In the New Word section, Apparat talks about recently emerging terms related to the new society. This release features a digital twin. A computer that knows everything about you and can imitate your behavior.

Digital twin

The technological dream of the futurist and founder of the Acceleration Studies Foundation, John Smart, that special computer programs able to mimic the behavior of specific people. Using various technologies for collecting and analyzing your data, the computer will be able to answer your letters instead of you and even communicate with your relatives while you yourself are busy. According to Smart, digital twins should appear within the next five years.

It is assumed that in order to create a digital twin, special software will analyze your correspondence in social networks and e-mail, browsing history and purchases in online stores, information from wearable devices, smartphones or smart watch, and any other available information. Based on this data, with the help of special algorithms it will be possible to program your behavior: how would you answer your partner to a business letter, so that you would tell the children to their message on Facebook. Computers already know a lot about our preferences, for example, advertising companies analyze our search terms And emails, create a profile for each person and try to show him only those ads that he will be really interested in.

Software that imitates human behavior is also beginning to penetrate our lives: digital assistants - Apple's Siri, Microsoft's Cortana and IBM's Watson - communicate with the user, answer his questions, and can keep up the conversation even on abstract topics. The first chatbots have been developed that successfully pass the Turing test - that is, they mislead people who communicate with them and make them believe that they are not artificial intelligence, but real human.

Scientists are also considering more fantastic options for creating a digital twin: the complete digitization of the brain, the so-called consciousness upload. But work in this direction is now only in its infancy: for example, as part of the Blue Brain project, by 2023, a digital version of the neocortex, the main part of the human cerebral cortex, should be fully simulated.

How to use the digital twin

Talk to a person after his death

One of the most ambitious plans is to create such a double that can replace a person after his death. “When you and I die, our children will not come to our graves. They will go and launch our digital twins and talk to them,” says John Smart. “Let this scenario sound a bit far-fetched,” he adds. “But people are already setting up a wailing wall on the pages of deceased relatives on social networks and continue to send them private messages.” Such perspectives are loved by science fiction writers and directors. For example, one of the plots in the Black Mirror series tells how a young woman replaced her husband who died in a car accident with a digital copy. Later, she "uploaded" her husband's consciousness into an android robot - that is, she practically revived him.

Personal assistant

This option is much simpler to implement and does not require such high level cognitive abilities. To some extent, this is already being implemented, for example, a digital assistant Google Now analyzes your mail and search queries, providing hints that make your life easier. However, the digital twin can not only tell you something, but also take on some of your tasks, albeit quite simple ones: make an appointment with a doctor, set up a business meeting, and in the store point to products that are most suitable for your diet in terms of content. useful substances.

Project Disadvantages

One of the main shortcomings of this concept, which even its ideologue John Smart recognizes, is a complete violation of privacy. The program will read all your correspondence, analyze purchases and, in general, in every possible way penetrate into what is called personal life. Large corporations that collect less data are already facing protests.

John Smart

the ideologue of creating digital twins

You know, I would like to keep my health and financial information in a small safe so that no one can access it. But such thinking is an atavism. You can't get much if you don't sacrifice your privacy. I am sure that as long as people feel in control of the technology, data privacy will be secondary.

Image: Edward Blake Edwards

Neural networks, digital twins, artificial intelligence. Industry 4.0 technologies will change the oil industry beyond recognition

Architects of the digital age

Usually the most technologically advanced areas are considered to be information technologies and biomedicine. The attitude towards companies in traditional industries, such as metal rolling or oil production and refining, is quite different. At first glance, they seem conservative, but many experts call them the main architects of the new digital age.

Industrial giants began to automate production processes in the mid-30s of the last century. Over many decades, hardware and software systems have been continuously improved and complicated. Automation of production processes - for example, in oil refining - has moved far ahead. The operation of a modern oil refinery is monitored by hundreds of thousands of sensors and instruments, and fuel supplies are tracked in real time by satellite navigation systems. Every day, the average Russian refinery produces more than 50,000 terabytes of information. For comparison, 3 million books that are stored in the digital storage of the Russian State Library occupy hundreds of times less - "only" 162 terabytes.

