Hydropower'15
Stavanger, Norway 15-16 June 2015
Failures and durability analysis of a small hydro power plant
Ales Drab, Jaromir Riha & Miroslav Spano
Institute of Water Structures, Faculty of Civil Engineering, Brno University of Technology, Czech Republic
ABSTRACT:
In the Czech Republic the environmental flow in streams is frequently exploited for power generation at small
hydropower plants located at dams. The performance of hydropower plants is significantly influenced by the
durability and service life of installed technical equipment, which in the case of small hydropower plants is often
simple and inexpensive. The study presents the results of durability analysis based on the analysis of defects and
failures of two turbines (Francis, crossflow Banki) installed on the bypass of the bottom outlets of the Sance Dam
in the Czech Republic. Firstly, the durability of the equipment and the main types of turbine shutdowns are defined.
Empirical probability distribution curves for time to failure are plotted for both turbine sets based on available
records of incidents and reasons for individual shutdowns from 20 years of operation. The analysis shows the
average periods between shutdowns and period length probabilities between shutdowns for each turbine set.
1. INTRODUCTION
The worldwide trend is for increasing use to be made of the hydropower potential of environmental flows for
hydropower generation. The goal is to effectively exploit the potential of the flow and available hydraulic head.
Small hydropower plants (HPP) located downstream of dams are frequently used for this purpose. The main profits
are coverage of self-consumption of electric power and economic income from electricity production for the dam
owner. As hydropower is a form of so-called "green energy", electricity distributors are obliged to buy the power
output by law in the Czech Republic. One of the key questions regarding investment in new technologies or
technical upgrades is the durability and service life of a given HPP’s components, information about which is
necessary for the assessment of economic return.
In this paper the results of the durability analysis of a small hydropower plant installed on the bypass of the bottom
outlets of the Sance Dam are presented. The analysis is based on operation records for two turbines (Francis,
crossflow Banki) installed at the HPP. First, the durability of the equipment and the main types of defects and
failure modes are defined. Then, empirical probability distribution curves for time to failure are developed for both
turbines.
2. DEFINITIONS AND METHODS
Service life is generally defined as the ability of a facility (a HPP in this case) to fulfil the required functions until a
limit state is reached if prescribed maintenance and repairs has been applied (Votruba & Heřman et al. 1993). The
service life can be expressed by technical life, mean technical life or mean period of usage in years. The limit state
is usually defined by limit values of predetermined parameters which describe technical, economical or operational
properties of the facility (Votruba & Heřman et al. 1993). For the period of the facility life the condition of
efficiency has to be satisfied, i.e. the ability of the facility to attain required goals. The main aim of an HPP is the
optimal technically possible exploitation of hydropower potential at the highest economic efficiency. To achieve
this it is essential for the HPP to be operational, i.e. to be able to perform all demanded functions and keep its
operating parameters within prescribed limits. If the equipment does not comply with these conditions it means that
the HPP is in failure mode. To bring the HPP back into operation requires recovery measures such as repairs,
refurbishment or the exchange of some components.
Durability is defined as characteristics that ensure that the serviceability, safety and environmental compatibility of
a device are maintained during the intended period of use (Blockley 1992). It can be quantified by the probability
that the object (equipment, machine or civil structure) will have a relatively long continuous service life without
requiring an inordinate degree of maintenance. If the durability of the HPP is over, the service life is over too.
In this study the system approach is applied to durability assessment. The HPP is defined as a system of
components. In principle the whole system (HPP) can be divided into two main parts, namely the civil structure and
its technical equipment (also including electrical equipment). Here, the Eurocode technical standard (CSN EN
1990) recommends a “design working life” of 100 years for civil structures and 10 to 25 years for technical
equipment. Therefore, the service life of an HPP is mainly determined by the condition of its technical equipment,
and so durability analysis has to be focused on that. During the first step in the analysis, detailed operational data
have to be evaluated and failures have to be specified (Ayyub et al. 1998). It must be stated that there were
practically no relevant data for the initial operating period of the Sance HPP. The analysis was therefore carried out
for the later periods for which data were available for individual turbine sets (see below).
