Genetic-based approach to achieve higher durability for stone repairs A. Isebaert + L. Van Parys

User menu

Since natural stone is the main construction material for built heritage, its alteration implies a challenge for conservation institutions. Repair of original stone can be performed with replacement stone but also with repair mortars. In both cases, the replacement material has to be compatible in short-term and long-term, to avoid damage or different alteration. In addition to mineralogical considerations, mortar compatibility can be determined through their properties (mechanical, physical) and their appearance. Creating a durable repair mortar by taking into account these parameters can be achieved through mathematical optimization, since genetic algorithms are suited to solve such a multi-objective problem [1].

Natural stones weather due to a combination of processes that catalyst or slow down one another [2,3,4,5,6,7]. This creates complex ever-evolving alteration patterns [8,9]. Part of these processes is due to extrinsic environmental factors, while other processes are caused by the intrinsic characteristics of the stone.
The extrinsic factors include climate, location of the monument (city vs. countryside), the orientation (protected or exposed to rainfall) and the location of the stone in the building (ground-level, protected or protruding) [10]. Water in general is an important catalyst for degradation, since it can transport harmful materials into the stone and cause stress into the stone when it freezes.
But in addition to these more natural causes of deterioration, other causes have appeared since the 19th century, such as pollution of the environment. Weathering due to this and other human interference is sometimes greater than natural weathering [7]. The famous black crusts for example, are caused by a chemical reaction of
deposition of sulphur oxides (combustion) and calcite (stone), creating gypsum (CaSO4. ½ H2O). In the porous gypsum, dirt particles such as carbon settle in, blackening the crust. Regardless of the fact that the specific amount of sulphur dioxide in the atmosphere is decreasing, pollution remains an important cause for degradation [11,12].
Additionally, there are also intrinsic causes for degradation. The composition and properties of the stone determine, for a large part, possible alterations [7,13]. Stones in general were formed through processes in the crust of the earth in a specific environment up until a stable condition was reached. Their exposure to the atmosphere brings them in another environment, leaving them in se in a physic-chemical disequilibrium [7]. Thus, the minerals and matrix of the stone are altered in composition or condition when exposed to the environment, e.g. they oxidise or dissolve [13,14,15]. Secondly, some stones in one and the same wall seem to be more ‘durable’ than others. This difference can be explained by the simple fact that stones have a spatial variability and heterogeneity [14]. As third aspect, the orientation of the rock fabric in the monument is also important to take into account, e.g. the orientation is partially responsible for the compression and flexural capabilities of the stone [14,16,17]. Finally, the appearance and finishing of the stone play a role in the alteration susceptibility as well: rough stones, with a higher surface area, allow more dirt and other particles to attach.

Repair mortar compatibility problems
After determining the deterioration patterns and their causes, restoration interventions are set up to preserve the building. One of the possible interventions is the repair of original stones (through replacement stones or repair mortars). However, the choice for an adequate repair material is of vital importance. Researches have pointed out severe deterioration of stones due to the inadequate use of repair mortars [18,19] or replacement stones [20,21,22].
An ideal repair mortar for natural stone should be durable enough, but self-sacrificing in the long run [23,24]. However, with variable deterioration processes and a heterogeneous material to start with, this isn’t easy to achieve. In the past, incompatible repair mortars were used, causing direct and indirect deterioration. Key aspects associated with the development of an “ideal” mortar are shortlisted here:

1. Perceptive aspects
Repair mortars have to be compatible with the stone in colour and surface roughness, so as not to disturb the general perception of the building. The main problem in terms of perception is designing a mortar with the right colour (in fresh and weathered state).

2. Physical and mechanical aspects
- Water transfer properties. Water plays an important role in the deterioration process; causing stress cracks and fractures, favour biological colonization, erode, dissolve and transport material. Consequently, the water transfer properties of the mortar are important for the durability of the stone.
- Modulus of resistance and/or elasticity. They can be significant in certain cases, when the repair mortar is used for the filling-in of large or structural parts taking into account the form of the filled in part and how it fits into the whole [25]. The rigidity of the material plays an important role and deterioration such as loss or cracks can occur to both repair mortar and stone.
- Thermal response. High temperature differences on sun-faced walls cause the minerals in stones to expand and contract. Consequent cycles of temperature differences cause internal stress, leading to detachment, deformation or cracking of mortar or stone [25].

