is any need for heat due to their highly increased efficiency over simply electricity production. De- pending on the site location and market for electricity and heat, which is often sold as district heat, the plants may be producing either heat or electricity as their main product.

The ratio of electricity and heat generated is known as the power-to-heat ratio (Gvozdenac et al. 2017).

CHP plants often operate on the “topping cycle” where the energy extraction required for condensing the turbine discharge steam is used as heat, for example for district heating (Speight 2013, p. 470; Breeze 2018, p. 25). Here the electricity production is the priority and the losses of the plant are minimized by selling the energy extracted during condensation as heat. This is a viable operation mode if electricity price is high.

Sometimes heat is the main product of the plant, and the plant is designed to provide heat with the side product of electricity with a turbine to minimize losses during heat production. Even in plants with the main product of electricity, the electricity demand and price might occasionally be low, but heat is needed, and plants designed to produce electricity might find it profitable to maximize their heat production. In these cases, the plant works on the “bottoming cycle” where the electricity production follows the heat production. (Speight 2013, p. 470; Breeze 2018, p. 25) During operation the required amount of district heat is produced, and the amount of electricity produced follows. In some cases, if electricity price is very low, it may even be financially viable to bypass the turbine completely to produce more heat or reduce the plant fuel consumption. This may be done by allowing the steam through a pressure-reducing and de-superheating valve, which simply lowers the pressure and temperature of the steam to the desired pressure and temperature while keeping the thermal energy of the steam as high as possible (Sarkar 2015, p. 336).

(18)

Steam power plants are also used in many process facilities and factories to provide processes with hot steam at the desired pressure. Process steam which is extracted from the process is the main product of these plants, but as side products, electricity and sometimes heat, are generated. (Pradeep Varma & Srinivas 2017; Breeze 2018, p. 26) This improves the overall efficiency and profitability of the plant. These plants operate with a slightly different mission, which is to provide a proper amount of process steam regardless of the steam demand. Electricity and heat are produced from the steam that is not needed in the processes. (Breeze 2018, p. 26)

The total energy conversion efficiency of a steam power plant is a good measure of plant performance. This is calculated with the ratio of energy produced and energy consumed.

(Breeze 2018, p. 1) The generated electric power, district heat power, and process heat power can all be determined from measurements done in the plant. In a combustion plant, the energy consumption may be determined from the amount of fuel combusted and its energy content, often determined with the lower heating value (LHV) (Speight 2013, p. 468).

The efficiency of a steam power plant is described with many different measures. The cycle efficiency is calculated based purely on the thermodynamic cycle and combustion energy conversion is not considered. It is calculated as the ratio of turbine work produced and the heat transferred to steam. (De Souza 2012, p. 4) In an electric generation power plant, the efficiency is often described with the fuel-to-electrical efficiency, heat consumption rate, or the heat-rate, which is the ratio of electric energy produced and the amount of fuel energy required (Speight 2013, p. 468; Breeze 2018, p. 1; Xu et al. 2019). This measure is essentially the inverse of the cycle efficiency but also considering the turbine mechanical and generator efficiency. In CHP plants the specific steam consumption, which is the ratio of the steam flow and the generator power output, is commonly used (Sarkar 2015, p. 4). The specific steam consumption does not consider heat production however, so it is not good in describing the overall efficiency of the CHP plant.

The availability of a plant is another good measure of plant performance. It measures the reliability and usability of the plant with the ratio of uptime to downtime. The availability of plants has been improving with preventive and condition-based maintenance.

(Speight 2013, p. 469) The availability of a plant has a direct effect on the profits made by the plant, since downtime does not produce any income, but upkeep costs are still running in addition to the repairing costs. Small improvements in availability can raise profits significantly.

(19)

The capacity factor is used to describe how well the plant potential is utilized. It is the ratio of the true production capacity to the maximum production capacity. (Speight 2013, p. 469) The greater the capacity factor is, the more revenue the plant makes during operation as it produces more products. Often improving the capacity factor also improves the overall efficiency also, improving profits.