Hydropower'15
Stavanger, Norway 15-16 June 2015
During durability analysis, the following reasons for “forced shutdown” are taken into account:
• stoppage of a machine due to any maintenance activity (including exchange, repair, dismounting or
mounting of any part of the HPP) necessary to put it back into operation,
• stoppage of the machine due to the defect or failure of any part which called for the repair and/or exchange
of part of its equipment (typically the exchange of a runner, shaft, generator, bearings, coupling, etc.).
The analysis was carried out separately for the two above-mentioned cases, both for all shutdowns and for
"significant reasons" only. Causes of shutdowns are linked to the specific components of the HPP, such as runners,
wicket gates, bearings, couplings, etc., and also to entire turbine sets.
In this study, durability is expressed using two characteristics. These are the average "no defect" period between
two consecutive forced shutdowns and the period of operation between forced shutdowns after which the
probability of failure is significant. Both can be expressed for a specific part or element or also for the entire HPP.
The average period T between forced shutdowns is defined as:
T=
tp
n
,
(1)
where t p is the observed operational period and n is the number of shutdowns recorded within t p .
The time Δt of operation between shutdowns can be expressed as the quantile:
∆t ( p ) = F
−1
(∆t ) .
(2)
The empirical distribution function is defined as follows (Van der Vaart 1998):
Fn (∆t ) =
1 n
∑1{∆ti ≤ ∆t},
n i =1
(3)
1
.
n
(4)
where Δt is the time of operation between two consecutive shutdowns (considered as a random variable), p is the
probability of occurrence (commonly considered as 0.95 in engineering practice), F(Δt) is a cumulative distribution
function representing the probability of failure occurrence after Δt time of operation from the last shutdown. In
formula (3) the vector Δt i consists of values sorted according their magnitude.
In this study the assumption of the equal probability of occurrence of individual Δt is taken into account. The
probability of individual times of operation or standby regimes between shutdowns was estimated as follows:
p i (∆t ) =
3. DESCRIPTION OF THE SANCE DAM AND HYDROPOWER PLANT
The Sance dam is located in the north-east part of the Czech Republic. It was built on the Ostravice River between
1965 and 1969. The dam body consists of rockfill shoulders and clayey core sealing. It is equipped with an
emergency spillway, two bottom outlets, an outlet tower and a stilling basin. The valve chamber at the downstream
edge of the bottom outlets is joined with the powerhouse of a small hydro power plant (Fig. 1 and Fig. 2). The
Sance dam is 62 m high, with a crest length of 342 m and a width of 8 m.
The bottom outlets are formed by two bypass galleries drilled at the right abutment.
The emergency spillway comprises a side weir, side channel, chute and stilling basin. The chute terminates in a ski
jump and the stilling basin that also serves to attenuate the energy of the outflow from the bottom outlets, the
sanitary outflow and the outflow from the turbines (Fig. 2). The stilling basin is 42.5 m long with a gradually
changing width ranging from 10.2 to 17.2 m. The depth of the stilling basin is 3.5 m.
The ski jump at the downstream edge of the chute forms the ceiling of part of the chamber containing the bottom
outlet valves, and of the powerhouse. The adjacent small house at the right downstream toe is an extension of the
chamber housing regulating valves for the bottom outlets, a power house with two turbine sets and generator units
including control valves, valves for raw water supply and an automatically operated bypass for environmental flow
(Fig. 3).
The layout of the Sance Dam is shown in Fig. 1. A detail of the entrance to the HPP is shown in Fig. 2.
The small hydropower plant is equipped with two sets of turbines, HC1 and HC2. The inflow to the turbines is
provided by an 800 mm-diameter diversion penstock which is driven from the left bottom outlet. The penstock is
equipped with a cross-shaped distribution piece which distributes the flow to both turbines and also serves as a
junction for the backup pipeline for raw water supply, see Fig. 3.
Hydropower'15
Stavanger, Norway 15-16 June 2015
Fig. 1. View of the Sance Dam
Fig. 2. Detail of the entrance to the HPP
Turbine set HC1 was installed in 1974 and is composed of a Francis F25 horizontal turbine with a runner diameter
of 525 mm, a capacity of 1.6 m3/s and a maximum power output of 810 kW. The turbine comprises the following
components: runner (stainless steel), shaft, coupling, flywheel, spiral case, draft tube, wicket gate, bearings,
regulating mechanism, brake and hydraulic aggregate. The asynchronous generator is connected to the turbine
through the coupling. HC1 is equipped with electrical protection devices, electrical distribution boards, an oil
transformer and an automatic control system.