3. Chemical aspects
Some repair mortars contain, create or attract harmful materials in the stone, such as salts that crystallize inside or outside the stone [25]. Deterioration patterns vary from efflorescence to cracks and scaling due to sub-florescence. Repair mortars can also be made with organic polymers, which are susceptible for biological organisms. Their presence will therefore speed up biological colonization on the stone [13].

A grand challenge lies in the perceptive compatibility (on the long run), and a large part of the compatibility problem lies in the physical-mechanical aspects where several objectives have to be considered simultaneously. Adherence is an emblematic topic where initial capabilities should be sufficient while water transfer or thermal response shouldn’t imply problems like freeze, crystallization or interface shearing in due course, the relative mechanical behaviour of stone and mortars playing there a fundamental role [26]. However, the care dedicated to mortar formulation should never overshadow the mastery of application. Several cases prove the necessity to use non-corrosive metal where dowels are needed and to finish the mortar as such that allows water to run off easily [27].

Mortar formulation through optimization
A repair mortar should be adapted to each stone specifically, but the theoretical ideal mortar isn’t achievable through current methods , literature pointing out three main paths towards an acceptable compatible repair mortar. The first one relies on commercial mixes for given types of stones: the task consists in pointing out which product could be used or adapted to meet the required standards [28,29,30,31,32,33]. The second approach builds up the mortar from scratch, aiming at repair mortar composition which is as close to the stone as possible: the same mineralogical composition is sought for and the binder is adapted to the demands of the stone [34,35,36,37]. In practice, tuning operations required by both these approaches are expensive in time and resources and justify the more recent third approach that tries to develop a modular system, where known base ingredients lead to the composition of a repair mortar for a specific stone [38]. This opens the way to computer-aided decision, but it is also merely valid for the researched components.
The current research starts from the same approach but takes benefit of mathematical tools. The perceptive, mechanical and/or physical properties of the desired mortar are linked with data gathered from the components, the system remains open: once the general framework has been developed, the method can be applied by any end-user who feeds the problem with data collected on the materials he intends to use.
Genetic algorithms are used for multi-objective problems. They rely on the natural selection principle. The Elitist Non-Dominated Sorting Genetic Algorithm (NSGA-II) allows multi-objective optimization. To obtain a mortar whose properties are close to the set targets, this intuitive approach can be used: many mortar formulations are the potential solutions x = [x1, x2,…, xn]T of the problem, expressed through n design variables that are the types and relative proportions of mortar constituents (type and quantity of sands, cement and water). A function is defined for each aspect the formulation takes into account (e.g. colour function and strength function), and this function will express the gap between the targeted value and the value corresponding to each individual. This implies one should already have an idea of the desired value for each of the variables (i.e. given proportion of sands, cement and water). Therefore, general mixing laws, based on available theories, are established which guarantee the universality of the method [39,40]. The tool will then create an initial population by a randomized set of values for each variable, with N individuals. This population of mortar recipes will then evolve, through a selection and replacement process, keeping the best individual. Then, crossing over and mutation processes combine the genes of these selected individuals. Through the evaluation process, the value of objective functions for each individual is calculated, considering simultaneously the whole of objectives (e.g. minimizing the gap between the strength/colour value of an individual and the targeted strength/colour). Through these evolutions and under certain conditions mainly associated with the initial population, it has been proven that the population converges towards an optimal formulation meeting the constraints and approaching as close as possible to the targets.