2.1.3 Steam turbines

Turbomachines are high speed rotational machines operating at high temperatures that are used to convert heat energy from a flowing fluid into shaft torque (Kadambi & Prasad 2015, p. 1). Steam turbines are turbomachines that extract heat energy from a flow of steam (Dincer & Zamfirescu 2014).

A steam turbine consists of multiple sequential stages where the flow is first either di- rected or accelerated by stationary, casing attached, stator blades and then the energy is transferred to the shaft through rotating rotor blades that the steam flow effects on.

Figure 5 shows how the stators and rotors are arranged and how the steam flows through the stages. (Kadambi & Prasad 2015, p. 2)

Figure 5. Flow through steam turbine stages (Kadambi & Prasad 2015, p. 2).

Two different principles, impulse and reaction, are used to convert the energy available in the steam into shaft torque (Speight 2013, p. 456). In the impulse principle, the steam velocity is increased in the stator and it hits the rotor blades with high velocity objecting a force on them (Korpela 2011, p. 138). The reaction principle is based on the pressure difference between the two sides of the rotor blade which causes a force on the blade.

Typically steam turbines work on a combination of both principles, with higher pressure stages relying more on the impulse principle and lower pressure stages more on the reaction principle. (Korpela 2011, p. 166; Sarkar 2015, p. 200)

(20)

Figure 6. The structure of a steam turbine (Leyzerovich 2008, p. 45).

Figure 6 shows the structure of a steam turbine. The main parts of the steam turbine are the casing, blading, shaft, bearings, and seals. The casing holds the stators and guides the steam flow. The shaft containing the rotor blades, often also called the turning gear, lies in the casing supported by the bearings. (Sarkar 2015, p. 205) There are two different types of bearings, journal bearings and thrust bearings. The journal bearings support the shaft in the radial direction, and the thrust bearing supports the shaft in the axial direction.

Both bearings require a constant flow of lubrication oil to work. (Perez & Lawhon 2016, p. 31) The seals are required to separate the oil flow bearings flow from the steam passages (Perez & Lawhon 2016, p. 40). Also needed for operating a steam turbine is a control valve before the turbine to control the steam flow through the turbines, and some device that can slowly rotate the turbine that is used both when speeding up the turbine and keeping the shaft rotating during cooling down of the turbine (Sarkar 2015, p. 212).

A steam power plant usually has multiple turbines following each other to increase the efficiency of the turbines and balance out axial forces forming due to the flow of steam, as is seen in Figure 6 (Leyzerovich 2008, p. 52). Often on the same shaft there is a high pressure (HP), intermediate pressure (IP), and low pressure turbine (LP). The HP turbine is seen in the left in Figure 6. After the HP turbine, the steam often goes back to the boiler to be reheated, and then goes on to the IP turbine, which is seen to the right of the HP turbine section, with slightly bigger rotors. From here, the steam flow in this case is split and goes to the LP turbines seen in the right. Here the steam flows in opposite directions to minimize axial loads, resulting from the large area of the LP blades, on the thrust bearing.

(21)

In between the different turbine sections or even in between stages there might be extractions from where steam is taken to different applications. The pressure of the extracted steam depends on the position of the extraction, and the closer to the end of the turbines the extraction is, the lower the extraction pressure. The extraction flow, which affects the pressure, may be controlled by valves at the extractions. (Korpela 2011, p.

135)

Steam turbines are typically categorized with three different categories, condensing turbines, backpressure turbines, and extraction turbines. Condensing turbines aim to extract the maximum possible amount of mechanical energy from the steam flow and the turbine discharge pressure is thus as low as possible. By taking the steam to a moist state, erosion of the last stage blades increases due to the increase of high velocity water droplets in the flow, and this erosion limits lowering the discharge pressure (Sarkar 2015, p. 21). Backpressure turbine discharge steam is often used for district heating or in processes, and it contains more energy than the discharge of a condensing turbine. A backpressure turbine does not extract as much energy from the fluid flow as a condensing turbine does, but the discharge steam energy is utilized at some other process before condensation. Extraction turbines, which are the most common in power generation, may have a discharge pressure as low as a condensing turbine, but when desired, higher enthalpy steam may be taken from the extraction. (Korpela 2011, p. 135)

