Evaluating energy efficiency on a coal power plant’s performance

November 20th, 2014, Published in Articles: Energize, Featured: Energize

 

Eskom has embarked on energy efficiency initiatives on its coal boiler plants in a bid to decrease the amount of coal burnt and in turn increase the electricity generated. This study focused on the analysis of the before and after outage data obtained from the unit cards in one of the Eskom’s “once through” 600 MW coal boilers with a mechanical conversion efficiency of 35%.

Multiple linear regression models were developed and built to predict the power generated and sent out in correlation with the predictors (average air heater temperature, average main stream super heater temperature, average high pressure heater temperature, the total mass of coal burnt, auxiliary power consumption, average of the cold well and hot well condenser temperature and pressure). The data obtained 3 months before and after an outage showed an average power generated of 434,95 MW and 502,08 MW respectively. The results also revealed that the cumulative energy gained was 44 000 MWh.

The industrial and commercial sectors in South Africa consume more than 80% of the electricity generated by Eskom [1]. Fig. 1 shows the electrical energy distribution in all the three sectors in South Africa.

Fig. 1: Electrical energy distribution  in South Africa.

Fig. 1: Electrical energy distribution
in South Africa.

Coal composition

Since the primary fuel used in every coal thermal power plant is coal, it is of relevance to describe its composition. Coal’s main constituents are carbon, hydrogen, oxygen, sulphur, nitrogen and ash. Fig. 2 shows the elemental percentages of the composition of coal [2].

Fig. 2: Shows the elementary composition of coal.

Fig. 2: Shows the elementary composition of coal.

Interestingly coal can be ranked depending on the percentage of carbon present in a particular type of coal. Generally, coal can be classified into six broad categories, namely:

  • Peat: This coal type has carbon percentage ranging from 20 to 25%. It is a very low grade coal.
  • Lignite: This is  real coal, but of low grade with a very high moisture content.
  • Sub-butiminous: This has lower carbon percentage as compared to bituminous coal and can also be termed as a low grade coal.
  • Butiminous: It has a high carbon content (65-85%) and low oxygen content. It burns with a yellow flame and can be termed as a high grade coal.
  • Semi-anthracite: It is usually referred to as hard coal and has coal content  between 80 and 85%. It is a very high grade coal, but softer in strength to anthracite.
  • Anthracite: This is the hardest of the solid coal, with a carbon content ranging from 90 to 98%. It burns with very little smoke. It is a very high grade coal.

The bulk of the coal in South Africa falls in the group of bituminous, semi anthracite and anthracite with carbon content in the range of 65 to 90% [3].

The challenge of generating sufficient electrical energy to meet the country’s demand requires that very efficient coal thermal power plants be provided. There are basically two categories of coal boiler plant; namely the “drum” and the “once through” types. The “once through” type is very cost effective and with a far better performance efficiency than the “drum” type.

Basic operation of a “once through” coal thermal power plant

The coal thermal power plant operates on the principle of Rankine’s cycle with water/steam as the refrigerant. The main components involved in the coal thermal power plant are the boiler, superheater, turbine, condenser, feed water pump and economiser [4]. The coal stored in the coal hopper is conveyed by conveyor to the pulveriser, where it is crushed and mixed with preheated air before finally blasted into the boiler. This fuel-gas mixture is ignited with the help of oil. Furthermore, with the aid of a water feed pump, demineralised water enters the boiler after being heated by the high pressure heater and the economiser. In the boiler, the desuperheated steam picks up latent heat at the evaporator and rises upward in the furnace, where it is further heated to super-heated steam by cascaded superheaters. This superheated steam is finally inputted into the turbine. The thermal energy in the superheated steam is extracted and converted to mechanical energy which rotates the turbine blades. The steam driven turbine is coupled to an alternator and acts as a prime mover. The rotation of the turbine blades drives the rotor which, as it rotates, cuts through the stator’s magnetic field lines. This results in an induced electromotive force due to the principle of electromagnetic induction. The generated alternating electricity is fed to a step up three phase transformer where the voltage is increased. This is sent to the high transmission lines via circuit breakers and busbars.

It is crucial to understand that after the extraction of thermal energy from the superheated steam as it flows into the turbine, it becomes wet steam and flows into the condenser where the condensation process takes place. The condensate exits the condenser at low temperature and pressure. This fluid is pushed into a low and high pressure heater and the economiser by water feed pump where it becomes preheated to a high temperature and pressure and finally fed back to the boiler. The flue gases produced during the combustion of coal are ejected via the stack chimney after passing through various heat exchangers, including superheater, reheater, economiser and air heater. The fly ash is collected by an electrostatic precipitator while the bottom ash is collected in an ash dump attached to the bottom of the boiler.

