Dr. Z’s Corner

Dr. Z

Ahmet Zeytinci, P.E., Ph.D., Fellow-NSPE, Fellow-ASCE is an award-winning professor, structural engineer, author and mentor living in Washington, D.C. Since joining academia, "Dr. Z", as he is known by his students and colleagues, has distinguished himself on campus and beyond. He is passionate about engineering, gifted in teaching, and is a true champion for professional licensure. Dr. Z. has extraordinarily high standards; has produced award-winning designs; is prolific in professional service; and infects others with these same values. He is the recipient of numerous local, regional and national awards, including recent national awards from the National Society of Professional Engineers (NSPE) and American Society for Engineering Education (ASEE). Since 2014, he has been regularly writing monthly articles for “Dr.Z’s Corner “ and offering hundreds of engineering problems, for free, every month for students, engineers and engineering educators worldwide. Dr. Z. also offers pro-bono Saturday classes for students and engineers; his free classes are open to all in the greater Washington metro area and cost nothing, nada, zilch! Starbucks coffee is always a must have for Dr. Z.

Dr. Z's Corner

Dr. Z’s Corner (202105)

Civil Engineering & Artificial Intelligence (AI) Applications from Netherlands

Dr. Eleni SmyrouDr. İhsan Engin Bal

This special issue of Dr. Z’s Corner will be the last before we break for the summer. This month I’ve decided to surprise our readers and invited two well-known engineers and scholars from Europe. My guest authors, Dr. Eleni Smyrou and Dr. İhsan Engin Bal, work together as a husband and wife team and currently both are faculty at Hanze University of Applied Sciences Groningen, Netherlands. I hope you will enjoy their interesting article.

Introduction

Technology is evolving at an unprecedented speed, by transforming the society, politics, governance, and professions. Civil engineering is no exception.

Computers significantly changed the way structures are engineered. The method of Hardy Cross from the University of Illinois UrbanaChampaign, for example, was revolutionary in the 30s, enabling structural engineers to design and build taller structures until the 60s. Similarly, the elastic design spectrum, proposed by Nathan Newmark who is another professor at Urbana-Champaign, revolutionized the seismic design of structures starting from the 50s. The implementation of computerized methods in civil engineering, however, was a total game changer. Now 90 years after the first publication of the Cross Method, and more than 60 years after the proposal of the Newmark Spectrum, our structural design problems are more complex than ever.

Although the civil engineering discipline adapted well to the early changes of computerization, the adoption of emerging technologies in the new millennium is slow. The use of brute-force when using computers to analyze and design larger, taller and more complex structures has become the main exploitation area of technology in civil engineering. Furthermore, structures have become much more complex in the last few decades, requiring interface with other disciplines via technologies such as BIM (Building Information Modeling), which is another technological development that found a place in practice. Apart from those, and despite the extensive research, other emerging technologies did not actually revolutionize the design and construction processes, yet; though this may change in the coming years.

Things are changing recently, although slow and limited. Some new technology applications in civil engineering are evident in the last few years. The momentum in new technologies spreading through our daily routines, and the increasing societal and economic demands, are forcing the civil engineering discipline to adapt.

In this article, we discuss one of the major emerging technologies, “Artificial Intelligence”, the magic word of the recent years.

What is AI?

Artificial Intelligence (AI) is a broader term that covers all sorts of applications where the intelligence is developed by a machine. Although the concept dates back to the 40s, everyday applications were only possible when the available computational power was enough to deploy large datasets for training models. This happened in the last several years, thanks to the use of GPU (Graphics Processing Unit), which was a major breakthrough for the realization of real-life AI.

The computers consist of CPU (Central Processor Unit) and GPU among other components. The processes are usually done by CPU and thus the computational power of your computer will heavily depend on the power of your CPU. GPU, however, is attached to the graphical unit of the computer and is used only for graphically demanding applications, such as playing videos, rendering 3D models, or playing games. Over the years the GPUs have become stronger and stronger, creating a sort of hidden power inside every computer, but used only for graphical purposes. It was not until a few years ago that it was discovered that the GPUs can be extremely useful for training AI models because they can run multiple training processes in parallel, something which can speed up the AI training significantly.

Today AI is almost everywhere. Entertainment platforms on the internet, for example, provide song, movie or series recommendations based on AI technology. The more time you spend on such platforms, the more data the AI algorithm will collect, get better trained and, in a way, get better acquainted to you. It will eventually provide you better and better recommendations. After a while these platforms become like a good friend who know your taste very well. That is the machine intelligence we are talking about.

Can AI Detect Structural Problems?