Two main types of turbine main control valves exist. The simpler, more traditional valve type is a typical throttling valve. With valve adjustment, the valve flow coefficient is reduced and the pressure after the valve drops with the enthalpy of the steam staying constant. This drop in the pressure difference across the turbine reduces flow. (Kadambi

& Prasad 2015, p. 145; Sarkar 2015, p. 212) Nozzle control valves are also used. The valve consists of multiple nozzles which are opened one by one to enable flow to only some part of the turbine. Nozzle control valves reduce throttling losses formed by the control valve. (Kadambi & Prasad 2015, p. 148; Sarkar 2015, p. 214)

The main task of the turbine is to expand the high pressure steam and extract energy from the steam flow. The turbine regulates the production of a steam power plant by varying the amount of mechanical energy extracted. A steam turbine in power generation drives a generator producing electricity at 50 Hertz (Hz) or 60 Hz depending on the grid frequency. For a 2-pole generator to give these frequencies, it must run at 3000 rounds per minute (rpm) or 3600 rpm, which are the most typical steam turbine operating speeds. (Speight 2013, p. 458) However, gearboxes between the turbine and generator are quite commonly used which allow differing rotational speeds. Due to constant grid frequency and thus constant generator speed, the rotational speed of power generation

(22)

steam turbines stays nearly constant with only small alterations according to grid frequency. (Sarkar 2015, p. 205)

High temperatures form the biggest challenge in steam turbine mechanical development.

The allowed turbine temperature is restricted by materials and is one factor that adds difficulty into increasing boiler discharge steam temperature and pressure which increases the Rankine cycle thermal efficiency. (Sarkar 2015, p. 235) Temperature changes are also the greatest source of stress on the turbine components. Both the casing and shaft expand and distort with rising temperatures which affects changes in tolerances. Thus, raising the temperatures of the turbine parts by altering the steam flow must be done carefully ensuring equal temperature distribution and component distor- tion. The mechanical structures of the casing and shaft determine how fast the load of the turbine may be altered. (Oakey 2011, p. 65; Sarkar 2015, p. 207) The evolution of the global power market towards renewable resources adds demand for steam turbines to regulate the production of electricity into the grid and the grid frequency which must be done by changing the turbine load, and thermal stresses complicate this fast load changing.

The unideal isentropic efficiency of the turbine results from thermodynamics losses of the steam turbine. Entropy of the flow is increased due to fluid friction, aerodynamic losses, and entrance and exit losses between stages. (De Souza 2012, p. 34) Other losses in the turbine are heat losses from the casing, mechanical losses due to bearing friction, and loss of steam potential energy due to internal leaking of steam seals. The losses due to leakage are heavily dependent on the pressures of the steam paths, but if the turbine is well axially balanced, frictional losses are only dependent on the shaft rotational speed. Depending on the size of the turbine, mechanical losses may range from 0.5% to 10%. (Kadambi & Prasad 2015, p. 14)

2.1.4 Turbine control methods

The steam flowing through the turbine, which in turn affects the turbine shaft power, is regulated by using the main control valve. The flow regulation is done either by throttle control, where the pressure of the steam entering the turbine is controlled, or by nozzle control, where steam passages are opened and closed and thus the flow is directly controlled. Nozzle control provides smaller throttling losses during partial load operation.

(Sarkar 2015, p. 213) In both control modes, some reserve usually must be left in the control valves for adjusting to changing conditions and possible supporting the grid frequency (Sarkar 2015, p. 232).

(23)

Steam turbines can be controlled by both purely mechanical and electrical systems, the former being obsolete in modern power generation systems. In both control systems the valve positioning is typically done with the aid of hydraulics which is regulated by actuators controlled directly by the control system. (Sarkar 2015, p. 220) Electronic control systems rely on measurements done to the turbine, and the controller, typically a PID controller, adjusting the control valve position depending on the measured values. Some electronic control systems are distributed control systems (DCS) where all the measurements are read to the process control servers where the control logic is implemented.

