The pace of change in the infrastructure industry is moving faster than it has in quite some time. What is driving this acceleration is a combination of cloud services and information mobility, and the permeation of this technology in everyday life. Indeed, this evolution of mobile devices and cloud computing is decreasing the amount of time consumers spend on the Web using a desktop or laptop computer.
That’s now reached the infrastructure world. But the cloud services and mobile applications being developed by infrastructure software vendors are different. They are for serious industrial computing and are augmenting the complex engineering models that are being created on the desktop. Vendors, like Bentley Systems, are developing technologies that work on the engineering data sets that are stored on servers and increasingly on cloud servers. That’s a big change. Software vendors are also building on top of those consumer technologies, and working on BIM models and construction models, and simulation models and datasets. This is all very new.
In addition, vendors are increasing their capabilities in analysis and simulation across the complete portfolio of engineering and infrastructure products, whether on land or in the ocean for offshore platforms and vessels; whether it is commercial, industrial, structural, fluid, or all of these. Simulation, for instance, is a critical decision support method that has, historically, taken place during the design phase. Increasingly, however, analysis and simulation provide decision support through the whole asset lifecycle. The biggest change to our business, though, is the emergence of cloud computing.
As local desktop performance and structural modeling capabilities improve, it is now becoming easier for people to create higher fidelity models, which include more detail and a more accurate representation of the real world. And, with this added detail, the compute time is increased. Therefore, having a more scalable compute capacity to apply to those more-detailed models has gone from being a “nice to have” to becoming a necessity. Another area where cloud computing is being applied is in reality modeling. This is an area where large data sets are collected from photos or laser scanning and the ability to compute in parallel is a big benefit for the user.
Engineers are creating models of higher fidelity such as this SACS offshore structural model.
Now, most of the discussion surrounding the cloud is about storage and communication. From an analysis and simulation perspective, the compute part is an interesting one, because when you have unlimited scalability in the cloud, the computing resources really change the game. There can be computers that are virtual computers, and that’s really the idea behind the cloud. A computer can be so scalable, so fast, and have so much memory that software can be written to make the computer look like 100 other computers or even a 1,000 computers. With these infinitely scalable computers, capabilities can be added that allow user organizations to run computations on another organization’s computer.
Each computer is able to run in parallel across dozens or hundreds of compute nodes to increase the number of analysis runs that can be carried out for design exploration. For the most intractable problems, the compute throughput increases so that any given problem can be computed in a short amount of time. Another area where cloud computing can be applied is in reality modeling, which employs point cloud data and LiDAR scans that accurately survey sites to create detailed documentation. The surveying is typically done digitally and helps to ensure the fidelity of the project. Equally important, by analyzing and researching the digital documentation, design teams are able to understand and replicate the techniques used for the initial construction.
Another extension of cloud computing that has come about recently is “optioneering,” which means investigating more design alternatives and comparing them to one another. This “design exploration,” if you will, can be done in several ways. One aspect is by creating geometrical alternatives by stretching, transforming, and creating different geometries in a given application. A more advanced way to perform this exploration is to use a computational application, such as Excel or MathCAD, to perform systematic transformation. But the most powerful way is to take it a step further and use a computational software program, such as Bentley’s GenerativeComponents, to drive design alternatives and generate different design geometries. They then become the design alternatives to input into the cloud computing service.
Bentley Cloud Computing Scenario Services allow more design alternatives to be considered.
GenerativeComponents is very much a part of this optioneering design exploration. It allows users to perform funky architecture with curved structures and transforming geometries, and it is also applicable to industrial infrastructure where it can be used to investigate different proportions and scaling, and help determine what size of plant is warranted – longer, wider, taller, or narrower. While a program like GenerativeComponents can do the heavy lifting, it still leaves the decision making in the hands of the engineer.
One of the more interesting areas that is now being explored is how to access analysis and simulation results in the field. For example, when simulating pipe flow or HVAC flow, it is possible that simulation will mix computed analysis results and gathered results from sensors in the field. This mash-up of generated vs. gathered data is an area where mobile devices for information access in the field will be invaluable for decision support in an operational context. Seeing the status of things while in the field will continue to advance, and accessing and updating models in the field will soon be an everyday practice.
Another area that is creeping into our workflows is artificial intelligence and machine learning. Machine learning is described as having a lot of computing capability, which one can use to generate a lot of data. Mostly, it is clever algorithms for pattern matching in large datasets.
Microsoft’s Machine Learning Service is an example of this in action, which it recently released as part of its Azure cloud platform. By using these machine-learning services, it is easy to start performing analytics on the large datasets and analysis on the models to find patterns in the data. For example, users can look at their structures and be alerted when the patterns are outside the norm. The services can help determine if their structure is, say, much heavier on the upper floors and other insights that might affect design quality. This is an area where machine learning is a real asset.
All of these services and technologies are just the beginning. As cloud services become ubiquitous in our world organizations, owner-operators will reap the benefits by creating assets that are of a better quality and increase their ROI. It is at this point that the industry will look for the next great technology to increase revenue and create superior assets.