This is the very “big data”, or Big Data, a flow comparable to the information load of the largest sites and social networks. The accumulated array of data is a unique resource that can be used in business management. But traditional methods of information analysis are no longer suitable for this. It is only possible to work effectively with such a volume of data with the help of Industry 4.0 technologies. In the context of a changing economic paradigm, a rich production “historical experience” is a serious advantage. Big data is at the heart of artificial intelligence. Its ability to learn, understand reality and predict processes directly depends on the amount of knowledge loaded. At the same time, industrial companies have a powerful engineering school and are actively involved in the introduction and improvement of new technologies. This is another circumstance that makes them key players in the "new economy".

Best of the week

Finally, domestic industrialists know the price of business efficiency. Russia is a country of great distances. Often, productive assets are located on far away from consumers. Under these conditions, it is very difficult to quickly respond to market fluctuations. Traditional technologies allow saving no more than a tenth of a percent. Meanwhile, digital solutions already today allow to reduce costs up to 10-15% per month. The fact is obvious: in the era of the fourth industrial revolution, those who learn how to most effectively apply new technologies in the context of accumulated experience will be competitive.

Petr Kaznacheev, Director of the Center for Resource Economics, RANEPA: “As a first step towards an “integral” artificial intelligence system in oil and gas, one could consider “smart” management and corporate planning. In this case, we could talk about creating an algorithm for digitizing all the key information about the company's activities - from the field to the gas station. This information could be sent to a single automated center. Based on this information, using artificial intelligence methods, forecasts and recommendations for optimizing the company's work could be made.

Digital transformation leader

Realizing this trend, the industrial leaders of Russia and the world are restructuring business processes that have been developing for decades, introducing Industry 4.0 technologies based on the Industrial Internet of Things, artificial intelligence and Big Data into production. The most intensive transformation is taking place in the oil and gas industry: the industry is dynamically “digitizing”, investing in projects that seemed like science fiction just yesterday. Plants controlled by artificial intelligence and able to predict situations, installations that prompt the operator for the best mode of operation - all this is already becoming a reality today.

At the same time, the maximum task is to create a system for managing production, logistics, production and sales, which would unite smart wells, factories and gas stations into a single ecosystem. In an ideal digital model, the moment a consumer presses the fuel dispenser, the company's analysts in the operations center are instantly informed about what brand of gasoline is being filled into the tank, how much oil needs to be extracted, delivered to the plant and processed to meet demand in specific region. So far, none of the Russian and foreign companies have been able to build such a model. However, Gazprom Neft has advanced the farthest in solving this problem. Its specialists are now implementing a number of projects, which should ultimately become the basis for creating a single platform for managing processing, logistics and sales. A platform that no one else in the world has yet.

Digital twins

Today, Gazprom Neft's refineries are among the most modern in the industry. However, the fourth industrial revolution opens up qualitatively new opportunities, at the same time presenting new requirements for automation. More precisely, it is not so much about automation, but about the almost complete digitization of production.

The basis of the new stage will be the so-called "digital twins" - virtual copies of refinery units. 3D models reliably describe all the processes and relationships that occur in real prototypes. They are based on the work of artificial intelligence based on neural networks. The "Digital Twin" can offer optimal modes of equipment operation, predict its failures, and recommend repair terms. Among its other advantages is the ability to constantly learn. The neural network itself finds errors, corrects and remembers them, thereby improving its work and the accuracy of the forecast.

The basis for training the "digital twin" is an array of historical information. Modern oil refineries are as complex as the human body. Hundreds of thousands of parts, tens of thousands of sensors. The technical documentation for each installation occupies a room the size of an assembly hall. To create a "digital twin", all this information must first be loaded into neural network. Then the most difficult stage begins - the stage of teaching artificial intelligence to understand the installation. It includes readings from sensors and instrumentation collected over the last few years of plant operation. The operator simulates various situations, makes the neural network answer the question “what will happen if one of the operation parameters is changed?” - for example, to replace one of the components of the raw material or to increase the power supply of the installation. The neural network analyzes the experience of past years and excludes non-optimal modes from the algorithm by calculation, and learns to predict the future operation of the installation.

Best of the week

Gazprom Neft has already fully “digitized” two industrial complexes involved in the production of automotive fuel — a catalytic cracking gasoline hydrotreatment unit at the Moscow Oil Refinery and a unit operating at the company’s oil refinery in Omsk. Tests have shown that artificial intelligence is able to simultaneously take into account a huge number of parameters of their "digital twins", make decisions and notify about possible deviations in work even before the moment when the trouble threatens to develop into a serious problem.