The same concept as entertainment platforms applies in almost all fields. In civil engineering for example, AI is already used in several areas. One of the most successful applications is detecting structural problems, anomalies, damage and deterioration based on photographs. Similarly to the entertainment platforms, the more data which is provided, the more accurate the model and the predictions become. Such AI-powered tools are more suitable for existing structures at the moment, thanks to the abundance of data for training models. Even if not, it is easier to collect data from existing structures rather than the new structures that are not even built yet. That is why many engineering firms are digging out their photographic databases to see if they can throw these photos into a smart AI model and replace the laborious engineering work of damage detection with computer codes.

In recent work1 with our colleagues from Hanze University of Applied Sciences in Groningen (Netherlands) and University of Leeds (UK), we showed that a simple photograph would be enough to detect a crack on a masonry brick surface. Earlier methods were able to only place the crack in a bounding box, telling us simply “there is a crack somewhere inside this box”. Our method detects the cracks pixel-wise, telling us the exact location, length, spread and width of the crack in a photograph. This was not achieved in masonry surfaces before, although AI-based crack detection is more advanced in concrete and asphalt surfaces, which are rather homogenous. Our work brings new opportunities such as regular scanning of historical buildings or old masonry structures based on simple photographs, an opportunity that will significantly reduce costs and time, and allow access to many more structures. What is superior in this method is that, even photographs taken by citizens and non-technical people can be used for extracting engineering information in a fully automatic fashion.

We support the crack detection technology with other emerging technologies, such as 3D scene reconstruction and near-infrared (NIR) crack width estimation, among others. The former is a method that can build a 3D computer model if enough photos are taken from a real structure, while the latter is a method where we developed invisible markers that reflect light only if special NIR cameras are used. Both of these technologies allow us to train a building responsible person or a citizen for taking suitable photographs for engineering purposes. Photographs can be taken regularly or after an event, such as earthquakes or deep excavations. We are already testing the combination of these technologies in a project funded by the Cultural Heritage Agency of the Netherlands (RCE).

AI is a promising concept and will certainly find further application areas in civil engineering, helping engineers make critical decisions for complex problems.

About the Authors:

Dr. Eleni Smyrou and Dr. İhsan Engin Bal work at Hanze University of Applied Sciences Groningen, Netherlands, in the Earthquake Resistant Structures research group. They both have degrees in civil engineering, as well as M.Sc. and Ph.D. degrees in earthquake engineering. Their work areas are seismic design, assessment, monitoring and strengthening of structures. Use of new technologies for structural safety has been a major research agenda topic for them for the last couple of years. They are also co-founders of “Senso Engineering – Vibration Solutions” and “Strintel – Structural Intelligence”, two startups which are providing services on the use of new technologies for structural safety.

1 https://www.sciencedirect.com/science/article/pii/S0926580521000571
(this is an open access article which is freely downloadable)

 

Dr. Z’s Corner (202104)

It’s the Economy Again, Stupid!

Sabine O' Hara

The phrase, ‘It’s the economy, stupid’ has been attributed to James Carville, Bill Clinton’s successful 1992 presidential campaign strategist. And indeed, it is. The economy is both blessing and curse. It creates both well-being and disaster. If we ever needed a reminder of the omnipresent impact our daily economic activities have on our lives, the global COVID-19 pandemic is just such a reminder. So, what then, is the economy?

Economic theory distinguishes between three sets of activity: production, resource allocation, and consumption. These distinct sets of economic activity can also be described as the ‘what’, ‘how’ and ‘for whom’ questions of economics. The first describes production activity and its many interesting issues related to the production of goods and services including production costs, efficiency, innovation, and competition. The second describes the recipe of production, and includes such interesting topics as labor inputs, energy, and renewable vs nonrenewable resources. The third describes the demand portion of the economy including what kind of products and services consumers need or want, and how responsive they might be to price changes.

This past year has highlighted all three of these economic spheres. For example, what goods and services do we produce in the United States and where do we produce them? Which ones do we import and how reliable are the supply chains of these imports? Take our food, for example. Eighty percent of our fruits and vegetables are produced in California where water is increasingly scarce. Most of our meat comes from Texas. Both are a far distance from the distribution centers in high population areas on the east coast. Much of our food is also imported and what is produced domestically requires the labor input of migrant workers who come to the US from Mexico and other central American countries. Without them, our strawberries, salad greens, and beans would not get harvested and would rot in the fields. We learned firsthand how vulnerable some of our supply chains are as workers in meatpacking plants and distribution centers became known as the essential workforce we depend on to get our food delivered to local grocery stores or to our doorsteps. These essential workers are some of the lowest paid and their working conditions make them highly vulnerable to COVID-19 transmission.