1. Deb K., 2001, Multi-Objective Optimization Using Evolutionary Algorithms, Wiley

Stone weathering problems
2. Přikryl R., Smith B.J. (eds), 2007, Building stone decay: Fom diagnosis to conservation, Geological Society Special Publications 271, London
3. Kourkoulis S.K. (ed.), 2006, Fracture and Failure of Natural Building Stones: application in the restoration of ancient monuments, Springer, Dordrecht.
4. Nicholson D.T. & Nicholson F.H., 2000, Physical deterioration of sedimentary rocks subjected to experimental freeze-thaw weathering. In: Earth surface processes and Landforms 25, 1295-1307
5. Rodriguez-Navarro C., Doehne E., 1999, Salt weathering: influence of evaporation rate supersaturation and crystallization pattern, In: Earth surface processes and Landforms 24, 191-209
6. Winkler E.M., 1994, Stone in Architecture: Properties, Durability, 3rd edn., Springer Verlag, Berlin
7. Amoroso G. & Fassina V., 1983, Materials Science Monographs 11 Stone decay and conservation, Elsevier
8. De Kock T. et al, 2012, Multi-disciplinary characterization of gypsum crust on lede stone, Proceedings 12th International Congress on the Deterioration and Conservation of Stone, 21-24 September 2012, New York
9. Papida S. et al, 2012, Enhancement of physical weathering of building stones by microbial populations, International Biodeterioration & Biodegradation 46: 305-317
10. André M-F. et al, 2012, ‘Quantitative assessment of post-restoration accelerated stone decay due to compatibility problems (St Sebastian’s abbey church, Manglieu, French Massif Central)’, Proceedings 12th International Congress on the Deterioration and Conservation of Stone, 21-24 September 2012, New York
11. de la Fuente D. et al., Mapping air pollution effects on atmospheric degradation of cultural heritage, Journal of Cultural Heritage (2012),
12. Kucera V. (coord), 2005, EU 5FP RTD Project Model for multi-pollutant impact and assessment of threshold levels for cultural heritage, Final report
13. Warscheid Th., Braams J., 2000, Biodeterioration of stone: a review, International Biodeterioration & Biodegradation 46: 343-368
14. Graue B., et al., 2012, Requirements for replacement stones at the Cologne Cathedral – A systematic approach to general criteria on compatibility, Proceedings 12th International Congress on the Deterioration and Conservation of Stone, 21-24 September 2012, New York
15. Moropoulou A. et al., 2006, Susceptibility of building stones to environmental loads: evaluation, performance, repair strategies, In: Fracture and Failure of Natural Building stones, S.K. Kourkoulis (ed.), 291-297
16. Belayachi N. & Hoxha D., 2012, On the effects of the spatial variability on the THM behavior of tuffeau stone of historical buildings, Proceedings 12th International Congress on the Deterioration and Conservation of Stone, 21-24 September 2012, New York
17. Warke P.A., McKinley J., Smith BJ, 2006, Weathering of building stone: approaches to assessment, prediction and modelling, In: Fracture and Failure of Natural Building stones, S.K. Kourkoulis (ed.) 313-327

Repair mortar compatibility problems
18. Dotter K.R. et al, 2009, Sacrifice and rebirth: the history of lime mortar in the north of Ireland, Proceedings of the 3rd Int. congress on Construction History, Cottbus, May 2009, 499-506
19. Pavia S, Bolton J (eds.), 2001, Stone monument decay study 2000. An Chomhairle Oidhreachta/ The Heritage Council
20. Quist WJ 2009, Replacement of natural stone in conservation of historic buildings – evaluation of replacement of natural stone at the church of our lady in Breda’, Heron 54(4): 251-278
21. Přikryl R., Smith B.J. (eds), 2007, Building stone decay: Fom diagnosis to conservation, Geological Society Special Publications 271, London
22. Blanc A, 1992, la restauration des monuments, in: terroirs et monuments de france, Pomerol C (eds.), 365-367, Orléans: Editions du BRGM
23. ICOMOS, 1964, The Venice Charter. International charter for the conservation and restoration of monuments and sites. Second international congress of architects and technicians of historic monuments, (14/01/2013)
24. ICOMOS, 1994, The Nara Document on Authenticity, (14/01/2013)
25. Hayen R., 2010, Herstelmortels breed uitgesmeerd, Versteend erfgoed – omgaan met herstelmortel en kunststeen. Studiedag 28 mei 2010, KIK/IRPA, Brussels, 4-19
26. Vanderauwera M. 2010. ‘Grenzen aan herstelmortels’. Versteend erfgoed – omgaan met herstelmortel en kunststeen. Studiedag 28 mei 2010. Brussel: KIK/IRPA, 29-32
27. Schwengelbeck O., 2002, Steinersatzmasse Naturstein, Bauhandwerk 6: 50-52

