Artificial Intelligence in Civil Engineering
Artificial Intelligence in Civil Engineering
Introduction:
Augmented Intelligence and Artificial Intelligence (AI) technique is a part of computer science and is applied in the field of research and design. The idea of AI in civil engineering includes Evolutionary Computation, Fuzzy Systems, Neural Networks(NN), Knowledge-based systems, Pattern Recognition (PR), Machine Learning (ML), and Deep Learning (DL). Out of all these methods, the major ones in civil engineering applications is Neural Networks (NN) and Fuzzy Logic (FL). The traditional method of optimizing a complex structure involves crucial formulas and it needs a huge amount of computing systems. Whereas, AI techniques provide efficient and alternative valuable solutions for the problems.
In each and every civil engineering project, there may be a risk involved in it. In the civil engineering domain, the damage factor was predicted using Artificial Intelligence. Meanwhile, the safety of the project as well the economical status of the project is also increased.
Development of Artificial Intelligence:
In 1940, John McCarthy originated the term “Artificial Intelligence” to describe human thinking as the mechanical handling of symbols. AI Techniques is a science and technology based on disciplines such as computer science, Biology, Psychology, and Engineering.
The major applications of AI in the civil engineering field are broadly classified as detailed plans and specifications, construction planning, and the management of infrastructures. Infrastructure management includes the construction and management of bridges, dams, roadways, airways. Railways, buildings, and utilities.
Also, the ANN is being used to predict the performance of the budget for construction, maintenance, and structural optimization.
Optimization and Automation of Adsorption Water Treatment in AI:
Water pollution has major effects on the food chain which affect the economy and ecology of an area. To overcome this issue, AI technology has been introduced into water treatment and desalination processes. Hence, AI techniques offer practical solutions to water pollution and water security. Using AI, cost optimization and material optimization are identified quickly.
Neural Network in water treatment:
To remove the pollutants such as bacteria, viruses, plastics, and phosphates from water, different adsorbents have been used. Using AI techniques we can analyze the performance of different absorbents. In particular, the Neural Network model was used to find the chemical composition of polluted water.
Though it has many advantages, there are some limitations, such as the number of applications and the availability and selection of data, which have slowed the widespread application of AI techniques in real-world water treatment systems.
These Neural Network techniques are beneficial for researchers from the civil or environmental engineering field, students, engineers, and stakeholders in the water industry.
Assessment of Highway Slope Failure using Neural Network:
The major reason for slope failure is less safe bearing capacity and self retain-ability of soil. This may cause dangerous driving conditions for road users because of the soil or rock debris. The reasons for the slope failure mainly depend on geologic and geomorphologic factors.
Also, the exogenetic and endogenetic processes and human errors or activities influence slope failure. Before the execution of AI, the limiting equilibrium method was used to find slope failure nowadays, the Artificial Neural Network (ANN) was used to identify the nonlinear relationships within geophysical systems.
We can also apply artificial neural to predict the amount of rock deformation, to analyze the landslide, and to assess the earthquake-induced soil liquefaction.
Back Propagation Neural Network:
Using ANN, the parameters such as stability of the slope, slope incline angle, the height of the slope, and cumulative precipitation are investigated. Also, average daily rainfall, material strength, joint strength, its number, the vegetation condition, and the direction of the slope are predicted.
ANN mimics the biological neural network based on the interconnection of nodes or artificial neurons.ANN learning is executed through a series of patterns and the input and target output values are adjusted through interconnection weightage.
To assess the slope failure of highways, Back Propagation Neural Network (BPN) is also used. In BPN analysis, the error from the output layer is directed in a backward direction as an input for the next learning process through hidden layers.
AI in Surface Settlement:
The most dangerous parameter in channel excavation is the surface settlement. Hence, it is very important to predict the Maximum Surface Settlement (MSS) to minimize the possibility of surface failure. Using the ANN cross model based on particle swarm organization, easy ro determine the MSS caused during mining.ANN is also used in the field of bridge cross cost estimation, Fibre Reinforced Polymers, Recycled Concrete Aggregate etc.,
Structural Health Monitoring using AI:
At present, Artificial Intelligence is involved in large-scale infrastructure projects such as bridges, dams, pipes, roadways, railways to maintain safety and durability. In Tamar Bridge, United Kingdom, the Genetic Algorithm was used to analyze the bridge and it was continuously monitored using sensors.
The computational Artificial Intelligence-based structural health monitoring system divides the overall software system into individual micro-controlled based systems and it is maintained at the site itself.
Conclusion:
The traditional methods for computing the solution for various civil engineering problems are based on tedious field assessments. AI has been successfully applied to many civil engineering domains. It predicts cost and risk analysis, decision making, construction management, etc.,
In the future, it is expected to produce sophisticated instruments based on algorithms to reduce risk and also to increase safety.
References:
1. https://doi.org/10.1016/j.cej.2021.130011
2. Journal of Zhejiang University: Science A, vol. 10, no. 1, pp. 101–108, 2009.