To determine dynamic response of mechanical system is always a complex issue. One of the core inputs for such a task is a coefficient of logarithmic damping decrement. There is no way how to determine this coefficient analytically; it needs to be measured on the final product. There has been a large amount of dynamic tests carried out on common structures and the results have been incorporated in design codes. Nevertheless, power plant boiler structure differs significantly from common structures. To provide a large number of dynamic tests on real facility is very difficult for many reasons. Using a scaled model is a way how to achieve relevant data based on repeatable experiment.
Physical model ; Scaling law ; Boiler ; Modal analysis ; Experiment
Structural conception of power plant boiler structure is driven by its technological purpose. According to boiler type there is a variety of types of boiler power plant structures. This article focuses to heat recovery steam boilers with vertical direction of flue gas (Fig. 1 ). This type of boiler is commonly known as a vertical HRSG boiler. Structural system of vertical HRSG boilers is quite simple and the differences compared to the common structure are well recognizable.

Fig. 1. Vertical HRSG boiler (real facility on the left, 3D FEM model on the right).

The majority of mass in structural system is made by tube modules and boiler casing. Either modules or boiler casing slang from boiler upmost part called “boiler ceiling”. Boiler ceiling is created by massive beams coming through the flue gas duct. These beams are in full contact with hot flue gas and that is why they cannot be fixed to the supporting structure. The suspended boiler body is fixed in its position by a system of vertical guiding and side stoppers. The detail of guiding allows thermal dilatation. This brief summary of boiler structure specifics shows that mechanical damping properties could have been bigger in HRSG structure compared to the common steel structure (Carpinteri et al., 2011 ).
Mechanical damping assessment is possible only by means of experimental testing (Cajka and Krejsa, 2013 and Li, 2010 ). Organizing this type of dynamic test on real facility is practically hardly feasible (Dusicka and Iwai, 2007 ). Moreover, the corresponding value of one unique dynamic test would have been limited. For that reason the solution on scaled physical model was chosen (Graczyk and Moan, 2011 ). Subject of the research was an experimental modal analysis carried out on one scaled physical model. This model was modified into five alternative designs during the testing. Three of the modifications represented one isolated specific in view. First and fifth modification represented simple steel structure without any technology. Logarithmic damping decrement applicable for real vertical HRSG boilers was determined on base of proportional change of property in view assessed for model with technology compared to a model without any technology (Maheedhara Reddy and Diwakar Reddy, 2014 ).
Fundamental principle of theory of similitude between real product and its scaled model is a conversion of the technical task in view into its nondimensional expression (Melcer and Kucharova, 2014 ). This way we can derive so called similitude invariant numbers which take the same value for both the product and its scaled model. The dimensional analysis of the specific task needs to be provided to perform these invariants. So the solution is not universal. There is a need to provide this type of analysis for every type of technical problems (Protivinsky and Krejsa, 2015 ).
Modal analysis of mechanical systems is defined by nine mutually independent physical quantities (Merczel et al., 2013 ). By means of these quantities we can express the modal analysis problem in a perfect way. All nine relevant quantities with their expression in technical system LTM (length, time and mass) are provided in Table 1 (Salajka et al., 2013 ).
Symbol  Scaled unit  Dimension in LTM 

l  Characteristic length  L 
ρ  Specific weight  ML^{−3} 
E  Modulus of elasticity  ML^{−1} T^{−2} 
t  Characteristic time  T 
ω  Characteristic frequency  T^{−1} 
σ  Stress  ML^{−1} T^{−2} 
F  Force  MLT^{−2} 
a  Characteristic acceleration  LT^{−2} 
c  Coefficient of viscous damping  MT^{−1} 
On base of relevant quantities we can make up dimensional matrix of its exponents in LTM system. Determinant of submatrix composed from last three matrix columns is different from zero. Therefore the rank of the matrix is equal to three. In agreement with PIBuckinghams theorem, the number of similitude invariant numbers is equal to the difference between the number of relevant quantities and the rank of dimensional matrix. Therefore the number of final similitude invariant numbers is equal to six. Representation of π parameters will be provided by using the formula (1) solving unknown exponents x_{1} to x_{9} . Since there are just three quotations for estimating nine unknowns we are choosing six linear independent choices for the six redundant exponents x_{4} to x_{9} .

