作者：高级编辑 Stephen Whitfield

Ryan Dawson 是 Corva 的首席 Corvanaut（创始人兼首席执行官）。

从一家技术驱动型公司的角度来看，您认为钻井承包商对创新的接受程度如何？

我认为钻井承包商越来越乐于接受创新，但他们仍然相当务实。当他们看到新技术能够明显提高安全性、效率或结果的一致性时，他们的接受度就会提高——对新技术的信任真正建立在经过验证的现场结果上，而不是空洞的承诺上。

似乎很多公司都不愿意成为新技术领域的先行者。当您与钻井工人和运营商谈论Corva的产品时，您认为这种犹豫会构成多大的障碍？或者您甚至认为这是一种障碍？

真正的挑战不在于技术，而在于激励机制的协调。钻井承包商身处一个成熟的行业，利润空间狭窄，且提高运营商收费的能力有限，这使得创新难以获得认可。

Corva之所以与众不同，在于我们与Nabors的合作，我们始终将效率作为价值驱动力。Nabors极具远见，围绕数据和自动化构建企业文化，利用技术减少非生产性时间，并大规模提升执行的一致性。通过这种合作，创新能够带来可衡量的效率提升，从而实现自身价值，并与钻井行业的经济现实相契合。

所以，没错，没人想抢占先机。但当他们看到某个方法奏效时，都会迅速跟进。运营商会发现一些机会，然后说我们需要快速行动，接着事情就会进展起来。之后，大家都在创新，而且没人会质疑为此付出代价。

这真是一个有趣的处境。你需要用新技术取得成功才能吸引别人的注意，但你却无法在没有别人愿意尝试的情况下取得成功。如果你还没有在这个领域积累过初步经验，你该如何向公司证明你的技术是可行的呢？难道只能靠运气吗？

是的，早期总是需要一定程度的冒险。我们经常进行这类讨论，特别是与钻井承包商，他们往往行事非常谨慎。考虑到风险和运营上的风险，这可以理解。而纳伯斯公司则愿意尽早行动，并将技术真正投入到现场应用中。

另一方面是成熟度。Corva 已经成立 11 年了，我们在这段时间里一直在夯实基础。围绕数据、可靠性和集成的大部分繁重工作已经完成。我们现在正处于一个不再是实验阶段的成熟阶段。系统是开放的、互联的，并且已经准备好扩展。

所以，当一家公司帮助实现预测性钻井之类的技术时，这不仅仅是证明某个工具有效，更是搭建起其他人可以在此基础上进行开发的基础设施。一旦基础设施搭建完成，任何第三方都可以加入并添加他们自己的功能，而这才是这项技术真正开始腾飞的开始。

在讨论钻井领域的创新和技术时，“信息孤岛”一词经常出现。石油和天然气行业的一些人认为，竞争能够促进创新，而对技术和理念过于开放未必是理想的选择。您在工作中是否也遇到过信息孤岛带来的挑战？当您遇到这种挑战时，您是如何克服的？

信息孤岛仍然是我们面临的一大障碍。它们限制了我们获取背景数据的途径，并减缓了学习进程。

我们正努力打破信息孤岛，整合来自多个来源的实时数据，并提升团队间的透明度。举个例子：目前，运营商正优先考虑自动化时间日志录入。在很多钻井平台上，人们仍然需要手动将这些信息录入日志。虽然有些钻井平台的人员非常擅长记录非生产时间（NPT），但绝大多数钻井平台并非如此。他们什么都不写，因为这对他们的业绩不利。因此，我们看到自动化进程正在加速推进。

在Corva，我们正在做的一件事是利用钻井平台上的传感器生成时间日志数据，然后自动将这些数据输入数据库。之后，我们可以添加钻井承包商的注释。也许公司代表会在一天结束时补充一些他们认为重要的背景信息，但数据本身已经录入完毕。

我认为这类事情非常重要，因为我们可以使用的逻辑逻辑模型和文本挖掘算法的基础部分将基于这些时间日志条目。如果性能数据能够自动录入，那就意味着我们即将打破许多信息孤岛。

你认为这个行业距离真正打破这些壁垒还有多远？

我不知道答案，但我认为打破部门壁垒是Corva的使命。我觉得我们在这方面处于领先地位。但我这个人比较乐观。一家服务公司可能会来找我说，我们可以做这件事，但一年后情况就完全不同了。

我们与压裂、电缆测井和地质勘探等众多服务提供商以及钻井承包商建立了合作关系。这是我们目前最成功的举措之一。过去20年来，大家一直在询问如何整合所有这些公司和信息。我们认为，这才是未来的发展方向。

