When it comes to automating processes in the petrochemical industry, many people believe that this is a very complex industry, so everything should be automated through process control systems. In fact, this is not entirely true.

The petrochemical industry is indeed fairly well automated, but this concerns the main technological procedure, where automation and minimization of the human factor are of critical importance. The high cost of process control system solutions means that none of the related processes are automated: they are all performed manually. Therefore, every couple of hours, it is typical for an employee to manually check whether a pipe has been properly heated, whether the switch has been switched on, whether the valve is closed, or whether the bearing vibration is normal.

Most non-critical processes are not automated, but this can be done using the Internet of Things (IoT) instead of process control systems.
Unfortunately, there is a problem associated with inefficient communications between customers from the petrochemical industry and hardware developers who do not have customers from the oil and gas industry and, therefore, do not acquire information about the equipment requirements in terms of its application in aggressive environments, hazardous areas, harsh conditions, etc.
In this article, we will discuss this problem and how to resolve it.

To check some parameters, we conduct walkaround checks for a visual and tactile inspection of the processing unit's non-critical parts. One of the most common problems is associated with steam supply. Steam is a heat transfer agent that is used for a variety of petrochemical processes, and it is supplied from the heating plant to the final consumer through long pipes. However, our plants and processing units are located in rather harsh climate conditions, and winters in Russia are severe, so some pipes occasionally start to freeze.

Therefore, according to the regulations, walkaround checks must be conducted every hour to measure the temperature of the pipes. Taking into account the scale of the plant, there are a large number of people who actually do just this — walk around and touch the pipes.

Firstly, it is inconvenient: temperatures are very low, and you need to walk long distances. Secondly, this approach does not allow you to collect data on the process and, especially, to use it. Thirdly, it is expensive: all these people could do more useful work. Finally, you need to consider the human factor: How accurately is the temperature measured, and how regularly are measurements taken? 

And this is just one of the reasons why plant managers are quite seriously concerned about the problem of minimizing the impact of the human factor on technical processes. 

This is the first useful case of the potential application of the IoT in manufacturing. 

The second case is monitoring vibration. The vibration of electric motors on various equipment must be monitored. This procedure is also still done manually — people make rounds once a day to measure the vibration level with special devices to make sure everything is fine. So, once more, we are again dealing with wasting time and resources, and again the human factor affects the accuracy and frequency of the rounds. However, the main disadvantage is that you cannot work with such data because there is practically no data to process and it is impossible to switch to condition-based maintenance of dynamic equipment.

Today, one of the main industry trends is to move from routine to condition-based maintenance, which involves active and detailed accounting of equipment operating time and comprehensive assessment of its current state. For example, when the time comes to check the pumps, you check their parameters and see that Pump A has reached the necessary number of operating hours for maintenance, and Pump B hasn't, which means that it does not have to be serviced yet – it is too early.

In general, it's like changing car engine oil every 15,000 kilometers. Sometimes, this happens every six months, and sometimes the interval is a year or more, depending on how actively a particular car is used.

The same is true for pumps. In addition, there is a second variable that affects the need for maintenance: the history of changes in vibration indicators. Suppose that the vibration did not deviate from the norm and there are no negative trends; in addition, the pumping time is still insufficient — this means that maintenance is not needed yet. But if the vibration history causes concern, then this pump should be serviced even with insufficient operating time. And on the contrary, even with an excellent vibration history, we provide technical maintenance when the required number of operating hours is reached.

If you take all this into account and perform maintenance in this way, you can reduce the cost of servicing dynamic equipment by 20% or even 30%. Considering the scale of plants, these are very significant figures that are achieved without compromising quality or safety. And this is a case for using the IIoT in an enterprise.

There are also a lot of meters which are currently read manually (i.e. an employee has to go, look, and record the data). But it would be more efficient to take these readings online to monitor what is used and how in real time. Such an approach will greatly help in solving the issue of efficient energy resource use: with accurate consumption data, you can, for example, supply more steam to Pipe A in the morning, and heat Pipe B more intensively in the evening. After all, boiler houses are being built with a large capacity reserve in order to reliably provide all consumers with heat. But it would be better to solve construction issues more wisely and allocate resources more efficiently.

This is the popular Data Driven Decision Making approach, when decisions are made based on the results of full-fledged work with the data collected. Clouds and analytics are very popular today, and participants of the Open Innovations Forum this year talked a lot about big data and clouds. Everyone wants to work with big data, process and store it, but you first need to collect this data. Yet, these issues are not widely discussed. There are very few hardware startups.

The third case for using IoT is tracking staff, providing perimeter navigation, etc. We use it to track employee movements and monitor restricted areas. For example, if any work is being carried out in an area where outsiders are not allowed, you can track it in real time. Or suppose a mechanic went to check a pump, but he has been at the pump for a long time and is not moving — maybe there is a problem and the mechanic needs help.

Another problem is that there are no integrators ready to develop solutions for industrial IoT. This is because these technologies still do not have well-established standards.

For example, let's see what happens in a typical household: if you have a Wi-Fi router, you can buy something else for a smart home — a kettle, a power outlet, an IP camera or light bulbs — and connect it all to an existing Wi-Fi access point, and everything will work. Everything will work for sure because Wi-Fi is a standard and all vendors follow it.

However, for enterprise solutions, there are no similar prevalent standards. The component base itself has recently become relatively affordable, which has enabled devices using such a base to compete successfully with human resources.