The modern coal thermal plant contains a cooling tower at the condenser unit which prevents the cold water from the river or pond which is fed to the inlet of the condenser from exiting at the outlet as hot, high pressure water and thereby polluting the river [5]. The cooling tower ensures that the heat gained by the induced water into the condenser is extracted by air and dissipated into the environment. The schematic diagram of a typical “once through” coal power plant is shown in Fig. 3.

Fig. 3: The temperature versus entropy diagram for a typical “once through” coal power plant.

Fig. 3: The temperature versus entropy diagram for a typical “once through” coal power plant.

The temperature versus entropy diagram illustrating the Rankine’s cycle for a coal thermal plant is shown in Fig. 4. The superheated steam from the superheater is fed to the high pressure turbine while the superheated steam feeding the intermediate and low pressure turbine comes from the reheaters. This process is known as bleeding and enhances the efficiency of the power plant [6]. The area enclosed by the cyclic process is proportional to useful thermal energy converted by the turbine. The larger the area, the better the performance of the plant.

Fig. 4: Schematic diagram of the coal power plant.

Fig. 4: Schematic diagram of the coal power plant.

 

Determining the thermal efficiency of a coal thermal plant

The main components relevant for the determination of thermodynamic efficiency of a coal thermal plant are the super heater (hot reservoir (T1)), the condenser (cold reservoir (T2)) and the turbine (cycle engine). The efficiency of a coal thermal plant with respect to the mechanical conversion ratio of the turbine can be defined as the ratio of the output mechanical work to the input thermal energy with respect to the two reservoir temperatures (superheater and condenser). The efficiency can be derived from the fundamental law of conservation of energy and is illustrated by equations 1 to 3.
06-GT-Fort-Hare-Eqn.01

(1)

 

06-GT-Fort-Hare-Eqn.02

(2)

 

06-GT-Fort-Hare-Eqn.03

(3)

 

 

where S is the entropy and T1 and T2 are temperatures in Kelvin (K).

Objectives

The primary objectives of the research were focused on:

l    Derivation of multiple linear regression mathematical models to predict the power generated and the power sent out using the following predictors:

i) The mass of coal burnt (M)

ii) Super-heated steam temperature (Ts)

iii) High pressure heater temperature (Th)

iv) Average condenser pressure (Pc)

v) Average condenser temperature (Tc)

vi) Air heater temperature (Ta)

vii) Auxiliary power (Pa)

  • To deduce an improvement in the power generated by the coal boiler unit by comparing the predictors’ results obtained before and after outage using the developed and built regression models.
  • To determine the energy savings from both the power generated and sent out power due to the energy efficiency intervention.

Methodology

Data for both the predictors and the responses are obtained from the installed unit cards of the coal boiler power plant from Eskom’s control and instrumentation stations for the period from January to March 2013 corresponding to before the outage of the coal boiler plant and from September to November 2013 designated as the after the outage. All the unit cards were configured to log every 30 minute interval.

This data was further processed into average weekday, average Saturday and average Sunday of each of the respective months included in the data set. Multiple linear regression models of the power generated and the power sent out before and after the outage were developed and built incorporating the inputs (air heater temperature, superheater temperature, high pressure heater, condenser pressure, condenser temperature, mass of coal burnt and the auxiliary power consumption).

Results and discussion

Determining multiple linear regression mathematical models

The models developed used the ordinary least square technique for solving the multiple linear regression [7].  The accuracy of the built models were tested by ensuring that a strong linear correlation existed between the modelled desired output and the actual measured output for the before and after the outage scenarios [8]. The equations for each of the multiple linear regression models can be represented by Eqns. 4 and 5:

Each of the parameters and their scaling values in the mathematical models given in Eqns. 4 and 5 are shown in Table 1 and Table 2.