At the same time, Gazprom Neft is testing integrated solutions that will minimize the impact of the human factor on the scale of the entire production. Similar projects are currently being implemented at the company's bitumen plants in Ryazan and Kazakhstan. Successful solutions found empirically can subsequently be scaled up to the level of large refineries, which will eventually create an effective digital production management platform.

Nikolay Legkodimov, Head of the Advanced Technologies Advisory Group, KPMG in Russia and the CIS:“Solutions that model various components, assemblies and systems have been known and used for a long time, including in the oil and gas industry. One can speak of a qualitative leap only when a sufficient breadth of coverage of these models has been achieved. If these models can be combined with each other, combined into a whole complex chain, then this will indeed allow solving problems at a completely new level - in particular, simulating the behavior of the system in critical, unfavorable and simply dangerous operating conditions. For those areas where retooling and upgrading equipment is very expensive, this will allow pre-testing of new components.”

Performance Management

In the future, the entire value chain in the logistics, refining and marketing block of Gazprom Neft will be united by a single technological platform based on artificial intelligence. The "brain" of this organism will be the Performance Management Center, established a year ago in St. Petersburg. It is here that information from the “digital twins” will flow, here it will be analyzed, and here, based on the data received, management decisions will be made.

Already today, more than 250,000 sensors and dozens of systems transmit information to the Center in real time from all the company's assets included in the perimeter of the Gazprom Neft logistics, processing and marketing block. Every second, 180,000 signals arrive here. It would take a person just to view this information about a week. The Center's digital brain does this instantly: it monitors the quality of products and the quantity of oil products in real time along the entire chain - from the refinery outlet to the end consumer.

The strategic goal of the Center is, using the technologies and opportunities of Industry 4.0, to radically increase the efficiency of the downstream segment. That is, it’s not just about managing processes – this can also be done within traditional systems, but to make these processes the most efficient: using predictive analytics and artificial intelligence at every stage of the business, reduce losses, optimize processes and prevent losses.

In the near future, the Center should learn how to solve several key tasks that affect the efficiency of business management. This includes predicting the future 60 days ahead: how the market will behave in two months, how much oil will need to be processed to meet the demand for gasoline at the current time, what condition the equipment will be in, whether the plants will be able to cope with the upcoming load and whether them repair. At the same time, in the next two years, the Center should reach 50% capacity and begin to monitor, analyze and forecast the amount of oil product stocks at all oil depots and refueling complexes of the company; V automatic mode monitor more than 90% of production parameters; analyze the reliability of more than 40% of process equipment and develop measures to prevent the loss of oil products and the reduction in their quality.

By 2020, Gazprom Neft aims to reach 100% of the performance management center's capabilities. Among the declared indicators are the analysis of the reliability of all equipment, the prevention of losses in terms of quality and quantity of products, and the predictive management of technological deviations.

Daria Kozlova, Senior Consultant at VYGON Consulting:“In general, integrated solutions bring significant economic benefits to the industry. For example, according to Accenture, the economic effect of digitalization could be more than $1 trillion. Therefore, when it comes to large vertically integrated companies, the introduction of integrated solutions is highly justified. But it is also justified for small companies, as efficiency improvements can free up additional funds for them by reducing costs, increase the efficiency of working capital management, etc. ”.

Discuss 0

More and more companies are showing interest in the topic of digitalization of production. The organizers of the regional scientific and technical conference “Digitalization of production processes. The use of industrial software for building digital enterprises”, which was recently held in Samara.

It was initiated by the SMS-Automation group of companies, known as a universal integrator specializing in the creation and support of industrial automation systems, together with the Digital Production Department of Siemens, one of the world's largest concerns in the field of automation and electrical products, with which Samara developers have been connected by more than two decades of fruitful cooperation.

The forum of manufacturers and developers of information systems was also supported by the Ministry of Industry and Technology of the Samara Region. Its specialists have repeatedly noted the success of the group of companies in the field of industrial automation and the construction of large information systems.

Representatives of industrial enterprises of the Samara region were introduced to the conceptual foundations and specific tools for building effective digital production. Industrial automation is only part of digitalization, or digitalization, as it is also called. Digitalization is the automation of processes throughout the life cycle of a product, equipment, enterprise. The project, and its functioning, and modernization fit into it.

