Spectroscopy Since 1975

Critical sampling in the cement industry: economic drivers

Martin Lischka

HERZOG Maschinenfabrik GmbH & Co. KG, Germany

The total global cement production in 2020 was around 4.1 billion tons, making it the industrial processes sector responsible for the highest single contribution of emitted CO2 worldwide, with no less than 27 % of the directly industrial-released CO2.1 Modern rotary kilns in cement plants have a production capacity of 5000–10,000 t per day, and for each ton of clinker produced, ~910 kg CO2 are emitted to the atmosphere.2 These emissions stem from three main sources: i) decarbonisation of limestone, ii) fuel for the rotary kiln and iii) fuel for the electricity consumption of the cement plant. There is a vital sampling role hidden away in this big picture, illustrated here with five scenarios for a critical process control parameter termed “LSF” (Lime Saturation Factor), the economic impact of which is the main focus here.

CO2 budgets

In order to meet international agreements on climate change targets, and with introduction of “CO2 certificate trading” in Europe in 2005, in addition to diligent process control, a new aspect for successful and economic cement plant operation arises. Due to CO2 certificate trading, the importance of reliable sampling in cement production must be considered from the point of view of the lowest possible CO2 production and the highest possible reliability of the data obtained.3 Studies have shown4 that a 5 % variation in the single most important process monitoring parameter, LSF (see Technical Info Box), leads to an increase in CO2 emissions of up to 16.4 kg CO2 / t clinker. Likewise, CO2 emission from carbon-based fuels, by a similar 5 % variation in LSF, increases by 17.2 kg CO2  / t clinker.

A sampling bias can very easily be introduced regarding the LSF, which can have severely amplified economic consequences.

 

Technical Info Box

Compared to many traditional mining and minerals processing industries based on heterogeneous mineralisations and materials (e.g. base metals, gold ores), cement production is based on relatively homogeneous raw materials (clay, limestone), supplemented by a few aggregates to ensure consistent product quality. Traditionally, therefore, rather less attention has been paid to the strictness of the TOS within this industry. Sampling of the clinker is typically performed from the running process stream with a cycle of one sample per hour. After sampling, the clinker is coarsely crushed in a jaw crusher to a grain size of less than 5 mm. This allows representative sampling to reduce the sample quantity to approximately 100 g. In modern plants, samples are transported to the laboratory by pneumatic transportation. In the laboratory, sub-samples are finely ground (< 45 µm) and prepared for automated X-ray fluorescence (XRF) and X-ray diffraction (XRD) analysis. To be able to use automated analysers, only about 10–15 g of sample material is needed, which is pressed into a steel ring (Ø 51.5 mm). Since the penetration depth of the analyser’s X-rays is only a few micrometres, in reality only a very small portion of these few grams is analysed. It is obvious that sampling plays a critical role in this measuring system context. The effective sampling rate (clinker-to-aliquot) is closely related to the clinker production rate (see Table 1) but can be estimated as ~1 : 50,000,000—which under all circumstances is daunting.

However, the subsequent sample preparation also has a considerable influence on the analytical result. A measurable parameter for the quality of sub-sampling and sample preparation is the standard deviation, used as a measure of spread between replicated sampling and analysis results.

In addition to the classical elemental breakdown of chemical analysis, three so-called moduli are used in the cement industry for chemical classification. The most important of these is the so-called Lime Saturation Factor (LSF) which is calculated as follows:5

LSF = 100 × CaO / (2.8 × SiO2 + 0.65 × Fe2O3 + 1.18 × Al2O3)

The three critical moduli are used to monitor and control the production targets. During the cement manufacturing process, heterogeneity of the intermediate products decreases continuously from the raw mixture to the finished product (good process control). The composition of the raw material mix and of the secondary fuels used are of significant importance for the clinker burning process efficiency, and also have a decisive influence on the composition of the clinker. Process control must, therefore, be carried out in such a way that the chemical and physical properties of the clinker remain as constant as possible. For this sensitive target, the quality, representativity and reliability of process sampling operations ARE of key importance.

The economics of it all

To illustrate the economic consequences of these technical relationships, one estimates the current financial impact based on a certificate price of €55 t–1 CO2 (even though increasing prices can be expected for the next years). The economic consequences of non-optimal LSF estimation are huge, as shown in Table 1. Here a relative error for the LSF ranging from 1 % to 5 % is considered, correlated to the simulation data given by Cao et al.4 for typical daily production rates.

Table 1. Estimated additional CO2 release for different production capacities caused by erroneously determined LSFs and the financial impact in terms of CO2 certificate price trading. These certificate costs could be saved by running the cement plant with a well-controlled process close to product specifications and with optimised power consumption.

 

Rel error (%)
LSF factor

Production in t/day

1000

2000

5000

10,000

Additional release (kg CO2 / day)

Clinker

1

3280

6560

16,400

32,800

2

6560

13,120

32,800

65,600

3

9840

19,680

49,200

98,400

4

13,120

26,240

65,600

131,200

5

16,400

32,800

82,000

164,000

Fuel

1

3440

6880

17,200

34,400

2

6880

13,760

34,400

68,800

3

10,320

20,640

51,600

103,200

4

13,760

27,520

68,800

137,600

5

17,200

34,400

86,000

172,000

Estimated costs for CO2 certificate (€)

Day

1

370

739

1848

3696

2

739

1478

3696

7392

3

1109

2218

5544

11,088

4

1478

2957

7392

14,784

5

1848

3696

9240

18,480

Year (300 days)