Pg = β0 + β1Ta + β2Ts + β3Th + β4Pc + β5Tc + β6M                (4)

Ps = β0 + β1Ta + β2Ts + β3Th + β4Pc + β5Tc + β6M + β7Pa    (5)

 

Table 1: The mathematical models of the power generated.
Input parameters Scaling constants Model before
scaling values
Model after
scaling values
Output
Ta β1 -0,212 -0,713 Pg
Ts β2 -0,714 3,604
Th β3 0,934 5,828
Pc β4 0,497 -9,239
Tc β5 0,407 1,849
M β6 2,030 1,654
βo 203,87 -2955
Table 2: Shows the mathematical models of the power sent out.
Input parameters Scaling constants Model before scaling values Model after scaling values Output
Ta β1 -0,181 -0,095 Ps
Ts β2 -0,032 1,362
Th β3 1,532 5,163
Pc β4 -3,187 8,51
Tc β5 0,614 -4,515
M β6 0,487 0,012
Pa β7 21,49 40,36
βo -417,2 -2166

Ranking of predictors according to their affect on the output

The relief-F statistical algorithm was used to rank the input parameters by weight of importance to the generated power. The relief-F algorithm ranked predictors by weight of importance from 1 (if the predictor is a primary factor to the output and shows a very strong correlation) to -1 (if the predictor is a secondary factor and shows a very weak correlation).  The relief-F algorithm demonstrated that all predictors were primary factors to the power generated with the high pressure heater and the superheater steam temperatures contributing the most, while the air heater temperature contributed the least. Fig. 5 illustrates the weight ranking of importance of the predictors to the power generated after the outage of the coal boiler unit. It was also observed that the mass of coal burnt contributed to the power generated more than the contribution from the condenser pressure and temperature.

Fig. 5: Illustrates the ranking of predictors to their contribution to the power generated.

Fig. 5: Illustrates the ranking of predictors to their contribution to the power generated.

Air heater temperature profiles before and after the outage

It was observed that the air heater temperature was a primary factor of the load generated and an increase in the temperature could result in a corresponding increase in the load generated. The profile of the air heater temperature of the average day for the final month (March 2013) before the outage and the first month (September 2013) after the outage was compared and illustrated in Fig. 6 over a 24 hour interval.

Fig. 6: The average day mean air heater temperature profiles.

Fig. 6: The average day mean air heater temperature profiles.

It can be clearly delineated that the profile after the outage (average temperature 273°C) was higher than before outage (average temperature 224°C) scenario. This confirms an improvement in the efficiency of the coal boiler unit.

Superheater steam temperature profiles before and after the outage

Fig. 7 shows the profiles of the superheater steam temperatures for the both scenarios. The average temperature was 533°C after the outage, and 459°C before the outage, of an average day over a 24 hour interval. Since this factor is one of the most contributors, the generated load after the outage will increase. The increase in the superheated steam temperature can lead to an increase in the mechanical conversion efficiency of the power plant.

Fig. 7: The average day mean super heater temperature profiles.

Fig. 7: The average day mean super heater temperature profiles.

Condenser temperature profiles before and after the outage

The condenser profiles of the both scenarios show no mean significant difference, as the average temperature before outage was 42,7°C and after the outage was 43,0°C. The both profiles are shown in Fig. 8.

Fig. 8: The average day mean condenser temperature profiles.

Fig. 8: The average day mean condenser temperature profiles.

Comparison of the coal burnt and power generated before and after the outage

Table 3 summarises the average day of the last month (March 2013) before the outage and the first month (September 2013) after the outage.

Table 3: Average month-day input and output parameters for the before and after outage. 
Parameters Average day
Last month before outage First month after outage
Ta 224,46°C 273,10°C
Ts 459,39°C 533,15°C
Th 230,87°C 233,05°C
Pc 7,58 Mbar 7,22 Mbar
Tc 42,74°C 43,04°C
M 212,25 t 219,63 t
Pa 22,63 MW 21,72 MW
Pg 495,86 MW 505,90 MW
Ps 473,23 MW 484,18 MW

where:

Ta = Air heater temperature

Ts = Superheated steam temperature

Th = High pressure heater temperature

Pc = Average condenser pressure

Tc = Average condenser temperature

Pa = Auxiliary power

M = The mass of coal burnt

Pg = Power generated

Ps = Sent out power

It can be confirmed from Table 3 that after the energy efficiency intervention, there was an increase in both the power generated and sent out from the coal thermal power plant. More importantly and without loss of generality, the average auxiliary power after the outage (21,72 MW) was lower than the average auxiliary power before the outage (22,63 MW).

Determining the average power generation increment for a month after the intervention

The average increment in the power generated in the month of September (after the outage) was calculated from the profiles derived from the mathematical multiple linear regression equations using the predictors’ data set for the said month. The solid line with star markers in Fig. 9 represents the actual measured power generated profile, the dash-line with diamond markers, represents the mathematical model of the after outage profile while the line with ring markers represents the extrapolated model profile of the power generated in the scenario where no energy efficient intervention was carried out at the plant.

