Artificial Intelligence Project report.pptx

Artificial Intelligence

1. Artificial Intelligence in Chemical Engineering

2. Content 01 Introduction 02 How AI helps in Chemical Engineering? 03 Evolution of Artificial Intelligence in Chemical Engineering. 04 Sources

3. 01 INTRODUCTION:

4. What is Chemical Engineering? Chemical Engineering is defined as the engineering discipline in which we design process to transport, transform and produce new material and it also involves modifying existing processes or methods. In chemical engineering we apply our basic science knowledge to convert our environment raw material into valuable products. With the help of chemical engineering we can adjust the rate of reaction according to our products need. This discipline uses the principle of chemistry, physics, biology mathematics and economics to design chemical equipment process and system to refine raw material into useful products through chemical manufacturing. This engineering discipline also requires the understanding of chemistry to produce valueable products like gasoline, jet fuel and monomers. Chemical engineers also works with inorganic and biological materials in order to produce pharmaceutical products and also fine chemicals. Chemical engineers also deals with industrial wastes in most efficient way to produce something usable or to discharge them with no or less harm to the environment. In chemical engineering we always try develop new methods or modified process for raw materials conversions so that it produces less or no waste. Many novel materials are developed by chemical engineers, some other them are polymer, biomaterials and nanostructures. They also work on composites, batteries, fuel cells , monitoring systems, food industries and pollution control devices. Chemical engineers focus mainly on chemical process design like designing a processing plant or designing a chemical reactor which is where a large scale chemical reactions take place and ofcourse more equipments are there.

5. What is Artificial Intelligence? Artificial intelligence is a intelligence of machines. Now a days machines are as intelligent as humans even more intelligent than humans. AI is a powerful tool which runs on the command given to it in the form of programs. AI is present everywhere in our present modern world like when we use our smart phone, We get our personal assistant which is only possible because AI is present, as it collects our data and gives us personalised results. When we search anything on Google AI uses our data to show us the most relevant results. Apps like Netflix, Amazon, Flipkart, Instagram etc also uses our data to get us the specific content which we are trying to find. Another advantage of AI is image processing, In this we are able to find any object form its image. Best Example to this is Google Lens. Some other AI based applications are also their like Truecaller , which identifies potential spam, Google translate, which translate any language to another language. Artificial intelligence is also the most important part of the modern robotics.

6. 02 How AI helps in Chemical Engineering?

7. How AI helps in Chemical Engineering? Artificial intelligence has power to bring revolution in Chemical Engineering by providing faster, more accurate and interpretable results or outputs. Artificial intelligence can process a large no of dataset at once. With the help of AI we can also optimise difficult process and get outputs based on the historical data given to it. In chemical engineering we use machine learning and deep learning for process optimization, quality assurance, predictive maintenance, supply chain management, energy efficiency, safety and risk assessment, novel materials development etc. With the combination of AI with the first principal model and data driven approach, the limitations and the future of AI in chemical engineering systems was addressed.Artificial intelligence is also used in material science for property predictions, material processing and also for the study of structure and materials properties.In process system engineering, AI techniques is used for process optimization, modelling, design and operations, which also includes the process monitoring and fault diagnosis .

8. 03 Evolution of Artificial Intelligence in Chemical Engineering.

9. PHASE I: (1983 – 2003) Expert System Era (Symbolic AI) Seperation of domain knowledge fro interface. Flexible execution order of program. If then rule for procedural knowledge. Semantic network for taxonomies.

10. Systems developed during this phase: MYCIN (1972 – 1982): Expert system for diagnosing infectious diseases. CONPHYDE (1983): Thermodynamic property prediction. DECADE (1985): Catalyst Design. MODEX (1986): Fault Diagnosis. DESIGN KIT (1987): Process design.

11. Diagnosis Tool Kit ( Dkit 1993-98) Dkit is designed to process the real time data from the system and to diagnose weather the system behaviour is normal or abnormal and if it is abnormal then dkit find the fault and suggest us the appropriate recommendations. It was implemented in G2, and tested at Exxon. Dkit successfully diagnosed failures even before the alarms went off (1/2 – 2 hours ahead ) It is successful for small scale but fails at large scale data.

12. PHASuite : Automated Causal Reasoning System (1995 – 2005) In Diagnosis ( Dkit ) we go from symptoms to cause, where as in prognosis we create abnormal causes to see their symptoms on the system so that such system was built. For that reason PHAsuite was built. It has a library of causal models and also had a reasoning engine. It saves 40-60% of time, effort and money.

13. PHASE II: Machine learning I (Neural networks 1990 - 2005) Too much time, effort, and specialized expertise Did not scale well for industrial applications Backpropagation algorithm (1986) Bottom-up strategy Automatically learned patterns between input and output vectors by adapting the weights. Expert System Drawbacks :

14. PHASE II: Machine learning I (Neural networks 1990 - 2005) The key breakthrough this time was the ability to solve nonlinear function approximation and nonlinear classification problems in an automated manner using the backpropagation learning algorithm. This novel automated nonlinear modeling ability spurred a tremendous amount of work in various domains, including chemical engineering. Researchers made substantial progress on addressing challenging problems in modeling, fault diagnosis, control, and product design. Notable contributions include the recognition of the connection between the autoencoder architecture and the nonlinear principal component analysis by Kramer and the nature of the basis function approximation of neural networks through the WaveNet architecture by Bakshi and Stephanopoulos.

15. So, why was Al not impactful in ChE during (1983-2010)? Researchers made great progress on conceptual issues Showed how to formulate and solve these challenging problems But we were greatly limited by implementational and organizational difficulties for practical impact Lack of computational power and computational storage Lack of communication infrastructure No Internet, WirelessLack of convenient software environment Lack of specialized hardware - e.g., NVIDIA GPU for simulations Lack of dataLack of acceptance of computer generated advice Costs were prohibitive

16. PHASE III: Machine Learning II (Data Science) Through the application of data-driven tactics in Phase III, chemical engineering has advanced by tackling issues such as speech recognition and picture recognition. These methods, however, call for large data sets that chemical engineering applications might not have. Phase I and II strategies can be employed to address this, but it is imperative to integrate first-principles understanding with data-driven models. Using examples from the industry, AI may help with defect diagnostics, process operations, and materials design.

17. PHASE III: Machine Learning II (Data Science) Convolution or Deep Nets. Reinforcement Learning Statistical Machine Learning Hierarchical feature extraction What really new is Data GPU and Software Big impact on Robotics, Vision, NLP

18. 04 Sources

19. Resources The Promise of Artificial Intelligence in Chemical Engineering: Is It Here, Finally? Venkat Venkatasubramanian Dept. of Chemical Engineering, Columbia University, New York, NY 10027 DOI 10.1002/aic.16489 Published online December 19, 2018 in Wiley Online Library (wileyonlinelibrary.com) Venkatasubramanian V. Artificial intelligence in process engineering: experiences from a graduate course. Chem Eng Educ. 1986;188-192. Venkatasubramanian V. Inexact reasoning in expert systems: a sto chastic parallel network approach. In: Proceedings of the Second Conference on Artificial Applications, pp. 191–195, 1985. Venkatasubramanian V, Chan K. A neural network methodology for process fault diagnosis. AIChE J. 1989:1993-2002. https://pubs.acs.org/doi/epdf/10.1021/acs.chemrev.1c00108

20. THANK YOU !!