Nature too suffers the consequences of a highly centralized food system characterized by long supply chains. Eleven percent of US greenhouse gas emissions are associated with the food system. Globally, agriculture is responsible for 25 percent of all CO2, 65 percent of methane, and 90 percent of nitrous oxide emissions and uses 70 percent of our freshwater. The positive impact of reduced travel was also noticeable during the pandemic as greenhouse gas emissions markedly declined.

Consumer demand shifted as well. Demand for home improvement related goods and services sored while tourism, hospitality and other customer facing service sectors suffered. The increasing bifurcation of our demand sector also came into stark focus as some households depend on food from local food banks and are forced to live in shelters or their cars, while others were able to bolster their savings. Women left the workforce in especially high numbers in part because they occupy a high portion of service sector jobs, and in part because they took over the bulk of home schooling and other child rearing responsibilities. Labor demands in and outside of the formal workforce are not evenly distributed.

So, what have we learned? My hope is that we were reminded of what economists call negative externalities. These are the unintended and usually negative side effects of our economic activity. These side effects show themselves in the pollution of our air, rivers, and soil, plastic islands and chemicals in our oceans, global climate change, growing inequalities, and social unrest. Negative externalities don’t just disappear. They are displaced over time and space. Their consequences are rarely suffered by those who create them, but by future generations and those least able to defend themselves.

My hope is that we will begin new collaborations. Economists from Malthus (1766–1834) to the Club of Rome (1968) have warned of this disastrous prospect of resource depletion and a rate of population growth that outpaces our rate of productivity growth and especially our ability to increase food production. Yet even as we face the prospect of a world population of ten billion people, we have reason for optimism. Technology and its efficiency increases has enabled us to produce more oil, more food, and more consumer goods. In fact, we have produced so much more that prices have fallen, and demand grows steadily. Yet new collaborations are needed to address the real frontiers of technology and efficiency.

The true limit to growth may not lie on the resource side of the economic process, but on the sink side. Sinks are the earth’s genius ways to absorb the emissions and waste by-products of our economic activity. As waste and emissions are released, they are processed, absorbed, buffered, and accumulated in a set of context systems we call our environment. Oceans absorb CO2, soils absorb water and the emissions it contains, the air takes up and dissipates NOx and SOx emissions. And there are social sinks as well as workers are loved and cared for in families and communities who absorb the stress and strain of their exhaustion and injuries.

This means that the field of economics, and the field of engineering, must be concerned with more than the resources necessary to sustain economic production and consumption. We must also concern ourselves with the sink capacities that provide the capacity to reduce the negative impacts created by growing emissions and waste. Can emissions not only be reduced but reversed? Can waste be reused with minimal energy inputs? Can buffer capacities of the soil be restored? Can we not only reduce CO2 emissions, but reabsorb excess CO2 without creating new unwanted side effects?

In other words, we need a new concept of economic activity that shifts our focus from sources to sinks and the value these sink functions create. Such a sustainable economy will increase efficiency so fewer inputs can result in the same or higher output levels of goods and services; it will reduce emissions and waste resulting from the production and consumption of these goods and services; and it will restore and improve the ecosystems services that deliver the sink capacities necessary to process emissions and waste and maintain the health and vitality of our physical and social environment. Such a new economy will certainly require circular designs and systems approaches.

Food is a good place to start. At the University of the District of Columbia and its College of Agriculture, Urban Sustainability and Environmental Sciences (CAUSES) we launched an Urban Food Hubs model that creates small scale circular food systems consisting of (1) food production, (2) food preparation, (3) food distribution, and (4) closing the loop through waste and water management. The urban food hubs are ideally located in neighborhoods that lack access to fresh food. They model a circular, decentralized food system that can supplement rural production with greens, tomatoes, peppers and ethnic crops so that the most perishable and nutrient rich food plants can be produced right where the majority of consumers live. I invite you to join in this journey of re-envisioning such a circular, decentralized economy that understands how nature works, and how people live well – one community and one product at a time. It’s the smart economy of the future!

About the Author

Sabine O’Hara, is Distinguished Professor and Program Director of the PhD in Urban Leadership and Entrepreneurship in the College of Agriculture, Urban Sustainability and Environmental Sciences (CAUSES) at the University of the District of Columbia (UDC). As founding dean of CAUSES she led UDC’s efforts in building a cutting edge model for urban agriculture that improves urban sustainability and the quality of life of urban communities.

 

Relevant Articles