Prediction analysis of the degradation and the service life building components in artificial method neural network and ISO factor 15686-2

The research study reported in this paper aims to combine the Artificial Neural Networks with ISO 15686 Buildings and constructed assets - Service life planning, a framework-based approach to offering a more reliable deterioration forecasting more reliable for building. This paper discusses the existing data and develop a close relationship definition between factors affecting the condition of the service life of the building, the value of building condition and determine the level of degradation of the building component, also predicted the age of the building components in accordance with a specific time variables. Data examination conducted in this research is building condition data of student dormitory at the Universitas Gadjah Mada, the data will be used to calibrate the model of damage to consider a number of factors that influence. To help demonstrate the concept, factors affecting the decline is considered in the analysis of the design level, the level of implementation of the work, the indoor environment, the external environment, the level of care and conditions of use. Predictive analysis with artificial methods of neural network (ANN) with ISO factor input variables and factors age of the building components and the severity level of degradation of the building components (Sw) as output, this will generate a calculation formula that shows the effect of each variable input to output. Predictive analysis carried out with the reverse approach in which after calculation formula obtained by ANN method, then the next step is to find the value of the variable age of the building components according to the value of degradation that has been determined.


Background
Another type of asset management infrastructure is the implementation of maintenance and care building. Any assets of infrastructure development need maintenance and care to ensure that infrastructure building they can properly function and to make sure that component structures meet the requirements during their economic life. The main problems encountered in asset management infrastructure include: 1. The management of infrastructure assets has not been able to provide detailed information on the condition of the assets owned so that the maintenance and maintenance plan of the asset can not be established based on the priority scale. 2. The types of maintenance and maintenance of each category of asset damage have not been uniform, so the potential in which the same damage, but behaviour can set different treatments.
3. Lack of maintenance standards and maintenance of assets that can be applied in making decisions in asset management. Based on these problems, it is necessary to identify early and predict the condition of building conditions regarding the preventive maintenance plan and corrective maintenance and regularly scheduled basis, in order to prevent the failure of building elements and minimize repair costs. Flores-Colen, I., & De Brito, J. (2010) explains that a poor design of construction details, a bad choice of the façade materials (e.g. plaster with high porosity in a marine environment), its inadequate application, and inexistent maintenance are the core of current problems in buildings' façades [1]. This research its substance referring to bridge management pattern that is developing better by now. Bridge infrastructure management based computer system developed by the Directorate General of Binamarga form of Bridge Management System (BMS raises the idea that the system could be applied to the management of the building. Therefore, there needs to be a study related to the method of building management especially the aspects of maintenance and the care of the building conforming by established standards and are effective and efficient. This research analysis predictions discusses the predictive analysis of the level of degradation and component service life of buildings taking into account the variables that affect the condition of the building, as a reference in determining the treatment plan. The methodology used to support the analysis in this study is the ISO 15686 standard, namely Buildings and constructed assets -Service life planning (future serviceability and asset building construction). International Standard ISO 15686 provides information on general principles, frameworks, and methodologies to estimate the service life of building components, as well as the prediction to determine the maintenance time and replacement of components necessary for building the infrastructure (ISO 15686 part 1 of the Introduction of the Standard). The method of calculation analysis to determine the predictive value of age serviceability of building components used in this study, conducted using Artificial Neural Network.in addition to determining the influence variable components used in the calculation analysis refers to the explanation in the standard document ISO 15686 part 2 of the Service life prediction procedures [2][3].

Problem Formulation
The problem can be formulated based on the description of the background that has been described above is how to identify the factors that affect the condition of the building, how to determine the level of degradation of a component of the building and how the formula can be used to predict the condition of service life of the building components on certain time.

Objective
The objective of this study is provide an explanation regarding: 1. Determination of the factors that affect the condition of the building service life. 2. How to calculate the value of the condition of the building and determine the level of degradation of the building components. 3. How to formulate a predictive value of a service life of building components in accordance with a specific time variable.

Benefits Research
Benefit of this research is to be able to determine the extent of degradation of a building component and predict the service life of a component of the building which is expected to be a reference to determine the best treatment strategy for the condition of the building.

ISO 15686, Buildings and constructed assets -Service life planning (2011)
The Kmax is the highest degradation to a component being examined.
A is the total area of a component.

Expert System Algorithm
According to a general description of a service life calculation analysis by ANN building components can be described as in the following Figure 2 [12]:

Regulation of the Minister of Public Works and Public Housing and the Building Condition Assessment of Governance How Building Maintenance
Regulation of the Minister of Public Works No. 24 Year 2008 contains an explanation regarding the implementation of the basic rules of maintenance and upkeep of buildings. In which includes the classification of the level of damage and the method of maintenance and care. As for the types of damage to buildings classified in the following levels:

Discussion
The discussion conducted on as many as 100 variations of a building wall component data degraded.

The calculation of the rate of degradation (Sw)
The calculation of the value of the degradation of the components of building walls is done by the following formula [12]: (3) From the coefficient at each type of damage as listed in Table 1 above, can further degradation values obtained for a building component. The results of the degradation value and the wall component damage level category for 100 sample data can be seen in the following Table 2: In addition, for the output variable is the value of the building component degradation (Sw) as in Table 2. The analysis of the degradation of the building wall components with ANN method is detailed as follows: 1. Incorporate variable input and output variables in Matlab program. 2. Running data by ANN composition analysis, hidden layers numbered 1 and the number of neurons numbered 2, as seen in Figure 3.  Table 3 below: The result from calculate of degradation is:

Conclusion
From the results of the analysis performed, we concluded the following results are from the analysis of 100 samples of the condition of building wall  1.4643 Z2).
From the results of the activity a condition by a wall with existing age some 38.5 years obtained the value of degradation is 43,912 % so that it can be categorized into the condition are heavily damaged. And the results of prediction components to a condition the wall that was 38.5 years such as the point of (2), has a maximum service life of 43 years as described in Figure 5.