Turbine set HC2 was added to the Sance HPP in 1992 and consists of a type CINK 3.4Bx312 Banki turbine with a
runner diameter of 340 mm, an effective width of 288 mm, a discharge capacity of 0.624 m3/s and a maximum
output of 243 kW. The turbine comprises the following components: turbine chamber, runner, shaft, coupling with
cover, frame, draft tube, bearings, regulating mechanism and transition inlet part. A horizontal asynchronous
generator is connected to the turbine through the coupling. HC2 is also equipped with electrical protection devices,
a transformer and an automatic control system. The gross head to the HPP is approx. 60 m.
Hydropower'15
Stavanger, Norway 15-16 June 2015
Fig. 3. Ground plan of HPP
Fig. 4. View of turbines HC1 Francis (left) and HC2 Banki (right)
4. DURABILITY ANALYSIS OF THE SANCE HPP
The service life assessment has been carried out separately for the structural part (powerhouse) and the technical
equipment.
The structural part of the Sance HPP consists of the outflow structure, the bottom part of the HPP and the
powerhouse. These structures are made of both massive concrete (the bottom part of the HPP) and reinforced
concrete (the frame of the powerhouse). Visual inspection of the civil works has confirmed they are in good
condition. The age of these structures is currently approximately 45 years. Even though the design service life of
the structures is expected to be about 80 years, practical experience shows that the real service life of concrete,
masonry and also wooden structures is usually longer, especially if they are properly maintained. No significant
incident or defect has been recorded for the structural part of the Sance HPP to date, and thus the durability analysis
was performed for the technical part of the HPP only.
Hydropower'15
Stavanger, Norway 15-16 June 2015
For the durability analysis each turbine set represents a system consisting of intake and outflow parts, the turbine
itself, electrotechnical components and the control system. To perform the durability analysis, a detailed review of
the operational history of the HPP was performed. Unfortunately, in the case of HC1 only limited data are available
from the period between 1974 and 1992. HC1 was completely renovated between 1995 and 1996. The analysis for
HC1 therefore covers the period of operation since 1996. HC2 was installed in 1992 and put into operation in 1994.
However, operation records are available for the period since 1997 so the analysis covers the operation of HC2
since that year. An overview of failures, repairs and the partial renovation of the HPP was compiled for each
turbine set based on available operation and service records.
4.1 The average period between shutdowns for significant reasons
The average periods between two consecutive shutdowns for significant reason at each turbine (HC1 and HC2) are
linked to the individual main components as well as the whole equipment (see Tab. 1). The results were obtained
via the evaluation of operation records for the period 1996 to 2012 for HC1 (Francis) and 1997 to 2012 for HC2
(Banki) according to formula (1). The results are shown in Tab. 1.
Tab. 1 Average periods between shutdowns for significant reasons (HC1 and HC2)
Components
Runner and/or wicket gate (chamber)
Bearings
Coupling
Generator
Whole equipment
Period between shutdowns for significant reasons [years]
HC1 Francis
HC2 Banki
2.6
2.5
4.5
2.1
6.0
15.0
9.0
7.5
1.13
0.94
The HC1 turbine set (Francis) has recently become rather obsolete, corresponding as it does to turbine design
trends common for the period of its installation (1974). The turbine is kept in operation by periodic maintenance
and service activities. Within the evaluated period of operation several significant repairs were carried out on this
set. These mainly consisted of general repair work on the turbine (1995), the exchange of bearings (2002, 2004),
the exchange of the generator, as well as repairs and the exchange of the pumping aggregate for a wicket gate
(1998, 1999), the modernisation of the control system, etc.
The HC2 turbine set (Banki) requires quite frequent service activities to be capable of operation. The main reason
is the quite high load on the runner, which is causing cavitation (Fig. 5). Based on the operation records it can be
concluded that on average the complete exchange of the runner is necessary approximately every 12 years, but
significant servicing and maintenance (the exchange of some parts, welding, balancing, etc.) is necessary
approximately every 7 to 8 years. The period between minor repairs is shorter: up to 2.5 years on average.
