ASSESSMENT OF THE STABILITY CONDITIONS OF A
CISTERCIAN CLOISTER
P.B. Lourenço, G. Vasconcelos, L. Ramos
University of Minho, Department of Civil Engineering, Guimarães, Portugal
ABSTRACT
This paper describes the evaluation followed to define the stability conditions of a
Cistercian cloister in Portugal. The adopted methodology of diagnosis results in
clear advantages in terms of safeguard of the authenticity of the construction and,
consequently, of its heritage and historical value.
1. INTRODUCTION
Preservation of the architectural heritage is considered a fundamental issue in the
cultural life of modern societies. In recent years, large investments have been
concentrated in this area, leading to developments in the areas of inspection, nondestructive testing, monitoring and structural analysis of monuments.
Nevertheless, understanding, analyzing and repairing historical constructions
remains one of the most significant challenges to the modern technicians.
Up to recently, it was normal to carry out any intervention without spending
proper resources (time and money) in studying the reasons why the intervention
was required. This lead often to either heavy interventions, which result in a
decrease of the historical value of the construction, or inadequate interventions,
which did not solve the problem that require the intervention. Fortunately, this
attitude has changed as more and more attention is being devoted to the tasks of
inspection and diagnosis. Indeed, the costs involved in the inspections and
diagnoses are easily recovered in the intervention itself.
In this paper, an example of a comprehensive diagnosis of a historical
construction is presented. The survey and characterization of the damage, together
with in situ and laboratory testing, allowed to gather the information necessary to
establish the need of an intervention and to bound this intervention. Additionally,
the obtained information allowed to analyze the construction in a computer, which
resulted in clear information on its structural behavior.
2. HISTORICAL FRAMEWORK
The Monastery and Church of Salzedas are located in Salzedas, Tarouca and the
church was recently classified as National Monument of Portugal. The church is
essentially set in a urban environment, whereas the monastery is set in a more
rural environment, see Figure 1a. The plan dimensions are very considerable, 75 ×
101 m2, making it the second largest monastery in Portugal, after the World
Heritage Monument – Monastery of Alcobaça. The monastery and church possess
a longitudinal irregular plan with different volumes, typical of a Cistercian Abbey,
see Figure 1b. The present study focus exclusively in the cloister dated from the
17th century (D – Large cloister, in the picture).
N
A
E
D
C
B
A – Church
B – Sacristy
C – Chapter room
D – Large cloister
E – Small cloister
(a)
(b)
Figure 1 – Monastery and Church of Salzedas: (a) aerial view and (b) plan
The large cloister is regular and substitutes part of the primitive cloisters, see
Figure 2a. It possesses crossed vaults in the 1st level and canon vaults in the 2nd
level. The walls, brackets and ribs are made of granite and the vaults are made of
brick masonry with clay filling. The Monastery is not own by the state and,
therefore, only limited interventions were carried out.
After repeated statements of the pre-collapse status of the cloister [1,2], the
General Directorate for National Buildings and Monuments (DGEMN) started a
restoration plan that was carried out in 1980/1981 and 1983, including the
following works, see Figure 2b-d:
• The vault of the 2nd level of the West wing was demolished, being
replaced by a reinforced concrete vault;
• The wall separating the large and small cloisters, between the 1st and
the 2nd levels was, dismounted and remounted plumb.
These interventions must be valued as, in their absence, it is likely that collapse
would prevail with a larger loss of the architectural heritage. However, it is noted
that the works do not comply with modern theories of intervention in historical
structures and, today, would be very debatable.
(a)
(b)
(c)
(d)
Figure 2 – Aspects of the cloister: (a) present view, (b) demolition of the canon
vault, (c) dismounting of the external wall between the large cloister and
the small cloister and (d) reconstruction of a reinforced concrete vault
3. SIGNS OF DAMAGE
The condition of the cloister is quite poor, including biological colonization
sometimes associated with moisture stains, deterioration of the bricks in the
vaults, cracks with variable thickness, crushing of stones and excessive
movements in walls and vaults.