The report of Andrey Sidorov, Chairman of the Board of Directors of the SMS-Automation Group of Companies, "Industrial Software as a Tool for Digitalization" aroused great interest among the conference participants. “We are on the verge of intellectualization of control systems,” said Andrey Sidorov (on the bottom photo). - Now equipment manufacturers in the West are changing the production model. Equipment begins to have a digital twin. The change in the business model will lead to the fact that a digital twin will be a significant factor in choosing a supplier.”

Digitization is, among other things, the development of situations on virtual digital models, which allows you to save huge amounts of money. Siemens is already at its digitalization site, without waiting for the arrival of a machine for the production of parts, having received its virtual image, it connects virtual robots to it and starts debugging technological processes without wasting time.

The topics raised by the experts related to the use of specific digital production tools were perceived with interest by the conference participants and raised a lot of questions and discussions. In addition to the reports, the attention of the conference guests was attracted by demo stands with practical examples of the implementation of the principles of digitalization in the realities of industrial process control systems in Russia. Special attention at the conference was given to the issues information security modern systems automation. Acquaintance with current trends in the development of enterprises within the framework of the Industry 4.0 concept, according to experts, can become additional tool in the process of increasing competitiveness in the era of "Industry 4.0".

In Russia today it is difficult to talk about the 4th industrial revolution, but we believe that it is necessary to speak. Among the technological drivers in industrial enterprises in the new generation, there will be platforms for the industrial Internet of things that implement the concept of a digital twin.

Forrester analysts define a digital twin as the creation of a real physical object in an abstract digital form that acts as an intermediary for any connection to a real device.

According to General Electric, the idea behind the digital twin is to go beyond just working with digital models. The company says that Maintenance will also occur in synchronization with the digital model with the real object through sensor systems and communication facilities.

Analyst agency Gartner predicts that by 2021, half of large industrial companies will use digital twins and, as a result, these organizations will receive a 10% increase in operational efficiency.

“Digital twins are driving the business impact of IoT by offering a powerful way to monitor and manage assets and processes,” said Alfonso Velosa, research vice president at Gartner. This is especially exciting for our team, since we in the SAYMON project are very closely involved in automated monitoring and control, including information systems and internet things. Of course, the competition in the IoT management platform market is quite strong - literally every large digital corporation today claims to have platforms, but not everyone has managed to make their own developments or acquire a company with a ready-made solution. Often a statement of availability is a tribute to decency - there is a technological trend, there is a statement of a corporation.

Today we do not work with digital models and drawings yet - we are open to partners with experience in this field. On this moment there is experience of cooperation with a company that forms photo-realistic copies of industrial facilities and as a result, a separate project VIOTR was born, combining the power of digitized space with the ability to obtain data from real sensors and video cameras, the ability to control switches, relays and dampers in the real world. The VIOTR project today has a focus on the educational technologies of the future, but in essence it is part of the concept of a digital twin.

This is exactly what our colleagues from Computer Weekly put it - a new approach involves managing communication between edge devices and internal systems and mirror reflection changes in the virtual model of the device - in other words, a digital twin appears.

The examples show that even for such simple operations as controlling door locks, you can get significant savings in operation. Dormakaba, which makes smart door locks, has been using ServiceMax field management software since 2012 to help it control its installations. Detailed data on the operation of each door helps Dormakaba and its partners manage buildings more efficiently. A recent Vanson Bourne study for ServiceMax found that industrial companies are losing $260,000 an hour due to unplanned downtime. Predicting failure using digital twins can help overcome this problem. The digital twin can provide the engineers at Dormakaba with an up-to-date record of every action or event that the sensors on the doors register, log the installation of components and firmware updates, and can be used by the Dormakaba service team to determine the life of the product along with detailed description a security log that is tied to the door. It is also important to ensure close connection with suppliers of parts and components and management life cycle product, providing an extremely precise level of control and service. By using digital lock prediction, Dormakaba expects to reduce customer calls and improve service quality. Together with Swisscom, a cloud platform for managing locks was created. Partner training is an important element of this innovation and business transformation, Dormakaba acknowledges.

In a Gartner report Digital Twins Will Impact Economic and Business Models, the analytics firm draws an analogy between the amount of data collection done by companies like Google, Amazon and Netflix and the amount of data digital twins in industrial firms will create to constantly monitor the operation of equipment connected to control systems.