1

110,880

221,760

554,400

1,108,800

2

221,760

443,520

1,108,800

2,217,600

3

332,640

665,280

1,663,200

3,326,400

4

443,520

887,040

2,217,600

4,435,200

5

554,400

1,108,800

2,772,000

5,544,000

Highly sensitive sampling

It is very easy to introduce a significant variability in process monitoring and control if proper attention is not brought to bear—making representative process sampling essential. This can be illustrated for the same LSF parameter, based on XRF measurements. Results are presented below from an analysis repeatability test (10 analytical results from the same sample). One re-analysis shows an “accidental” higher amount of Fe2O3 which, however, changes the average LSF magnitude significantly, from 105.44 to 102.15. This single sample preparation variation is consequently responsible for a relative error of ~4 % for the LSF, Table 2. With the economic impact of even small LSF variations as shown in Table 1, all sampling, sub-sampling and sample preparation variability is decidedly unwanted. TOS to the fore!

Table 2. Routine XRF analytical results from a simple replication experiment (10 analytical aliquots prepared from the same sample) showing how easily the LSF can be impacted by non-representative sampling, preparation or analytical inconsistences. The primary clinker sampling variability must be added to this error, which is solely due to sample preparation and analysis.

Test

Al2O3

SiO2

CaO

Fe2O3

LSF

1

4.39

20.17

67.27

2.84

105.93

2

4.38

20.19

67.21

2.81

105.78

3

4.42

20.33

67.40

2.79

105.40

4

4.41

20.33

67.41

2.80

105.41

5

4.42

20.33

67.43

2.83

105.39

6

4.43

20.24

67.33

2.79

105.67

7

4.42

20.34

67.33

2.81

105.21

8

4.44

20.39

66.63

4.48

102.15

9

4.48

20.46

67.54

2.77

104.92

10

4.46

20.38

67.51

2.81

105.26

Mean

4.42

20.32

67.31

2.97

 

SD

0.03

0.08

0.25

0.50

 

RSD

0.6 %

0.4 %

0.4 %

16.9 %

 

Insight leads to greater climate responsibility

The above economic relationships define three main goals for continuing vigilance regarding optimised cement production control to be in optimal compliance with increasingly stringent climate policy efforts, which today should be included in sustainability reports from all forward-looking cement manufacturers:

  • Process and product specifications, as close as possible to minimum climate impact demands
  • Design of alternative, more climate-friendly cement products
  • Low-energy operation and low-CO2 cement plant emissions

Thus, today there are both environmental, technological, economical (plant scale, global climate scale) as well as somewhat “hidden” sampling drivers for a continuously evolving cement industry—no longer mainly driven by narrow economic incentives alone. The TOS has a role to play nearly everywhere, and the economic costs for even a minor lassitude can be substantial, as was shown above (Table 1), in which a LSF uncertainty of 4 % (rel) results in estimated potential additional certificate cost of €4.4 M per year.

There are other, non-optimised sampling issues in cement production, first and foremost primary clinker sampling. Often scoop sampling is applied in this stage, a sampling method that critically needs to be reconsidered, because a complete cross-section of the process stream is traditionally considered “almost impossible” to achieve. Remarkably there are not many publicly available clinker sampling rate estimates, nor assessments of the associated sampling errors.

References

  1. IEA, Technology Roadmap - Low-Carbon Transition in the Cement Industry. IEA, Paris (2018). https://www.iea.org/reports/technology-roadmap-low-carbon-transition-in-the-cement-industry [accessed 30 June 2021]
  2. P. Stemmermann, U. Schweike, K. Garbev and G. Beuchle, “Celitement – a sustainable prospect for the cement industry”, Cement Int. 8(5), 52–66 (2010). https://celitement.de/wp-content/uploads/2020/07/2010-10-26_Celitement_a_sustainable_prospect_for_the_cement_industry-1.pdf
  3. C. Wagner and K.H. Esbensen, “A systematic approach to assessing measurement uncertainty for CO2 Emissions from coal-fired powerplants—missing contributions from the Theory of Sampling (TOS)”, Chem. Eng. Res. Des. 89(9), 1572–1586 (2011). https://doi.org/10.1016/j.cherd.2011.02.028
  4. Z. Cao, L. Shen, J. Zhao, L. Liu, S. Zhong and Y. Yang, “Modeling the dynamic mechanism between cement CO2 emissions and clinker quality to realize low-carbon cement”, Resour. Conserv. Recy. 113, 116–126 (2016). https://doi.org/10.1016/j.resconrec.2016.06.011
  5. VDZ, Zement Taschenbuch. Verein Deutscher Zementwerke (2008). https://www.vdz-online.de/wissensportal?tx_vdzknowledgebase_pi1%5Baction%5D=detail&tx_vdzknowledgebase_pi1%5Barticle_preview%5D=6392&tx_vdzknowledgebase_pi1%5Bcontroller%5D=Article&tx_vdzknowledgebase_pi1%5Btype%5D=0&cHash=790040ed0c1e7f3fd35d7eaa91c51373 [accessed 30 June 2021]
 
Martin Lischka

Martin Lischka

Mr Martin Lischka (MSc Geosciences and Environment) has more than ten years of experience in the field of sample taking and sample preparation. He is currently working in the R&D department at HERZOG Maschinenfabrik GmbH & Co. KG. Projects he is involved range from special sampling systems, large scale raw material applications, down to final aliquot preparation—like pulverisation, pressed pellet preparation, borate fusion for XRF analysis and many more. His recent activities focus on precious metal recycling, copper-related commodities and sensoring methods applied to sample taking and preparation routines as a quality measure.
m.lischka@herzog-maschinenfabrik.de

Rate this Article: 
No votes yet