3.1. Observed Cracks
The largest and widest set of cracks occurs in the canon vaults of the South and
East wings of the 2nd level, as well as the SE and NW corners, see Figure 3a. The
cracks occurs mostly in the longitudinal direction, even if some transversal cracks
also occur, up to a crack opening of 40 mm. The vaults of the first level exhibit
also cracks, in the South and West wings, up to a crack opening of 15 mm, see
Figure 3b. The West wing is supported on temporary wooden poles.
(a)
(b)
nd
Figure 3 – Mapping of cracks for the (a) 2 level and (b) 1st level
The walls exhibit widespread cracking at the 2nd level and almost no cracking
at the 1st level. With the exception of a few localized areas, cracking is minor
(crack openings in the range of 1 to 5 mm).
3.2. Observed Displacements
Vertical displacements up to 35 mm were measured at the key of the crossed
vaults of the 1st level. But all the walls of the cloisters exhibit large horizontal
movements that lead to the separation between the vaults and the walls, in a clear
lack of verticality, see Figure 4. The out-of-plumb displacement of the internal
walls reaches values of 0.18 m, 0.14 m, 0.09 m and 0.07 m in the wings West,
South, East and North, respectively.
(a)
(b)
(c)
(d)
Figure 4 – Aspects of the wall movements: (a) separation between wall and vault
(intrados of 1st level), (b) detail, (c) separation between wall and vault (extrados of
2nd level) and (d) horizontal movements in the direction of the court
3.3. Other Damage
The brackets supporting the crossed vaults of the first level show signs of
compressive crushing, particularly in the West wing, see Figure 5a. This can be
explained by the tilting movement of the walls. The absence of connection
between the infill of the crossed vaults and the walls resulted in a very localized
area to transfer the load, i.e. only the brackets. Also, a significant number of
bricks show deterioration, particularly around the cracked areas, see Figure 5b.
This occurs in both levels and can be explained by frost-thaw cycles and water
infiltration, as the amount of rainfall per year in the region is high and the
temperatures in the winter are excellent for ice formation (daily cycles with ±0º).
(a)
(b)
(c)
Figure 5 – Aspects of the mechanical deterioration: (a) crushing of the stone
brackets of the 1st level and (b,c) decohesion of the bricks
Other perturbing signs, less relevant from the structural point of view, include
damage of the stone due to freeze-thaw cycles and overall biological colonization,
see Figure 6.
(a)
(b)
(c)
Figure 6 – Other aspects of the materials deterioration: (a) degradation of stone
carving and (b,c) biological colonization
4. IN-SITU SURVEY AND LABORATORY INVESTIGATION
In order to characterize the construction, to justify the observed damage and to
define corrective measures, an experimental in-situ and laboratory testing program
was carried out.
4.1. Soil and Foundation Survey
This survey consisted of seven borings and three pits to allow the mechanical and
physical characterization of the soil and foundations. It was possible to define a
layered soil consisting of embankment of clayey nature (1.1 m), organic soil (0.30
m), alluvial soil with medium large stones, naturally wounded and worn by the
action of water (0.60), alluvial soil with pebble (0.50 m). Between 2.5 and 2.7 m
depth, the soil is granular with some clay and below 2.7 m depth large stones with
0.30 m to 0.40 m are found, see Figure 7.
Pit 3
Medieval clay tile pavement
(original soil level)
Medieval
foundation with
granite ashlars
XVII foundation
(irregular
masonry bonded
with clay)
Pit 2
Embankment
Organic soil
Alluvial soil with stones
Alluvial stones with pebble
Granular soil with clay
Large stones
Figure 7 – Detail of the soil survey
The foundation soil exhibits moderated resistance and large heterogeneity for
depths between 1.0 and 1.8 m, were supposedly all the cloister foundations are
set. The foundations for the walls seem to be medieval and of good quality, but
the foundations of the cloister columns are inadequate. These irregular masonry
foundations are unable to distribute the loads over a significant soil area and the
foundations depth around 1.0 m seems to indicate that the foundations were built
on top of the original pavement level, directly on organic soil.