The process control servers then control the actuators which regulate the turbine hydraulics. (Wu et al. 2015)

The traditional method of governing or controlling a steam turbine is speed control, where the governor or control system seeks to keep the speed of the turbine constant. During shaft load increase the speed begins to drop and flow must be increased, and during shaft load decrease the flow must be decreased to prevent the speed from rising. (Perez

& Lawhon 2016, p. 60) Turbine overspeed, a speed above the maximum allowed speed, is a dangerous condition for both the machine and personnel at the plant, as high centrifugal forces may result in the machine dismantling and machine pieces being thrown around. Turbine controls require fast speeds and response times as dangerous overspeed conditions happen very quickly if for example the generator is cut from the grid at full load and turbine flow. Overspeed in normal conditions is prevented by the speed control, but for redundancy and safety a turbine must have an overspeed protection system which is one of the most critical parts of the turbine control system. In case of by- passing the maximum allowed speed, the overspeed protection system closes the turbine control valves which stops turbine acceleration. (Perez & Lawhon 2016, p. 77) A modern steam turbine is run to grid speeds with specified speed ramps using speed control. When a power generation turbine is in the grid, the speed of the generator is bound to the grid frequency. (Sarkar 2015, p. 205) Thus, speed control is not necessary unless frequency support is being done, and other control methods may be used.

In steam power plant configurations, it is sometimes desired to have the turbine controller regulate the pressure of the boiler discharge steam by using the turbine control valve (Sarkar 2015, p. 220). In this control scheme, the power output of the generator is altered by changing the amount of fuel combusted (Wu et al. 2015). The fuel amount is increased, and the steam pressure begins to rise which results in the turbine controller opening the turbine control valve. The steam flow through the turbine increases, which affects an increase in generator power.

(24)

Sometimes it is desired to have the turbine control system control the generator load directly. When more electricity is desired, the turbine controller opens the turbine control valve and the steam flow increases increasing generator load. The increase of flow results in pressure reduction of the boiler discharge steam, and the boiler compensates to this by increasing the fuel power. This control scheme results in faster generator load alteration, but the steam pressure is not as stable as having the turbine control the steam pressure. (Wu et al. 2015)

In a backpressure turbine the turbine control system often controls the turbine backpressure. If low pressure steam extraction increases, the backpressure begins to drop. The turbine controller opens the turbine control valve allowing more steam to pass the turbine and raise the backpressure. The generator power varies as the controller keeps the backpressure stable.

In extraction turbines, the extraction pressure may be controlled with a separate controller that adjusts the extraction control valves. This control often works parallel with the turbine main control method. Opening the extraction valve increases the flow that is extracted resulting in a pressure drop in the turbine casing and a pressure rise in a possible extraction manifold or medium pressure steam reservoir. The main control mode of the turbine reacts to the flow leaving the extraction by opening the turbine control valve.

When steam turbines are producing electricity to the power grid, the control system must be able to support the grid frequency in some way. There are three reserve types in the control system that help in supporting grid frequency. Primary control is the fastest, and it reacts immediately by altering the generator power given to the network when it detects change in frequency. The primary control reserve is limited, and it must be restored at some point however so it can react to new fluctuations in the electric grid consumption.

Secondary control takes care of this, freeing up the primary control reserve by altering the power given to the grid in a longer timeframe of seconds to 15 minutes. Secondary control must also be restored so it can be used in the future, and this is the task of tertiary control, which is often done by rescheduling power generation to the grid and is not nec- essarily a task of the turbine control system. (Spliethoff 2010, p. 96-97)

2.1.5 Steam turbine failure modes and condition monitoring

The main components of a steam turbine are the casing, blades, shaft, bearings, and seals. Failure in any of these components typically results in downtime of the turbine and need for maintenance. However, failures in auxiliary devices and systems such as the control system, control valves, and lubrication oil system including its pumps can cause unavailability or performance degradation of the turbine also.

