513335690-AI-applications-in-civil-engg.pptx

513335690-AI-applications-in-civil-engg.pptx

1. AI applications in Civil Engineering By. Piyush Bhandari

2. 1. Structural Design AI-based applications will enable the designers to collaborate with computers to create complex designs quickly. Designers can utilize software tools to input different design goals and criteria. Software solution utilizes AI to generate options for designers to select from alternatives and make necessary changes as needed.

3. 2. Estimation: Generative design combined with data on costs and schedules of similar projects will provide great value for estimators. Machine learning systems can analyze the data from similar projects to develop preliminary cost and schedule estimates. Combining AI with BIM allows estimators to develop accurate estimates with greater accuracy within a quicker turnaround time.

4. 3. Safety management AI-based applications utilizes visual processing algorithms as a important risk monitoring and prevention tool for safety managers. Photos and videos from the jobsite are analyzed for safety hazards such as missing workers, not wearing appropriate PPE gear etc. Similar to project managers, many safety managers are managing multiple projects and cannot be at the jobsite all the time. However, the field team is capturing photos at the jobsite. Safety monitoring solutions that utilize AI works will scan large amounts of photos and quickly. It will identify workers and instances that are not following safety protocol.

5. 4. Project management Project managers utilize autonomous devices such as drones, sensors, and cameras to monitor job site activity. AI-based applications will utilize data to measure the quantity of materials consumed by analyzing the job site in 3D. The learning algorithms track the progress in real-time against original plans, budget, and schedule. Project managers use the progressive information to track labor productivity and make adjustments to keep the project on schedule.

6. 5. Construction foremen The AI-based applications are used to measure the quantity of materials used. The same learning algorithms will measure the quality of work. The 3D model captured by drones and cameras is compared against the original design to detect any errors or inconsistencies. Foreman can receive timely alerts to immediately work on corrections before the scope for rework increases.

7. 6. Water Resources Engineering For Time series Prediction Rainfall-runoff prediction Rainfall forecasting Stream-flow prediction Reservoir inflow prediction Reservoir operation

8. 7. Irrigation engineering Evaporation prediction Evapotranspiration prediction Irrigation return-flow prediction

9. 8. Structural Analysis Automated conceptual design of structural systems Structural dynamics involving Earthquake Non-linear analysis of plates Detection of damages Vibration analysis

10. 9. Transportation Engineering Traffic flow simulation Automatic signals Priority of highway maintenance

11. 11 10. Geotechnical Engineering Hydraulic conductivity Soil thermal resistivity Strain-rate dependent behaviour of soils Prediction of settlements during tunneling Predict settlement of shallow foundation Assessment of damage of pre-stressed piles Capacity of piles in cohesionless soils

12. Applications of ANN 1. Rainfall Forecasting: Multi layered feed forward network can be used along with Back Propagation training algorithm for conversion of the remote sensed signal into rainfall rates. It is further used to determine the runoff for a given river basin. ANN technique can be applied for rainfall-runoff modelling. Daily runoff can be forecasted by giving input of daily precipitation, temperature and snowmelt into multi layered feed forward network with Back propagation algorithm. Hourly rainfall values can be disaggregated into sub hourly time increments with the help of ANN. Feed forward back propagation steepest descent learning algorithm is used along with unipolar activation function.

13. 2. Runoff Prediction: ANN with linear and non-linear regression techniques can be used for spring runoff prediction For real time forecasting of hourly flood runoff and daily river stage and to the prediction of rainfall sufficiency for India. Runoff Forecasting can be done with back propagation, conjugate gradient and cascade correlation training algorithm. ANN model can predict both runoff and sediment yield on a daily as well as weekly basis from simple information of rainfall and temperature.

14. 3. Reservoir operation: ANN based model can be used to monitor the monthly inflow data series and their performances. ANN is found better for predicting the high flows. Also ANN is an powerful tool for input-output mapping and can be used for reservoir inflow forecasting and operation. ANN based simulation optimization can be used for reservoir operation. For multivariate reservoir forecasting, different types of neural network architectures, i.e. input delayed neural network (IDNN) and a recurrent neural network (RNN) and multi layered perceptron (MLP) are used Dynamic programming based neural network model is used for optimal reservoir operation.

15. 4. Stream flow prediction Flows in streams are main input for design of any hydraulic structure or environmental impact assessment. Cascade correlation algorithm is used instead of designing and trying different types of architectures for predicting daily stream flow data. This type of algorithm uses incremental architecture in which training starts with minimal network and goes on increasing size as proceeds. To forecast level of water in river. Different types of algorithms like hack propagation, conjugate gradient and cascade correlation can be used. Among them, Cascade correlation algorithm is the fastest for the training of the network. To establish the rating curves. Three layered back propagation training algorithm is used. ANN is significantly better than conventional curve fitting techniques. For forecasting river stage and for the sensitivity analysis

16. 5. Estimation of evapotranspiration For estimation of daily grass reference crop evapotranspiration. Multi layered feed forward network with back propagation is normally used. Single hidden layer is capable for developing non-linear relationship between climatic variables and corresponding evapotranspiration.

17. 6. Draught Analysis A draught is generally defined as an extreme deficiency of water available in the hydrologic cycle over an extended period of time. Draught forecasting plays a vital role in the control and management of water resources systems. ANN is used to forecast draught. Conjunctive models are used to significantly improve the ability of ANN to forecast the indexed regional draught. Three layered feed forward network with back propagation training algorithm is used. Accurately predicted draughts allow water resources decision makers to prepare efficient management plans and proactive migration programs that can reduce draught related social, environmental and economic impact significantly 

18. 7. Soil water storage ANN is used analyse the soil water retention data. A three layered feed forward network is used in input layer, one neuron represents the matric potential values and the output neuron represents corresponding moisture content. Hidden layer having 5 neurons and sigmoid transfer function is also used. Then Back propagation training algorithm is used. 

19. 8. Flood Routing Information theory and neural networks are used for managing uncertainty in flood routing. Parallel ANN model uses state variables, input and output data and previous model errors at specific time steps to predict the errors of a physically based model. ANN models can remove the errors of physical based models also reduce the prediction uncertainty.

20. 9. Model Drainage Pattern ANN is used to automatically determine the drainage pattern from digital elevation model (DEM). Three-layered network with back propagation algorithm is used for training

21. 10. Classification of River Basins Pattern clustering and pattern mapping capabilities of ANN are used for classifying river basins. An unsupervised ANN architecture viz. Adaptive Resonance Theory (ART) is used for pattern clustering that is grouping of basiss of hydrological homogeneity. Multi layered perceptron is used for pattern mapping.

22. 11. Tidal Level Forcasting Tidal level record is an important factor in determining constructions or activity in maritime areas. Back propagation neural network is used to forecast the tidal level using the historical observations of water levels. However, their model is useful only for the instant forecasting of tidal levels, not a long-term prediction.

23. 12. Earth Retaining Structures Neural network model is used to provide initial estimates of maximum wall deflections for braced excavations in soft clay. The input parameters used in the model are the excavation width, soil thickness/excavation width ratio, wall stiffness, height of excavation, soil undrained shear strength, undrained soil modulus/shear strength ratio and soil unit weight. The maximum wall deflection i s the only output. Neural network model is a time-saving and user friendly alternative to the finite element method.

24. 24 Application of Genetic Algorithm (GA) 1. Pipe network Design of networks, and analysis 2. Ground water management problems Quantity and quality management models

25. 3. Reservoir operation For Single purpose single reservoir For Multi-purpose single reservoir For Multi-reservoir systems For Multi-purpose multi-reservoir systems