Fig. 5 Cavitation on the HC2 runner (source: M. Kozelsky, Povodí Odry, s.p.)
Hydropower'15
Stavanger, Norway 15-16 June 2015
The period between two consecutive shutdowns for significant reasons lasts on average 1.13 years for HC1 and
0.94 years for HC2. Such periods according to current demands on similar technology can be regarded as
inappropriate. The quite short intervals between shutdowns for significant reasons indicate the low durability of
both turbines. The producers generally state their turbines’ service lives as being in the order of decades without
significant maintenance.
4.2 Period of operation between all shutdowns
Empirical distribution functions describing the probability of all shutdowns and also shutdowns due to significant
reasons as a function of the operating time between shutdowns were developed according to equation (3) for both
turbine sets (HC1 Francis and HC2 Banki-crossflow) based on detailed records for the period from 2004 to 2013
(see Fig. 6). The most important source of data for the calculation was operation records which mention the reasons
for the shutdowns of both turbines. The events are indicated as “breakdown, repair, exchange, dismount” and
represent shutdowns which have to be fixed at least by an operator. These events last from tens of minutes up to
several months. If the service staff has to exchange any part of the equipment the case is recognized as a
"significant reason" for shutdown (Fig. 6).
Fig. 6. Distribution functions describing the probability of shutdowns in relation to the time of operation
between shutdowns
Based on the results shown in Fig. 6 it can be concluded that for a high exceedance probability (95 %) a forced
shutdown should be expected after approximately 160 days (0.4 years) of operation from the last repair at HC1, and
about every 250 days (0.7 years) at HC2, while a shutdown for a significant reason (a major defect) should be
expected after approximately every 1300 days (3.5 years) at HC1 and after approximately every 660 days (1.8
years) at HC2.
CONCLUSIONS
In this paper a durability analysis was carried out for two turbine sets located at the Sance HPP using operation
records. Unfortunately, practically no records exist for the Francis turbine during its first 19 years of operation. The
following decade also provides only poor information about the performance of both turbine sets. The analysis
therefore provides relevant results only for the period of the last 17 years of operation. However, the contribution of
the presented method is that it can act as a guide in the durability analysis of any HPP.
The results of the durability analysis show there are quite short periods between forced shutdowns, which in
general is not a favourable state of affairs for the HPP owner. Shutdowns for significant reasons that call for the
exchange of parts of equipment may occur practically every year, on average (Tab. 1). Probability analysis
indicates that common forced shutdowns can be expected with relatively high probability (70 - 80%) within about
100 days of a previous repair, though one may even happen much sooner. The quite short intervals between
shutdowns for significant reasons indicate the low durability of both turbines. Although the standard periodic
maintenance have been performed the current technical state of the Sance HPP enables further operation only if
Hydropower'15
Stavanger, Norway 15-16 June 2015
quite frequent repairs are performed on the technical equipment. Both turbine sets (Francis and Banki) have
recently become rather obsolete. HC1 (Francis) corresponds with turbine design trends common for the period of
its installation (1974). The installation of a Banki turbine for a head of close to 60 m causes significant detrition to
the machine.
A further step in the analysis would be the assessment of the economical efficiency of possible variants for future
remedial works at the Sance HPP.
ACKNOWLEDGEMENT
This study is the result of projects entitled The utilization of probabilistic methods for safety survailance of dams
with respect to its safety during global climate change, project code TA04020670, and Advanced Materials,
Structures and Technologies, project code LO1408 AdMaS UP.
REFERENCES
Ayyub, B. M., Kaminskiy, M. P., Moser, A. D., 1998. Reliability analysis and assessment of hydropower
equipment, IWR Report 98-R-6. U.S. Army Corps of Engineers, 75p.
Blockley, D. I. (ed.), 1992. Engineering Safety. McGraw Hill, 1992. 494 p.
ČSN EN 1990 ed. 2, 2011. Eurocode: Basis of structural design. 97 p.
Van der Vaart, A.W. (1998). Asymptotic statistics. Cambridge University Press. p. 265.
Votruba, L., Heřman, J. et al., 1993. The reliability of hydraulic structures. Czech Technical Society, Vol. XCIX
1993, 488 p. (In Czech).