(25)

Table 1 includes the most typical failure modes of steam turbines and their auxiliaries that either cause downtime or performance degradation of the turbine. The table is a collective table gathered from multiple different sources (Beebe 2003; Leyzerovich 2008, p. 458; De Souza 2012, p. 229).

Table 1. Typical steam turbine failure modes.

As is seen from Table 1, many of the failure modes can be seen in vibrations which are typically monitored in condition monitoring solutions. Vibration-based condition monitoring is good for detecting faults present in the turning gear which cause unbalance on the shaft or abnormal shaft behavior at the bearings.

Component Damage mode Symptoms Root causes Control

valves Valve stem bend-

ing Leakage, jamming Fluid forces, thermal stresses Wear Leakage, jamming Steam impurity, normal wear Fouling Flow restriction Steam impurity

Nozzles Erosion Efficiency drop,

leakage Steam impurity, corrosion Fouling Efficiency drop,

flow restriction

Steam impurity Blading Erosion Efficiency drop,

change in flow characteristics

Steam impurity, corrosion, high water content in steam

Fouling Efficiency drop, flow restriction

Steam impurity High cycle fatigue Warping, cracking Thermal stresses Centrifugal force

fatigue Rupture, disman-

tling Overspeed

Rotor Warping Vibrations Thermal stresses

High cycle fatigue Vibrations Excessive vibrations, thermal stresses

Bearing surface

wear Vibrations Improper oil supply, excessive thrust force, excessive vibrations Bearings Wear, rubbing

and melting

Vibrations, high bearing temperatures

Improper oil supply, excessive thrust force, excessive vibrations Casing Deformation Leakage, vibrations Thermal stresses, high tempera-

tures

Rupture Leakage Thermal stresses, high water content in steam, corrosion

Seals Wear Leakage, total effi-

ciency drop, change in flow characteristics

Rubbing of shaft due to thermal stress deformation, particle erosion

Lube oil

pumps Wear Change in perfor-

mance Normal wear, lube oil impurities Piping Fouling Flow restriction Steam impurity, corrosion

Wear Leakage Flow erosion

Generator Insulation faults Efficiency drop Component aging Mechanical rotor

and bearing faults

Vibrations Improper bearing lubrication, for- eign particle damage

(26)

Vibration-based condition monitoring for turbines has its own features that must be understood. The turbine shaft is essentially rotating withing the journal bearing which is simply a round metal support filmed with oil for the shaft. The bearing has some play and the shaft can move around in the bearing. (Perez & Lawhon 2016, p. 32) Thus, measur- ing casing vibrations is not the best approach as the rotor should not be directly in touch with the casing. Rotor displacement measurements are typically used which are formed of two perpendicular displacement sensors installed at each bearing. From the rotor position within the bearing, rotational patterns and vibration amplitudes can be read, and a good insight of rotor behavior is gained. (Wilcox 2016)

Based on the rotor position measurement, the accurate rotor movement inside the bearing can be followed. The shaft centerline tells the position of the shaft inside the bearing, and the orbit tells how the shaft is vibrating inside the bearing. Following certain frequency vibrations allows to monitor the emergence of different faults, as different fault result in different frequency vibrations. Vibration-based condition monitoring depends on following and analyzing the shaft movement inside the bearing with different methods.

(Wilcox 2016)

Analyzing vibration behavior is quite difficult as the process surroundings affect the vibrations significantly, and root cause evaluation for vibrations must always consider the surrounding process. However, the development of startup and shutdown vibrations is typically followed as the conditions in these situations are relatively stable. The vibrations during startup and shutdown are typically visualized as a bode plot showing the shaft vibration amplitude and phase change at each rotational speed. (Wilcox 2016)

From Table 1 it is also seen that many faults and degradation are not seen in vibrational performance. The attempt to detect the emergence of these faults is performance analysis. Performance analysis consists of analyzing the thermodynamic properties of the turbine, and it is typically done with periodic “valve wide open” tests, or with high accuracy periodic measurements attached. The efficiency development of different turbine sections is typically followed which can help notice fouling, material buildup in the different components, or increase of leakage passing the blading. (Beebe 2003)

2.2 Product technology

The scope of the thesis is to propose a product concept improving steam turbine performance that is utilizing Valmet Industrial Internet capabilities. Thus, it is necessary to understand generic features of the Industrial Internet of Things but also specific features and capabilities of Valmet Industrial Internet.