Fig. 9: Actual measured after outage and before outage average month power generated.

Fig. 9: Actual measured after outage and before outage average month power generated.

The demand increment was determined from the difference between the model’s profile after the outage and the model’s profile before the outage using the predictors’ data for of the month of September. The average increase in the power generated for an average weekday, average Saturday and average Sunday were determined to be 63,00 MW, 71,14 MW and 110,0 MW, respectively.

Determining the total energy increment over the three month period

Owing to the intervention by the energy efficiency initiative, the energy increment was determined from the 1 September to 30 November 2013. Table 4 summarises the average weekday, average Saturday and average Sunday of the energy generated and energy sent out from the actual measurement, the model of the actual measurement and the model of the baseline adjustment scenarios. The cumulative energy increment of the from the power generated and sent out energy for the three months was 4,40 x 104 and 7,40 x 104 MWh respectively.

 Table 4: Summary of the actual and models energy generated and sent out after the outage.
Generated energy x 104 MWh Sent out energy x 104 MWh
Time Actual M after M before Actual M after M before
September AWD 1,23 1,20 1,19 1,19 1,17 1,04
ASA 1,21 1,20 1,15 1,16 1,16 1,00
ASU 1,20 1,20 1,05 1,15 1,15 0,92
Month 36,7 36,2 34,8 35,4 34,9 32,8
October AWD 1,15 1,18 1,13 1,09 1,07 1,03
ASA 1,16 1,15 1,04 1,10 1,09 0,91
ASU 1,13 1,12 1,11 1,07 1,05 1,02
Month 35,6 35,5 32,8 33,7 33,5 31,4
November AWD 1,26 1,25 1,22 1,19 1,20 1,05
ASA 1,26 1,25 1,19 1,18 1,19 1,02
ASU 1,25 1,25 1,23 1,18 1,17 1,08
Month 37,8 37,6 36,3 35,6 34,2 31,5
M = Mathematical model, AWD = average weekday, ASA = average Saturday,
ASU = average Sunday

Furthermore, it can be deduced that there was an increment in both power generated and sent out for each of the three months. The actual measured power generated and sent out showed a negligible difference to the after-outage modelled results.

Conclusion

The increase in power generated and energy gain achieved was determined with a high level of confidence due to the development and building of the multiple linear regression models. The additional power sent out was also reported by the national grids, since the amount of the power consumed by the auxiliary components was far lower than the power generated. It can also be affirmed that auxiliary component power consumption was between 20 to 30 MW. Energy efficiency initiatives in power plants also improve the life of the plant’s performance and enables a systematic testing of an individual unit’s performance.

Acknowledgements

Fort Hare Institute of Technology and Eskom are acknowledged for their financial support of this research project. The authors also acknowledge Eskom’s control and instrumentation engineering station for providing them with the data from the power plant.

This paper was presented at the industrial and commercial use of energy conference (ICUE) in August 2014, and is published here with permission.

References

[1]    Department of Minerals and Energy: “Integrated Energy plan for the Republic of South Africa”, 2003.
[2]    www.btinternet.com/~ian.rivett/imic/combust.htm: “Elementary percentage composition of Coal”, 2012.
[3]    www.eia.doe.gov: “World recovery return statistics”.
[4]    S Sivanagaraju, M Balasubba Reddy and D Srilatha,: “Generation and Utilisation of Electrical Energy”, 2010.
[5]    DP Kothari and IJ Nagrath: “Modern power system analysis”, 2003.
[6]    Babcock and Wilcox: “Steam: Its Generation and Use” 2005.
[7]    S Chatterjee, and AS Hadi: “Influential Observations, High Leverage Points, and Outliers in Linear Regression”, 1986.
[8]    M Robnik-Sikonja and I Kononenko: “Theoretical and empirical analysis of ReliefF and RReliefF Machine Learning”, 2003.
[9]    Math Works Corporation, Matlab and Simulink: “Math work cooperation 2012b”
(Version 7.12), 2012.

Contact Stephen Tangwe, Fort Hare Institute of Technology, Tel 078 307-6922, stangwe@ufh.ac.za

Related Articles

  • South African Government COVID-19 Corona Virus Resource Portal
  • Ministerial determinations propose 13813 MW of new-build by IPPs, none by Eskom
  • Crunch time for South Africa’s national nuclear company, Necsa
  • Dealing with the elephant in the room that is Eskom…
  • Interview with Minerals & Energy Minister Gwede Mantashe