Today I would like to discuss with you current status and problems which I see in the industrial equipment simulation software area. First of all, I will briefly describe the problem as I see it. To highlight the details, I will tell you how engineers deal with the equipment simulations, what they call process, how different related to difficulty technological processes can be. We will take a look what currently happens with typical simulation software and its main restrictions. Also, I will describe a real example of gas compressed simulation software, which was implemented by me and show out of boundary way how it integrated into existing modeling structure. Last but not least, I will tell you why all of this is important and what possibilities I see in area of equipment simulation. So let's start with problem itself. There are a lot of industrial companies in various sectors of industry. Some sectors are more advanced, some of them are conservative and dogmatic, but all of them want to be effective and successful. To do so, they want to plan better. Better company plan better. It spends its resources, saves where it can save and spends only where it is necessary to spend. Any company want to understand its technology. Often industrial companies are big and complex with a lot of different people involved in operation, such as engineers and managers. Bigger companies have very complex and often interconnected technologies in operation, and clear understanding of the underlying technology is a key to companies successful operation. During last years, there was a clear trend in industrial companies to be more digital. Even companies from conservative areas such as oil and gas, for example, cannot ignore benefits which modern software can give to them. And often when such companies faces their challenges, sometimes which nobody encountered before, they find that they don't have proper instrument which can help them to solve their problems. And software tools, which they have, doesn't have enough features or such features cannot be directly used in their cases. In other words, industrial companies missing software tools. But let's go a little bit deeper to understand a problem. To get better understanding, first I have to explain based terms which used in engineering across industries. And I would like to speak about a thing we call process or sometimes technological process. In engineering. Process is a series of interrelated tasks, which takes something on its input, do some transformations and provides a desired output. Processes can be very different, but all of them follow this paradigm. For example, converter on car plant takes parts and assemble them into the car, breweries convert hop into the beer, chemical plants mix chemicals into medicines, and oil plants distillate oil into petrol. And all of these processes can be described and represented as mathematical models, which later can be implemented in related modeling software. But let's look what we can do right now. So in general, we are very good in mathematical modeling. We have a strong knowledge of mathematics, physics, chemistry and other sciences. It is not a problem for us to do any calculations with base physical parameters such as temperature, pressure, flow rates, energy, et cetera. We can do such calculation fast and without anything special required, for example, using software tools like Excel, especially if correlations between parameters are simple and linear. But also we are quite good in more complex mathematical models, especially in well described areas such as thermodynamics, hydrodynamics, or chemical reactions. We start to see some difficulties here, mainly with calculation complexity, when our process includes many interconnected parameters, or we have several interconnected processes which need to be simulated simultaneously, which actually leads us to things we model not so well. One of such example is industry rotary equipment. In my case, I would like to focus on a modeling process of such equipment as a compressor. If we speak about gases or pump in case of liquids, compressors and pumps are quite similar machines with common principles, but everything I say also applies to any other complex industrial rotary equipment. For better understanding why compressors are so difficult to simulate, I shortly will tell you about its structure and application. So the main purpose of pumps and compressors is to increase pressure of medium. Process of increasing the pressure is used in various applications, starting from quite simple air conditioning, where relatively small machines are used, and goes up to natural gas transportation processes, where series of huge machines which has physical size of typical country house do a lot of compression work. Such machines differ in structure and in size, which is measured in maximum consuming power, in watts. And since this is an active equipment which consume power to do work, it is always accompanied with some kind of engine, electric motor or gas turbine, for example. But if the main process of such machine is just an increasing of pressure, which is one simple physical parameter, why is it so difficult to simulate it in software? The answer here is variation. So due to nature of such equipment, the process cannot be stable every moment of time. There are a lot of parameters change every moment of time, and pressure is only one of them. Each model of such machine has its own unique fingerprint, a chart called performance curves. This is a chart which includes ranges of various parameters such as pressure ratio, flow rates and power calculated for specific impeller and for specific rotation speed. Change of rotation speed parameter leads to change of the wall map and all interconnection of parameters will also change, all of this multiplied by all possible variations in hardware designs. And do not forget that machines can run in parallel or in series influencing each other. So literally it is almost impossible to do universal software solution. But is it really that nobody tried to do so? Actually, not really. There are process simulation software exists and some of tools are actually very powerful, but usually not with rotary. They help engineers to do quite complex calculations. But I see one big issue here, flexibility. First of all, almost all such software are proprietary with closed sources. If you want change something in the way how it simulates your equipment, no, you cannot. Often such software do a lot of simplifications and also do not pay any attention to rotary equipment complexity. What companies can do, they can write messages to support with request to implement required changes and wait for next visionis version, sometimes for years. Or there is another option. Try to implement its own model and software and sometimes find very unintuitive ways how to embed it into existing modeling infrastructure. Here I want to show one example. I worked in company which had a complex simulation model of all assets in the one of its projects. There was no problem with simulation of any equipment except rotary. To do precise calculation we had to implement our own Python software which made a precise modeling of particular installed compressor type. I had to investigate what can be simulated in each part of the machine. Since it was a series of huge 16 mw machines driven by gas turbines with various impellers installed. I had to find proper mathematical models for each part as well as find their correlations. All this work was implemented in separate Python models which later was combined. At the end, my software were able to calculate precise fuel gas consumption per machine in all apparition range by approximation of performance curves, converting apparition parameters to required calorific value through mechanical and heat power and fuel gas consumption. The funny thing was that it was also possible to interconnect it with petroleum expert software through interface called Openserver from their site, which originally was used only to extract some reports. Our software led us to do precise prediction of compressor power consumption, which is actually a company loss on fuel gas. And actually this leads us to the question, why is it important? Actually the demand of such precision is questionable until you don't try to look for long term planning. There can be situations, especially on very big and long processes, when cumulative effect of nonoptimal operation can lead to significant losses of money later. For example, here is a chaos of typical oil and gas production profile. It shows how much product will company get during its operation time. On this chart, company plans to operate for 20 years, which is quite normal for typical oil and gas project. Blue line here shows overall production figures without any losses in real operation. Some of product is used for own consumption, for example for energy generation or as a fuel for compressors. Currently companies just take some fixed percent value and deduct it from overall production to get final figures. It is shown by orange line. The issue here is that condition of operation will definitely change a lot. And without precise modeling there definitely will be an error. The green line here represents simulated results and all the difference between green and orange lines on chart is actually potential money losses due to incorrect planning, which actually can be avoided. For the first view, the gap between lines is not so dramatic. But don't forget about cumulative effect and the scale. Often we are speaking about up to hundreds of millions dollars per year. What is important here? This effect can be seen already on the middle of the project life. And you can see this time point as blue vertical line on chart. Also there are another important time point exists in any such project. The moment when you have to stop your production so called abandonment phase. Any production in this phase cannot be profitable in any way. So you are interested to start the construction works as soon as you get to this point. Otherwise you just will lose money. Without precise simulation, you can miss this time point easily. It is clearly seen on the chart as it's shown by two red vertical lines. For almost a year, company's cost will be higher than profits in total. It's not only lost money, but also additional and avoidable spends. So in a nutshell, there are options. Companies can do simple use, conservative approach and mostly likely lose some millions in future. Or if they want to be safe, they can do investment in modeling software development now and save a lot money later. As a conclusion, I can say that I see a great potential in modeling software development. There are a lot of blank spots on this market and definitely more discussions are required. I personally would like to see more movement from industrial companies to open source world and vice versa. I hope this presentation as well as described experience can be as such step. So thank you very much.