(27)

2.2.1 Industrial Internet of Things

The concept of Internet of Things (IoT) is relatively new and business developing around this concept is rapidly evolving and developing. Industrial companies have developed their own concept of Industrial Internet of Things (IIoT), or simply Industrial Internet, around IoT that aims to serve industrial purposes and aid in daily business. (Gilchrist 2016, p. 3) Companies developing Industrial Internet solutions seek to improve the efficiency of their operations and production, which in the end means improved profits (Gilchrist 2016, p. 8).

The Internet of Things can be described as a network of different devices that can com- municate with each other through the internet (Ramgir 2020, p. 1). Different Industrial Internet solutions typically have very similar structures. Data is gathered by sensors which often are readily existing in a DCS system. From the DCS system, the data, and perhaps some ready-made analysis results, are sent to cloud servers. At this stage the data is filtered in some way such that its size on the server is not unnecessarily big, but the sampling frequency is still good enough for the desired use. (Ramgir 2020, p. 82) From the cloud servers, different online applications analyze and visualize the data giving information about plant, machine, or device operation to aid in decision making. The results are typically visible online from anywhere where internet access is available and may thus be accessed easily in different parts of the plant. (Gilchrist 2016, p. 4; Ramgir 2020, p. 5) Industrial Internet solutions focus on collecting collective data from multiple sources to a cloud server and analyzing and visualizing it to provide information about processes or devices that helps in daily operations.

The structure of an Industrial Internet platform can be described with three different tiers depicted in Figure 7. Tier 1, or the Edge Tier, consists of the sensors and data transfer such as wiring installed into processes that give information about how processes are behaving. Tier 2, or the Platform Tier, takes care of reading the sensor data, doing simple analysis, and transferring the data into the cloud. Tier 3, or the Enterprise Tier, is where the Industrial Internet applications are running, and the data is further processed and visualized to the desired user group. (Gilchrist 2016, p. 76; Ramgir 2020, p. 3)

(28)

Figure 7. The structure of an Industrial Internet platform (adapted from Gilchrist 2016, p. 77).

A typical DCS system has the first two tiers nearly complete. Sensors are installed and the measurement data is being analyzed in the process control servers. Often the data is filtered and sent to a local history database. Transferring the data to a cloud server from the local database is enough to complete the first two tiers. The third tier requires a platform where the data can be analyzed and visualized in the cloud.

The strength of Industrial Internet solutions is the access to large quantities and timeframes of data. All the data available about the process is accessible in one online server and further analysis can be done easily. Often the analysis is done almost real- time, depending on the speeds of the data transfer connections, and online solutions analyzing machine performance as the machine is running are possible. As data is stored to a server, analysis can also be done on history data and long timeframes of operation can be analyzed. Advanced data analysis methods such as artificial intelligence (AI) are often utilized as part of the analysis done. Long timeframes of data allow implementation of artificial intelligence algorithms such as machine learning which aims to learn from data with minimal human intervention. (Ramgir 2020, p. 232)

The major challenges of IIoT applications are data safety, integrity, and validity. When data is transferred and stored online, proper means must be used to secure the data as it is essentially connected to all devices that may access the internet. Data security during storage and transfer must be payed close attention to when creating Industrial Inter- net platforms and applications. (Ramgir 2020, p. 6) Data validity is also critical for successful IIoT solutions. The quality and accuracy of the data should be known and acknowledged before doing analysis on it. Filtering the data improperly before sending it to the server can also result in loss of data accuracy and usability. (Gilchrist 2016, p. 87) Small errors and even failures in measurements may sometimes cause great errors in

(29)

future processing, and this must be considered when using measurements to calculate for example the process performance.