4.2. Visual Inspection with Rigid Endoscope
In order to characterize the inner constitution of vaults and walls, a few bore holes
and several cracks were inspected with a rigid endoscope, see Figure 8. The
inspection allowed several conclusions, among which:
• The vaults are made with clay brick masonry with 0.22 m thickness and
clay filling. Separation between the two materials was not found;
• The walls are made with large granite stones, with dry joints or a thin clay
joint. Clay seems to be washed out in most cases due to weathering. An
internal core of weaker mechanical characteristics was not found;
• Internal longitudinal cracks that would compromise the stability of the
walls under vertical loading were not found.
As a result of the inspection with the rigid endoscope, it was concluded that the
granite walls of the cloister are adequate and there is no danger of collapse due to
decohesion under vertical loading. Coring and other techniques to estimate the
strength of the walls were considered not necessary and it was decided to carry
out a simple flat-jack stress control.
(a)
(b)
(c)
Figure 8 – Aspects of the inspection with rigid endoscope: (a) drilling machine
(b) visual inspection and (c) photograph
4.3. Flat-jack Testing
Flat-jack testing was carried out in the spirit of the norms ASTM C1196 and
C1197, even if the selection of a place to execute the test was difficult due to the
irregular nature of masonry. To open the slot, both a sawing machine and a
drilling machine were used, see Figure 9a,b. A rectangular flat-jack with 406 ×
102 × 4.2 mm3 and shims were adjusted with a pressure of 3.0 MPa. Afterwards,
two tests to measure the in situ vertical stress were carried out, utilizing the three
pairs of marks illustrated in Figure 9c and denoted as row 1, row 2 and row 3.
(a)
(b)
Figure 9 – Flat-jack testing: (a) Sawing the slot, (b) stitch-drilling the slot and
(c) measuring operation
The results of one of the tests is represented in Figure 10 but the two tests
results were very similar. It can be observed that the row 2 and row 3 are more
adjusted with the slot, whereas row 1 shows a clear lower stiffness. The average
in-situ stress obtained with the latter rows is 1.2 MPa. The value expected with a
simple numerical analysis, is around 0.6 MPa. The 50% difference between these
values might be due to the irregular shape of the masonry blocks, indicating that
only half of the stone block is inactive. It is stressed that the value of 1.2 MPa
obtained can be considered as relatively low for the particular type of masonry.
0.200
0.150
0.100
Displacement
[mm]
0.050
Stress [MPa]
0.000
-0.050
1.0
1.2
1.4
1.6
1.8
2.0
Row
Fiada 11
Row
Fiada 22
Row
Fiada 33
2.2
-0.100
-0.150
Figure 10 – Stress-relative displacements results of flat-jack testing
4.4. Coring
In order to confirm the internal constitution of the vaults and to characterize the
mechanical behavior of the brick masonry, three φ75 mm cores were extracted
from the vault, see Figure 11. The cores confirmed the borehole observation.
(a)
(b)
Figure 11 – Samples for mechanical testing: (a) coring and (b) sample
4.5. Chemical, physical and mechanical characterization
The plaster, vault infill and mortar from the brick masonry of the vaults were
characterized with X-ray diffraction, non-soluble residual and burn loss tests, see
Figure 12a. The bricks were characterized with absorption tests and uniaxial
compression tests, see Figure 12b,c,d. The mortar from the brick masonry was
also characterized with uniaxial compression tests, see Figure 12e. The
representative samples were extracted from the construction or the cored samples.
The tests indicated the composition of the plaster and mortar (1:3 in volume)
and the composition of the vault infill (clay). The bricks are of low quality and
non-durable, with an absorption in cold water around 20% and a volume mass of
1560 kN/m3.
