Pages: 1
May 28, 2020, Waterflood diagnostics with advanced well-testing and well interference studies
Topic: Waterflood diagnostics with advanced well-testing and well interference studies

Waterflood can be tricky when dealing with complicated reservoir - compartmentalized, stacked, heterogeneous etc. We will go through several field cases to illustrate how advanced reservoir surveillance - single and muliple well testing - can help understand the real inter-well communication pattern and adjust waterflood for better efficiency.

Date: May 28, 2020

Presenter: Vladimir Krichevsky
Company: Sofoil
Questions and answers after webinar.

1) Question: In this case above, how many cells of MRT are needed to be able to suggest fine-tuning the waterflood?

Answer: It was a field-wide analysis, so we took all the wells in the field. We only see one deposit but in fact, we had I believe about sixty MRT cells here but, in most cases, we would just focus on the problem areas, so it is possible to give some recommendation for only the problem areas of the field and then the time of the analysis and of course, the cost would be much lower

2) Question: What represent the numbers into boxes on the plot?


The numbers inside the boxes are the formation pressure around producing wells in bars, we see about 200 bars in the southern part, up to 270 in the center part, but this is caused by the water breakthrough from well number X05, and in the northern part  - down to 70 bars and this is the area of the most pressure drop.

The initial pressure here is about 280 bars.

3) Question: As you explained the producers were converted to injectors and some oil produced was due to unswept oil. What year did the waterflood start?

Answer: The pay is quite thin here, so I think it's about eight meters and the well spacing is quite tight also the permeability would say it's low but considering many Russian fields it's quite normal it's about 20 or even 40 millidarcy. Once we start to inject in this well we had a response in only about a month, so it's a quite fast reaction which we were able to see you in no time

4) ​Question: Do we need shut in when using PCT?

Answer: Number one priority is to create a big enough pressure disturbance in a generating well and sometimes the only way to do it is to shut in the injection well. But sometimes – and it’s defined by fluid mobility, distance, PVT properties, we can perform a series of increases/decreases that would mean that there is no need to shut neither generating nor receiving wells. If we get an expected formation properties, expected PVT and SCAL we can pre-model the test to understand at least the order of magnitude of our pressure response, and if we can register this response with downhole gauges we can afford in some well.

5) ​Question: ​Did you conduct any tracer survey to get an idea for an optimal pattern?

Answer: The results of pressure interference studies MRT and PCT are often compared to tracer studies and this is not strictly correct because the tracer is only telling you if you have any water in this well coming from that well or if you have any communication between the wells at all. If you get water from the producer in your injector then they are indeed communicating but the amount of this communication and is it a good or bad communication is often unclear with the tracer result. So to summarize in many cases you will see that if you don't get any tracer in a well but still these wells are closely connected because there is another well between them for example and this could be a producer or an injector and you would not get any tracer in an offset producer but still when you shut this injector you would see a pressure drop in the producer. So this means that pressure response does not work like a mirror of the tracer results so, these are technologies that complement each other, and we need to know what can be achieved with different methods here. We had tracer studies which were run right before our test and the main conclusion we had have been confirmed with tracer studies.

6) ​Question: Any specific reason most of the generating well i.e. injectors are located at crestal location?

Answer: It does not matter, it was just a coincidence here. We have many interesting cases of both technologies, some of them are quite recent, and they are not published yet and unfortunately due to NDA pictures can't be shown at the moment, but they are soon to be released to the public, and then we can run another webinar where we will see a set of other interesting cases. So, in this case, it was a pure coincidence that injectors of PCT are located in the crestal area; it could be an injector somewhere outside the aquifer zone, it only depends on the distance and the mobility and is a matter of predesign.

7) Question: Could the generating well be a producer?

Answer: Yes it can. The idea is to generate pressure disturbance so if we don't have any injector for example, and we can afford to reduce the rate of the producing well than we can make a producing well a generator. We would perform a series of reduction/increment operations and in such way that we can generate the same pressure disturbance, and then we can get pressure response in the offset wells. So, technically we can run the PCT in the field where there is no waterflood at all (no injection well at all).

8) ​Question: Do you recommend conducting the production logging before or after the PCT test?

Answer: In most cases to properly design the test it is better to have results of production logging beforehand, but sometimes due to the operational issues you may find yourself restricted to production logging operation and sometime you would perform production logging after the test, it would only mean that during the PCT design we will have an additional risk of for example thief injection in the generating well. And it would mean that we need to put some safety to the design and overall test duration would be longer, but we would be certain that we’ll get results even if we have some amount of crossflow injector and only after we run production logging to find out what is going on and interpret results considering the effective net pay and injection.

9) ​Question: How do you design the period (i.e. how long it takes) say for example right now for injector to perform the series of injection and shut in operation and how long does it usually take?

Answer: In most cases, we would just create a model, it would be a simple model in some cases, it could even be an analytical model, sometimes it would be 2D numerical, sometimes when the field is quite complicated we need to go through 3D modelling, and then we just pre- model the test, so we create this pressure disturbance in the model and see what kind of response can we receive, and what kind of gauges we need to decompose this response considering the type of noise we have in the receiving wells.