2.2.2 Valmet Industrial Internet

Valmet has its own Industrial Internet of Things platform called Valmet Industrial Internet (VII). Valmet Industrial Internet advertises its offering as solutions that help the customer improve their production in the pulp, paper, and energy production industries. One of the main marketing points of Valmet Industrial Internet is effective combination of data analysis and process experts which provides different benefits for customers. (Valmet 2020A) The Valmet Industrial Internet platform is like the typical IIoT platform. The customers have an existing DCS system, which might be Valmet DNA or a competitor solution, from where data is filtered and pre-processed and saved to a local history database. From the history database, the data can still be pre-processed before being sent to the VII cloud database which can read data from many different plants and sources. From the cloud the data can be easily accessed and processed by different VII applications. Implement- ing Industrial Internet to a DCS plant is quite simple, and only the data transfer to VII cloud servers must be added. Once the connection to the VII platform exists, adding new applications can typically be done purely online unless new measurements are required.

The Valmet Industrial Internet platform is typically connected to a local history database existing in the DCS system. Filtering of the data is typically done before storing it into the local database. From the local database, the data is sent and stored in the Snowflake cloud data platform. From there, data analysis can be done with multiple different applications and programming languages. Data visualization is typically done with Tableau, an external solution aimed for developing Business Intelligence applications.

VII applications aim to be developed in cooperation with process experts from elsewhere in the company and data scientists from the VII team (Valmet 2020A). The aim of this cooperation is to combine process understanding and data-analytics such that the applications truly benefit the customer. In many cases, the cost of a VII application to the customer may depend on the benefit the customer gets from using the application, and thus the business model is performance related.

Existing Valmet Industrial Internet applications analyze data uploaded to servers to aid in plant monitoring, predicting future behavior, and optimizing production (Valmet 2020A). VII solutions are based mainly on analysis and visualization of data in online user interfaces, and typically the aim is to give easy access to process behavior information to aid in decision making. Different VII applications analyze and visualize the data

(30)

in different ways to give deeper insight into what is happening in the process and how it is performing.

Valmet Industrial Internet markets its data analysis capabilities that can be utilized to identify possible improvements in process operation. With enough data, the operation of a plant can be analyzed, and possible improvements identified with the help of process specialists. This expertise is offered as a product in the VII offering. (Valmet 2020A) Most of the VII products are based on data analysis, and this is a strength of Valmet Industrial Internet that should be utilized in new VII product development. Existing VII applications utilize artificial intelligence and machine learning algorithms as part of the data analysis they do.

Valmet Industrial Internet offers solutions that aim to improve energy production in the power generation industry. However, no VII solutions exist that focus on improving turbine performance that could be offered alongside the Valmet DNA turbine control solution, and the thesis aims to propose a concept to fill this empty spot.

2.3 Markets to be entered

In addition to understanding the customer and technology to be used, the current markets to be penetrated and their future must also be understood for a product concept development project to be successful. In this thesis, turbine automation is considered a larger market, under which lies the market of turbine performance improvement solutions. From both markets it is necessary to understand both competitor and Valmet offering. When the markets and their future development are truly understood during product concept development, the outcome entering the markets will have advantage over the competitor offering and it will not conflict with, but rather supplement, existing Valmet offering. In addition, in order to help understand the customer business model and possible savings coming from the proposed concept, it is good to understand the market steam power plants are working in.

2.3.1 Steam power plant product markets

In CHP plants and steam plants producing multiple products, controlling the steam turbine regulates and controls the production of the plant. The amount of electricity, heat, and process steam produced can be adjusted by changing the operation of the turbine.

The electricity market is quite similar globally. For an electric grid frequency to stay stable, the consumption and production to the grid must always be balanced (Lin & Mag- nago 2017, p. 212). In most electricity markets around the globe, the electricity pricing is

Improving Steam Turbine Performance with Industrial Internet

Topias Leppänen