If we look in more detail, the numbers will be approximately as follows. One industrial sensor for a process control system costs about USD 2,000. One LoRaWAN sensor costs RUB 3,000–4,000. Ten years ago, only sensors for process control systems were available, without any alternatives, and LoRaWAN sensors appeared about five years ago. However, we cannot simply take and use LoRaWAN sensors throughout our enterprises.

Home Wi-Fi and office equipment are quite understandable.

That said, there are no popular and commonly used standards for industrial IoT. There are, of course, a bunch of various industrial standards that companies develop for themselves.

Take, for example, WirelessHART, which was made by Emerson. It also uses 2.4 GHz, i.e. almost the same Wi-Fi. Coverage from point to point reaches 50–70 meters. And we can easily feel defeated when we consider that the surface areas of our plants exceed the size of several soccer fields. In addition, one base station in this case can only reliably serve up to 100 devices. For example, we are now equipping a new processing unit, and there are more than 400 sensors already in the initial stages.

And there is also NB-IoT (NarrowBand Internet of Things) provided by telecom operators. Again it is not intended for use in production — firstly, it is too expensive (the operator charges a fee for traffic); secondly, it is too dependent on telecom operators. If you need to install such sensors in a bunker where you can't get a mobile signal, then you need to install additional equipment. This means you will have to contact the operator who will charge an additional fee and will not give specific time frames for providing network coverage for the facility.
You will not be able to use "pure" Wi-Fi at production sites. Even domestic channels are too congested, both at 2.4 GHz and at 5 GHz, and we have a production site with a huge number of sensors and equipment, rather than just a couple of computers and smartphones in an apartment.

Of course, there are proprietary standards of good quality. However, this does not work when we build a network with a variety of disparate devices, because we need a unified standard that is not something closed, which again makes us dependent on one supplier or another.

Therefore, the LoRaWAN alliance seems to be a reasonable solution, as this technology is actively developing and, in my opinion, has every chance to become a full-fledged standard. After expanding the RU868 frequency range, we have more channels than in Europe, which means that network capacity provides us with the widest possibilities and makes LoRaWAN an excellent protocol for regular data collection, say, once every 10 minutes or once an hour.

Ideally, there are a number of sensors that should provide us with data every 10 minutes so that we can effectively monitor the processes and status of the equipment. But in the case of patrol personnel, this frequency is an hour at best.

What else are we missing?

There is no effective communication between hardware developers and customers in the petrochemical or oil and gas industries. As a result, IT specialists produce excellent equipment from their point of view, but it is not suitable for mass use in the petrochemical industry.

For example, there is a LoRaWAN-based device for measuring the temperature of pipes: you just need to install the sensor on the pipe, fasten it with a clamp, then fix the radio module next to it, and close the control point — that's all.

From the point of view of IT, the equipment is absolutely suitable, but from the point of view of the industry it has certain problems.

The battery capacity is 3400 mAh. Of course, the battery is not so simple: it is thionyl chloride, which gives it the ability to work at -50 degrees Celsius without loss of capacity. If such a sensor sends data every 10 minutes, it will discharge the battery in six months. If this is a one-off solution, then there is nothing wrong with that — you can remove the cover from the sensor and insert a new battery for RUB 300 every six months.
But what if you have tens of thousands of sensors at a huge production site? This will result in an extraordinary loss of time. After eliminating man-hours spent on walkrounds, we need the same time to maintain the system.

There is a fairly obvious solution to the problem — use a battery that doesn't cost RUB 300, but RUB 1,000 with a capacity of 19,000 mAh, which will only have to be changed once every 5 years. That would be fine. Of course, this will slightly increase the cost of the sensor itself. But industry can afford it, and it is truly needed.

Nobody does cusdev, so no one knows about the needs of the industry.

The main problem is a lack of communication — plain and simple. Any petrochemical plant is a hazardous production facility where local gas leaks and formation of an explosive cloud are possible. Therefore, all equipment must have explosion protection. It must also have appropriate explosion protection certificates in accordance with the Russian CU TR 012/2011 standard.

But the developers just don't know about it. At the same time, explosion protection is not a component that can be simply added to an almost finished device, unlike a pair of additional LEDs. It is necessary to redesign everything, from the board itself and the electrical circuit to wire insulation.

It's simple — we need to communicate. We are ready for a direct dialogue. I'm Vasily Ezhov, the owner of an IoT product at SIBUR. You can send me a private message here or to ezhovvs@sibur.ru. We have ready-made Terms of Reference. We will tell and show everything so that you will understand what equipment we need and why and what needs to be taken into account.

Right now, we are already implementing a number of projects using LoRaWAN in the green zone (where explosion protection is not mandatory); we will look at how this equipment works as a whole and whether LoRaWAN is suitable for solving problems on such a scale. As for small pilot networks, we really liked this technology, and now we are deploying a network with a high density of sensors, where we plan to install about 400 sensors for a single process unit. In terms of quantity, this is not much for LoRaWAN, but in terms of network density, this is already quite significant. So we will check it out.

At a number of high-tech exhibitions, I have been the first person to tell equipment manufacturers about explosion protection and its necessity.

So this is, first of all, a communication problem, which we want to solve. We are very much in favor of cusdev, as it is useful and beneficial to all parties: the customer receives the necessary equipment for their needs, and the developer does not waste time developing something unnecessary or completely redesigning existing devices from scratch.

If you are already doing something similar and are ready to enter the oil and gas or petrochemical sector, just let us know.