تجزیه و تحلیل حساسیت از خاک ریزی لاستیک ضایعات با استفاده از نمودارهای نفوذ بیزی

Scrap tires have several properties that make them preferable to other materials as fills for embankment construction, including light weight (the dry unit weight is 1/3 that of soils), high hydraulic conductivity (up to 23.5 cm/s), and low thermal conductivity. These properties of scrap tire fills result in low lateral pressures on the abutment wall and in reduced design and construction costs. The low thermal conductivity helps to prevent permafrost action of soil layers beneath it and failure of the subgrade due to frost penetration. However, scrap tires possess high compressibility, a property that leads to settlement of the fill and consequent failure of the embankment. Other undesirable attributes of scrap tire embankments are susceptibility to internal heating and leaching of substances into surrounding water. An efficient means of controlling such undesirable attributes in the field is by comparing them with those simulated from a model embankment developed using Bayesian influence diagrams. In this work, the essential responses simulated using the Analytica® software program are the temperature, lateral pressure, settlements, and leachate characteristics. The most critical embankment characteristics, based on the maximum probability densities, are the settlement and horizontal pressures, which are relatively low at 0.428 and 0.0034, respectively, because the likelihood that these values will be exceeded in the field is high. Temperature response was not considered critical because the maximum probability density simulated was 0.9301. Limits for leachate concentrations were also obtained for the model embankment based on ASTM D 6270 (1998) standards.

A number of alternative materials to soil have recently been considered for the construction of embankments such as fly ash, cement kiln dust, high-quality wood waste, and construction and demolition debris. Presently, scrap tires are generating a lot of interest due to their versatility as a construction material and useful properties such as high permeability, low thermal conductivity, and lightweight. However, accompanying these favorable properties are three major disadvantages – large settlement, potential for internal heating, and leachate effects. To forestall failure, the behavior of the embankment is simulated based on the loading and exposure conditions. The response of the embankment is varied using the decision software program Analytica and compared with the actual system as well as standard specifications to detect any anomaly. The responses simulated are the settlements, vertical pressures, horizontal pressures, temperature, and possible leachate concentrations. These are intrinsic characteristics of an embankment which depend on predictor variables that influence one another and the overall performance of the embankment during use. The levels to which the variables affect one another are expressed with the aid of influence diagrams. An influence diagram can be described as a network for representing the probabilistic relationships between the variables that constitute the system behavior [2]. In this case, the system is the embankment behavior. The nodes represent the variables that have information stored in them such as uncertainty, decisions and objectives. The arrows represent the causal relationships between variables. The influence diagrams provide a descriptive way of analyzing the uncertainty between the variables, which is represented by a probability distribution over the states. Influence diagrams can be used to analyze a problem so as to arrive at the best solution for it. They can be used to monitor the entire response of a system and simulate the likely occurrences if one or several of the factors are added, altered or removed. Consequently, a model that properly represents the operation of the problem can be developed. The influence diagram shows the dependencies between data and the state of knowledge at a decision point and enables evaluation of the outcome under various scenarios [3]. The aim of this paper is to formulate and solve field behavior of scrap tire embankment using Bayesian influence diagrams. The geometry and other inputs of the simulation model are similar to the field construction and the materials.

The temperature profiles down the embankment due to variations at any depth other than at the surface are similar; this is evident from the fact that their probability distribution curves have similar gradients. For any two or more points at different depths within the embankment, the temperature across the embankment section does not vary. This means that only a few sensors are required to study temperature variations within a scrap tire embankment. In addition, it is very unlikely that temperature variations within an embankment of similar geometry on site could pose any reasonable risk considering that the maximum probability density for the simulated temperatures is considerably high (0.9301). • Since the relative settlements are highest near the surface of the embankment (10.15 cm) and the corresponding density of the tire shred materials is lowest here (0.126), very sensitive sensors will be needed to capture the information at these locations. At points further within the embankment such as at a depth of 1.1 m, where the corresponding density of material under deflection is greater (i.e., 0.428) and the settlement is lower (2.987 cm), information will be easier to obtain here using sensors that function reasonably well. • The horizontal pressures in a hot climate should be considered critical when designing the abutment wall, since they are greater in magnitude than the horizontal pressures in a cold climate. The resulting normal distributions of pressures at different depths within the embankment are similar, and the peaks occur at an average pressure corresponding to a depth of 5.49 m, i.e., the bottom of the embankment. The temperature conditions under which the model is analyzed for horizontal pressures should be further varied to consider how other intermediate temperatures affect the pressures. Further studies into how this aspect of the model can be developed to analyze horizontal pressures at different temperatures will be required. This will help to produce distinct differences between the normal distributions of the horizontal pressures and ensure better interpretation of the results. • The characteristics of the embankment that require critical assessment are the settlement and horizontal pressures because their corresponding maximum probability densities simulated by the model are relatively low at 0.492 and 0.0034, respectively.