10) ​Question: I think it's worth making a dedicated presentation on what one can see from PDG against what is observed later from PLT and/or workovers in the same or offset wells.

Answer: It could be made into a webinar; it would be rather short, but it could cover this important topic.

11) ​Question: In a field of single injector-multiple producers, how many OP wells do we need to monitor the response during PCT test?
Answer: The area which would test is defined not by the technical possibilities of PCT but by the issues and problems you have in your field so if you know the issue of the field, then we may work together to understand which well to include, which minimal quantities of wells we need to include in the test program and then understand what can we can achieve and how.

12) ​Question: How is this method different from CRM method?

Answer: This is a wide topic and deserves its webinar. We also build CRM models, and we understand perfectly the difference. If we talk about CRM vs MRT and they are quite similar in a way that they take similar input, the history of pressure and the rate, and try to build the interference model, the CRM is a simpler version of MRT, this once again asks for another webinar because we have now work in progress that compares results of both technologies. But the bottom line is that CRM outcome is a lot poorer than MRT outcome and in most of CRM model realizations we can't even understand the interference between producing wells, where MRT we can run on a field with no waterflood at all, so it only produces, while CRM will not work with this type of fields, so we will build a complete matrix of cross-well interference covers early, middle and late times and also proving all the same results besides which the CRM will get you.    

13) ​Question: ​If all wells are equipped with PDG, can the same analysis to PCT be performed?

Answer: Yes. It is perfectly possible, there will be a need to see what is the pressure resolution of your PDGs and then when we will design the test; we will try to fill restrictions that this pressure resolution we give us, on PCT test we are trying to use our sister company’s gauges which have a very high resolution of around 20 pascals as of now which is very precise but yes of course if you have existing working PDG's then it would be a lot better if we run the PCT on your PDG’s but it could mean a slight increase in the test duration. But considering that there is no need to perform a series of running and pulling out of the hole operation – a little increase in duration seems to worth it.

14) Question: I have a plan to design an interference test to find the value of communication between wells we have SCADA system

Answer: We need to see all the data, we can further discuss it on a forum I think there is a link to the forum just below this video or if not, we will just post it in no time and I think we can further discuss this question because I think this one is pretty specific

15)  Question: Is pressure disturbance is still valid for the tight reservoir with 1-2md?  

Answer: Yes, of course, it is because again the amplitudes of the pressure impact are not defined by the permeability only, you should also consider the viscosity and the compressibility and of course the distance between the wells, also it strongly depends on the time of the test so when we inject and then shut in, the longer is this pulse the higher response you get. In one of the PCT cases that I showed you today, we had the dynamic permeability of only one and a half millidarcy.

16) Question: What is the typical duration of the test to get good data from PCT? And what factors are affecting the required duration?

Answer: There are many factors, I just talked about which define the time of the test and the typical duration would be I would say from 10 days to in some cases up to 3 months but if we're talking about typical time - I think it would be about one month long; but you know it's like measuring the average temperature of all the patients in a hospital, so each case is unique, and we need to consider it individually.

17) ​Question: How was the well-spacing initially optimized as it seems that it is not consistent everywhere? Which have caused poor Ev in some wells (as shown in your plot)?

Answer: Of course, I have seen a picture a while ago of a reservoir engineer standing in front of the wall with two paintball guns: one of them is red and another one is blue. The red gun is called producer gun and the blue one is called injector gun, so he is just standing and firing into the wall and thus he is forming the development system. So yeah, I think this field in a way seems like a result of such actions but as a matter of fact, while drilling, while getting more and more information - the development system was being adjusted to this new information and eventually it came looking like this.

18) Question: Any comparison of MRT with RTA?

Answer: I think it would also be a nice topic for another dedicated webinar where we would compare our MRT to both RTA and CRM but briefly I would say that the RTA is a straightforward method.

So when you perform rate transient analysis in many different softwares you can run a numerical analysis where you add offset wells and you try to model the response of your focus well. That works pretty fine but this is a straightforward calculation so it means that you have to model it and if you don't get a good match you change the parameters, you model it again; and you are free to change many parameters like porosity, permeability, thickness, if you have fracks in you well - you can change the direction of this fracs and all this would affect what you see as a result.

MRT is a reverse calculation so it's a reversed task solution, once we already have the data - we are trying to fit it automatically, this is the beauty of MRT - we have a very complicated problem in the field of diffusion modelling, we convert it into a mathematical field, then we decompose it into simple tasks like which pressure response will I get when I launch the well with the unit rate; which response will I get in another well when I launch this world with the unit rate,  this is a simple problem, and we when we go back with this problem into the diffusion modelling we will be able to solve it in no time.

19) ​Question: What if gas is already elaborated in reservoir condition, considering the high magnitude of gas compressibility will it affect response?

Answer: Yes, of course, it would affect your response and we need to properly pre-model it because in some cases it could create some problems for us, we need to understand how many free gas do we have in a layer and where is it situated and then we may pre-model what we will get as a response because of course the secondary gas cap will affect the calculation and it will affect the test result.
Pages: 1
To participate in the forum you must Register or Login