( 1) 
By solving the Eq. (1) for i = 6 exponent choices x_{4} to x_{9} we receive six dimensionless parameters π_{1} to π_{6} . Since the π parameters are compound by relevant quantities and at the same time their numerical representation shall be equal for the original product, its scaled model can derive scaling laws for every relevant physical quantity just knowing π parameters. Unfortunately the scaling laws computed from particular π parameters contradict each other very often. It is on a researcher to neglect the weak laws not significant for main focus of his research. The contradiction comes from the fact that we choose a limited number of scales equal to the rank of dimensional matrix only. Moreover, choosing the length scale and model material we run out of all select scales. The rest of scales are computed.
PMMA material was chosen as the most suitable material for a physical model (Fig. 2 ). PMMA is a commonly used material in modeling practice (Protivinsky and Krejsa, 2012 ). PMMA has linear response to loading, quite small internal damping, and connections provided by the PMMA glue behaving as basic material (Pospisil et al., 2014 ). A brief summary of PMMA material properties is shown in Table 2 .

Fig. 2. Physical model (on the left – X direction view, on the right – Y direction view).

Symbol  Physical quantity  Value  Unit 

ρ  Specific weight  1190  kgm^{−3} 
E  Modulus of elasticity  3200  MPa 
f_{y}  Yield point  60  MPa 
Scaling laws of relevant physical quantities are shown in Table 3 . The gravity law was considered as weak law.
Symbol  Physical quantity  Scaling law (scale)  Note 

L  Characteristic length  1:20 = 0.05  Geometry scale 
P  Specific weight  1.19:7.85 = 0.15159  Determined by the material choice 
E  Modulus of elasticity  3.2:210 = 0.015238  Determined by the material choice 
T  Characteristic time  1:6.341 = 0.15770  Computed field 
Ω  Characteristic frequency  6.341:1 = 6.341  Computed field 
A  Characteristic acceleration  1:2 = 0.5  Computed field, considered as weak law 
A complete set of dynamic tests was carried out in the experimental research center in ITAM Prague. Modal analysis was performed using industrial computer Dewetron DW 2010 and exciter Brüel&Kjaer, type 8202. The tests were carried out repeatedly on one single model in five particular modifications. A brief description of the model modifications is shown in Table 4 . There were twelve measurement tests for every single modification in each of horizontal directions. There were six accelerometric singleaxial sensors fixed to the model. Location of fixing points was designed for evaluating both flexural shapes of natural vibration (Kralik, 2013 ).
Modification number  Model description 

K_{1}  Unloaded model – structure only 
K_{2}  Hanging technology fixed to structure by springs 
K_{3}  Hanging technology – no springs 
K_{4}  Hanging technology – no springs, beams supporting slung technology fixed on ballbearing 
K_{5}  Unloaded model – structure only 
Evaluation of the data recording of tail vibration shows that the PMMA scaled model is applicable for the specified task. In the first phase of tail there is significant mutual influence of adjacent natural shapes in evidence. Nevertheless, damping of the PMMA system is low enough to get in the late phase of tail where only the first dominant natural shape occurs.
Logarithmic decrement was determined from the late phase of tail according to formula (2) (Melcer and Kucharova, 2014 ).

( 2) 
The n in (2) formula represents number of periods and the v_{0} represents amplitude of the measured physical quantity (Table 5 ).
Modification number  Logarithmic damping decrement  Rate to modification K_{1} (rate to common structure) 

K_{1}  0.059  1.0 
K_{2}  0.135  2.3 
K_{3}  0.154  2.6 
K_{4}  0.48  8.1 
K_{5}  0.065  1.1 
This paper brings results of the first part of the experimental research on mechanical damping of power plant boiler structure. Time data recording of the vibration tail shows two phases of vibration tail. In the first phase there is visible significant mutual influence of the adjacent natural frequencies. In the second phase there is only one dominant natural frequency visible. All the evaluation was done on the basis of this second phase. The cardinal task for every research based on scaled physical models is how to converse the experimental results gained on scaled model to a real structural design. The methodology used in this case was very simple. The damping decrement gained for every particular model modification was divided by the decrement pertained for modification K_{1} . This modification represents common steel structure. Every other modification represents some specific of vertical HRSG steel structure. This methodology is clear and universal. Results determined in this way are applicable for structural design of real vertical HRSG boiler structures ( Brozovsky and Dufka, 2014 , Jendzelovsky and Balaz, 2014 , Protivinsky and Krejsa, 2014 , Vican and Janik, 2014 and Protivinsky and Krejsa, 2015 ).
In next steps of this research it is planned to evaluate the vibrating tail from harmonic force with frequency of loading equal to first dominant natural ones. These measurements are to give countercheck of the recent conclusion. Moreover, it will be possible to evaluate the damping even based on the first phase of vibration tail.
Although there are some more tests ahead we can claim that vertical HRSG structures have significantly bigger damping than that applicable for common steel structures.
The authors declare that there is no conflict of interest.
This paper has been completed thanks to the financial support provided to VSB – Technical University of Ostrava by the Czech Ministry of Education, Youth and Sports from the budget for conceptual development of science, research and innovations for the year 2015.
Published on 05/10/16
Licence: Other
Are you one of the authors of this document?