Mortar formulation through optimization
28. Torney C., et al., 2012, „Plastic“ repair of natural stone in Scotland: perceptions and practice, Structural Survey 30, 4: 297-311
29. Steenmeijer R., 2010, Kunststeen of natuursteen. Voorbeelden uit de praktijk, Versteend erfgoed – omgaan met herstelmortel en kunststeen. Studiedag 28 mei 2010, KIK/IRPA, Brussels, 33-38
30. Bromblet P., et al., 2005, Approach for compatible mortars for restoration purposes: Stone repairs of the roman amphitheatre of Arles (France), International RILEM Workshop on Repair Mortars for Historic Masonry. Delft The Netherlands, 26th -28th January 2005, 73-81
31. Marie-Victoire E., Bromblet P., 1999, A new generation of cement-based renderings : an alternative to traditional lime based mortars ?, Bartos P., Groot C., Hughes J.J. (eds.), International RILEM Workshop on Historic Mortars: Characteristics and Tests. Paisley, Scotland, 371-393
32. Szemerey-Kiss B., Török A., 2011, Time-dependent changes in the strength of repair mortar used in the loss compensation of stone, Environmental Earth Sciences 63:1613-1621
33. Blauer C., et al., 2010, Repair mortars for the sandstones of the cathedral of Berne, Valek J., Groot C., Hughes J.J. (eds.), 2nd Historic Mortars Conference HMC 2010 and RILEM TC 203-RHM Final Workshop. 22-24 September 2010, Prague, Czech Republic, 909-916
34. De Kock T., et al., 2012, Compatibility study and adaptation of stone repair mortars for the Lede stone (Belgium), Geophysical research analysis 14, EGU2012-9550-1
35. Stefanis N.A., Theoulakis P., 2010, Design of conservation mortars for the restoration of the Piraeus stones at the monuments of the Acropolis of Athens, Valek J., Groot C., Hughes J.J. (eds.), 2nd Historic Mortars Conference HMC 2010 and RILEM TC 203-RHM Final Workshop. 22-24 September 2010, Prague, Czech Republic, 1181-1188
36. Beck K., et al., 2008, Critères de compatibilité entre des mortiers à base de chaux et des pierres calcaires a forte porosité (tuffeaux blancs), Actes des 25e Rencontres de l’AUGC, 23-25 mai 2007, Bordeaux, (20/01/2013)
37. Bromblet P., 2000, Evaluation of the durability and compatibility of traditional repair lime-based mortars on three limestones, Internationale Zeitschrift für Bauinstandsetzen und Baudenkmalpflege 6, vol. 5: 513-528.
38. Ramge P., Kühne H.C., 2012, Development of repair mortars fort he restoration of natural stone in cultural heritage, Alexander et al. (eds.), Concrete Repair, Rehabilitation and Retrofitting III, London: Taylor & Francis Group, 882-887
39. De Larrard F., Sedran T., 1999, Une nouvelle approche de la formulation des bétons, Nantes : Laboratoire Central des Ponts et Chaussées
40. Leoskool L., et al., 2012, Compatible mortar for masonry restorations : discrete optimization for equivalent strength and coour prescription, Proceedings ICDS-12 Int. Conference Durable structures : from construction to rehabilitation, Lisbon, Portugal
41. Sasse H., Snethlage R., 1997, Methods for the evaluation of stone conservation treatments, Baer N.S. & Snethlage R. (eds.), Saving our architectural heritage: the conservation of historic stone structures. Dahlem Workshop Report, London: John Wiley & Sons Ltd., 223-243