在这个行业，创新和竞争优势之间似乎总是存在着一种平衡。在Corva，你们开发新技术，这些技术可以造福整个行业，但你们仍然是一家努力维持市场份额的企业。您如何看待自己的工作在这种平衡中的位置？您对竞争优势有多担忧？

我们提供预测性钻井服务的时间越长，就越意识到总有新的领域可以让我们去探索。你会发现自己还有很多未知的东西。像这样的系统包含太多不同的组成部分。

一是钻机控制。我们需要这台钻机钻井过程像黄油一样顺滑，并且要避免任何可能发生的故障。这些问题本身就相当棘手，需要丰富的经验和反复的迭代。为了避免故障，必须在一定的限制范围内操作；如果能够将这些限制自动化，使其越来越严格，就能获得更大的压差，从而实现更顺畅的钻井。这是一项复杂的工作，我认为目前还没有其他人做到，它需要大量的努力和迭代，但这是我们已经集成到系统中的一项功能。

这一切都说明，这个行业有很多难题需要解决，我们总会遇到一些问题。

正如您之前提到的，Corva在过去几年里一直与Nabors保持着合作关系，两家公司宣布将于2025年将合作扩展到Nabors的RigCLOUD平台。在将您的系统与他们的技术整合的过程中，您最初是否遇到过任何挑战？

我认为最初集成方面的挑战主要集中在架构的协调和确保低延迟数据流上。

在尽可能接近实时地获取数据方面，您认为存在哪些挑战？

我认为主要问题在于，包括我们公司在内的许多公司都有自己开发的系统，而这些系统的集成并非总是顺畅无阻。对于某些 WITSML 提供商，我们的系统会因为构建在云平台上而导致服务器过载甚至崩溃。当我们需要在云端增加服务器空间时，只需添加即可。但如果要与物理服务器上的系统集成，那就困难得多。现在，我们有专门的 DevOps 工程师负责监控并调查每一个异常情况。这些实时关键任务系统容不得半点延迟。

例如，如果你向 API（应用程序编程接口）发出请求并收到数据，你可以立即关闭请求，这当然没问题。但如果你出于某种原因想要保持连接打开——比如你想基于该请求执行某些操作，稍后再返回该接口——那么当你返回时，90% 到 95% 的时间都花在了确保连接仍然安全上。这累积起来相当可观。

很遗憾，这是我们不得不面对的事情。

听起来您是开源API的拥护者。您是否对业内日益增多的关于互操作性的讨论感到鼓舞？

我深受鼓舞。我乐观地认为我们都在朝着这个方向前进。只是有些人走得比其他人快一些。

你认为这个行业在这方面的进展速度是否达到了应有的水平？

不，我认为它需要更快。我刚入行的时候，看到WITSML，觉得它简直糟糕透了，这完全是因为我之前是软件工程师。但现在，我意识到它就是这样。它能用。它是一个开放标准，而且人人都知道。我认为它是真正推动这个行业发展的关键因素之一。我注意到，例如在压裂领域，人们在这方面有所落后。

关键在于与标准制定机构合作。如果你是像微软或谷歌这样的大公司，你可以采用标准并添加自己的定制功能。我们目前还没有看到足够的此类合作——也就是说，我们没有看到企业采用 WITSML 1.4 并添加这些功能来改善世界。我们需要加大力度推动这项工作。

您提到使用WITSML处理压裂数据。这让我想到另一个问题：Corva提供预测钻井服务，但该公司也大力宣传该系统在完井作业中的应用，以及它如何帮助企业从被动完井转向预测完井。您如何定义预测完井？

预测性完井——或者我们这里称之为预测性压裂——是一种长期策略，包括直接连接设备。我们目前还没有这项能力，但我们正在研发一种叫做引导式压裂的技术。我们希望实时导入所有泵的运行计划并主动监控泵的运行，这与其他服务公司的做法类似。

我们所做的事情与以往不同的地方在于，我们希望实时叠加过去的阶段，这样你就可以看到，例如，压力响应与过去阶段的比较情况。

我们也在努力预测压力响应，并围绕异常检测构建了许多功能。一旦检测到压力峰值，我们就能迅速做出反应，系统将能够生成警报，并添加标签和注释，以便压裂作业人员稍后进行查看。

我们之前已经稍微谈到了数据，您也谈到了行业需要更好地最大化利用现有数据。您认为像Corva这样的公司在哪些特定领域最能帮助行业以一种前所未有的方式理解数据？

我们提供所有这些分析，但从我们的角度来看，我们并不是那种会主动帮你找到你最需要知道的那件事的公司。如果你刚打了一口井，你需要从这口井中了解哪三件事？你从中吸取了什么教训？这些都是客户需要决定的。