Analysts warn that this will require even more control over components and software upgrades, and will require car manufacturers to become software vendors. “Asset operators will need to add skills to work with software tools to their operating divisions as they add smarter assets as well as add ownership software and data into support contracts,” analysts warn.

Image from https://www.ge.com/

More recently, German Gref, president of Sberbank, said that in 5 years artificial intelligence will replace many people: 80% of decisions will be made by machines, and this will lead to the fact that tens of thousands of people will lose their jobs.

Machine learning expert and artificial intelligence Pedro Domingos goes even further: he suggests that people will acquire a computer psychological model of their personality. What will she be?

Sex, lies and machine learning

The digital future begins with the realization that when you interact with a computer, whether it be your own smartphone or a server thousands of miles away, you do it on two levels every time. The first is the desire to immediately get what you need: an answer to a question, a product you want, a new credit card. At the second level, strategic and most important, you tell the computer about yourself.

The more you teach him, the better he will serve or manipulate you.

What model of your personality do you want to offer to the computer? What data can be given to him so that he builds this model? These questions need to be kept in mind whenever you interact with a machine learning algorithm - just like when interacting with people.

digital mirror

Think about all your data that is stored in all the computers in the world. These are emails, MS Office documents, texts, tweets, Facebook and LinkedIn accounts, web search history, clicks, downloaded files and orders, credit history, taxes, phone and medical records, driving information recorded in the on-board computer of your car , a map of movements registered by your mobile phone, all the pictures you've ever taken, brief appearances on security camera footage.

If the future biographer had access only to this “data exhaust” and nothing else, what picture would he have? Probably pretty accurate.

Imagine that you took all your data and gave it to the real Master Algorithm of the future, which already has knowledge about human life that we can teach it. He will create your model and you can carry it on a flash drive in your pocket. Of course, it will be a great tool for introspection - how to look at yourself in the mirror. But the mirror would be digital and would show not only your appearance, but everything that can be learned by watching you. The mirror could come to life and talk.

Benefits of the digital twin

What would you like to do, what tasks to entrust to your digital soulmate? Probably the first thing you would want from your model is to instruct her to negotiate with the world on your behalf: release her into cyberspace so that she will look for all sorts of things for you.

Of all the books in the world, she'll recommend a dozen that you'll want to read first, and the advice is so deep that Amazon couldn't even dream of. The same will happen with movies, music, games, clothing, electronics, whatever. Of course, your refrigerator will always be full. The model will filter your email and voice mail, news on Facebook and updates on Twitter, and when appropriate, reply for you.

It takes care of all the annoying little things of modern life, like checking credit card bills, appealing wrong transactions, planning schedules, renewing subscriptions, and filing tax returns. She will select the medicine for you, check with your doctor and order it in the online store.

The model will tell you who you like. And after you get to know each other and like each other, your model will team up with the model of your chosen one and choose restaurants that both of you can like. And this is where it gets really interesting.

Model Society

In the very fast approaching future, you will not be the only person with a "digital half" who runs your orders around the clock. A similar model of personality will appear in everyone, and the models will communicate with each other all the time.

If you are looking for a job and a company has X employees, then its model will interview yours. Their "conversation" will in many ways resemble a real, "live" one - your model will be well instructed, for example, it will not give out negative information about you - however, the whole process will take only a fraction of a second.

In the world of the Master Algorithm, "my people will contact yours" will become "my program will contact your program." Each person will have a retinue of bots, designed to make their way around the world easier and more enjoyable. Deals, negotiations, meetings - all this will be organized before you lift a finger.

Your digital soulmate will be like a power steering: life will go where you want it, but with less effort on your part.

This does not mean that you will find yourself in a “filter bubble” and only see what you are guaranteed to like, without any surprises. The digital person will be much smarter, they will be instructed to leave room for chance, let you touch new experiences, look for happy accidents.

As the models improve, the interaction will become more and more similar to what would happen in the real world, but it will happen in silico and a million times faster. The cyberspace of tomorrow will turn into a very vast parallel world, which will choose all the most promising to try in reality. It will be like a new, global subconscious, the collective "Id" of humanity, or "It".

Today's world is notable for the fact that theories of mind have begun to appear in computers. So far, these theories are still primitive, but they are developing rapidly, and we will have to work with them no less than with other people to get what we want.

Based on the materials of the book "The Supreme Algorithm"