The uniaxial compression tests were carried out in samples of 45 × 45 × 45 mm3,
tested with greased Teflon layers to avoid the plate confining effect. The obtained
Young’s modulus and strength for the bricks were Eb = 7.3 GPa and fb = 5.2 MPa,
respectively. These values are quite low and confirm the poor quality of the
bricks. The obtained Young’s modulus and strength for the mortar were
Em = 8.6 MPa and fm = 3.8 MPa, respectively. This strength value can be
considered normal for the mortar composition.
With these results it is possible to estimate the strength of the brick masonry as
f k = K × f b0.65 × f m0.25 = 0.60 × 5.2 0.65 × 3.8 0.25 = 2.4 MPa ,
(2)
according to Eurocode 6. The Young’s modulus of the composite might be
obtained from homogenization procedures as
E=
t m + tu
0.02 + 0.04
×ρ =
× 0.5 = 3.8 GPa ,
tm
tu
0.02 0.04
+
+
8.6
7.3
E m Eu
(3)
where tm represents the thickness of the mortar, tu represents the height of the
brick and ρ represents an efficiency factor associated with the deficient bond
between the two materials (assumed equal to 0.5).
The above values allowed to perform a numerical analysis of the structure [3].
1900
Mass (g)
1800
1700
Sample
Provete 11
1600
Sample
Provete 22
(a)
336
240
192
144
24
1500
Time (h)
(b)
14000
Force (N)
12000
T2
T3
10000
8000
T1
T1
6000
4000
2000
Displacement (µm)
0
0
200
400
600
800
1000
1200
1400
1600
1800
2000
(c)
(d)
(e)
Figure 12 – Examples of chemical, physical and mechanical testing: (a) X-ray
diffraction test on vault infill material, (b) absorption test on brick,
(c) compressive failure of brick, (d) uniaxial compression test on
brick (two samples) and (e) compressive failure of mortar
5. RECOMMENDATIONS
The conclusions of the numerical analyses of [3] together with the inspection
described in the previous section allowed to conclude that: (a) the non-symmetry
between the internal and external walls of the cloister result in movements in the
direction of the court, as observed in the construction; (b) a linear elastic analysis
of the construction results in very limited displacements (in the order of the
millimeter) and moderate stress values (maximum tensile stress of +0.25 MPa and
maximum compressive stress of –0.6 MPa). The large displacements observed in
the construction require a geometrical and physical non-linear analysis; (c) in
order to obtain horizontal displacements in the order of the observed in the
structure, it is necessary to consider the soil-structure interaction. It seems
therefore that the foundations play a key-role in the observed damage; (d) the high
movements recorded in the construction and the deterioration of the brick vaults
indicate that the safety level of the structure is not compatible with any use and
immediate intervention is necessary.
With the study described it was possible to recommend the following
intervention, see details in [3]: (a) Immediate (Winter of 2000) – Covering of the
roof to stop water infiltration and further damage in the brick vaults; temporary
shoring of the SE corner and the vaults in the South wing; (b) Short Term –
Strengthen vaults in the 2nd level (all wings) and vaults in the first level (South
wing); strengthen foundations of the columns in the internal walls of the cloister.
6. CONCLUSIONS
A comprehensive program of inspection and diagnosis of a Cistercian cloister is
presented. The results of the inspection, together of a numerical simulation of the
construction, allowed to understand the damage of the construction, to assess its
safety and to propose a set of corrective measures.
It was possible to demonstrate the value of the study to define a strengthening
project that minimizes the intervention itself and maximizes the performance of
the intervention. This strengthening project is currently under execution.
REFERENCES
1. Leitão, A., 1963, The monastery of Salzedas (in Portuguese)
2. Cocheril, M., 1978, Guide of the Cistercian Abbeys in Portugal (in French),
Calouste Gulbenkian Foundation, Lisbon
3. Lourenço, P.B., Vasconcelos, G., Martins, J.B., Ramos, L., Jalali, S., 2000,
Diagnosis of the stability conditions of the XVII century cloister of the
Monastery of Salzedas, in Tarouca (in Portuguese), Report nº LEC 31/2000,
University of Minho, pp. 200