但我想说的是，我们已经转变了思路，不再只是简单地提供所有分析结果然后就进入下一阶段。借助 Corva，我们开发了大约 200 个应用程序，可以帮助您根据需要对数据进行尽可能精细的提炼。但这最终还是要回到弄清楚哪些信息真正重要，哪些信息能够创造价值的问题上。

当你和业内人士谈论自动化时，话题似乎总是会回到钻井平台上人的作用上。没有人希望完全取消钻井平台上的人。从科技公司的角度来看，人的作用在你们的决策中占据多大比重？

人是创新速度的核心。从技术角度来看，我们可以快速开发产品，但真正的进步发生在产品能够自然地融入人们的工作方式时。这就是用户体验至关重要的原因。界面至关重要。如果它直观易用，并且符合现场工作流程，那么一切都会更快。

目前，我们正在大力推进方向性和地理位置整合，例如将曲线直接融入我们的界面。技术已经成熟，演示效果也很棒。现在的重点是确保它能以一种流畅自然的方式呈现给用户，让他们在日常使用中感到舒适。我们将不断努力改进这项技术在用户面前的呈现方式。

更广泛地说，我们看到自动化和人工智能正在改变钻井平台上人类的角色，而不是取代他们。人类的角色更多地转向监督、策略制定和决策。

自动化的目标是增强而非取代人类的专业技能，从而使人们能够将更多时间用于运用判断力、经验和创造力等关键领域。

By Stephen Whitfield, Senior Editor

Ryan Dawson is Chief Corvanaut (Founder and CEO) at Corva.

Coming from the perspective of a technology-driven company, how accepting do you feel drilling contractors are of innovation in general?

I think drilling contractors are increasingly open to innovation, but they’re still pretty pragmatic. When they can see new technology that clearly improves safety, efficiency or the consistency of outcomes, that acceptance rises – trust in new technology really builds through proven field results rather than promises.

It seems like many companies don’t want to be the first mover on new technology. When you’re talking with drillers and operators about Corva’s offerings, how much of a barrier do you find that hesitancy to be? Or do you even see it as a barrier?

The real challenge is not technology. It is incentive alignment. Rig contractors operate in a mature business with tight margins and limited ability to charge operators more, which makes innovation difficult to justify.

What differentiates Corva, through our partnership with Nabors, is a deliberate focus on efficiency as the value driver. Nabors had the foresight to build a culture around data and automation, using technology to reduce nonproductive time and improve execution consistency at scale. Through that partnership, innovation pays for itself through measurable efficiency gains and aligns with the economic realities of the drilling business.

So, yeah, no one wants to be the first to move. But when they see things working, they all jump on it quickly. Operators will see something and say we need to move fast on this, and then things move. Then everyone’s innovating, and no one has a question about paying for it.

That’s an interesting position. You need success with new tech to gain interest, but you can’t get that success without someone willing to give it a try. How do you show a company that your tech is viable if you haven’t had that initial work in the field? Is it just a leap of faith?

Yeah, early on there’s always some level of leap of faith. We’ve had these conversations a lot, especially with rig contractors, and they tend to move pretty carefully. That’s understandable given the risks and what’s at stake operationally. With Nabors, there was a willingness to step out early and actually put it to work in the field.

The other piece of it is maturity. Corva’s been around for 11 years now, and we’ve spent that time building the foundation. A lot of the heavy lifting around data, reliability and integrations has already happened. We’re getting to a point where this isn’t experimental anymore. The systems are open, they’re connected, and they’re ready to scale.

So when a company helps enable something like Predictive Drilling, it’s not just about proving one tool works. It’s about putting infrastructure in place that others can build on. Once that’s there, any third party can come in and add their own capabilities, and that’s when it really starts to take off.

The word “silo” often comes up when discussing innovation and technology in drilling. There are some people within the oil and gas space who say that competition breeds innovation and that it is not necessarily ideal to be too open with technologies and ideas. Have you found silos to be a challenge in your work? How do you overcome that challenge when you see it?

Silos remain a major barrier for us. They limit access to contextual data and slow down the learning process.

We’re trying to break silos by integrating real-time data from multiple sources and promoting transparency across teams. I’ll give you an example: Right now, operators are prioritizing the automation of time-log entry. On a lot of rigs, people are manually putting this information into the log. There are some rigs in the world where people are really good about putting in NPT, but there’s a vast majority that aren’t. They don’t write anything because it wouldn’t look good for them. So, we’re seeing a push toward automating this.

One thing we’re doing at Corva is we’re using sensors on the rig to generate time-log data and then automatically inputting that data into a database. From that, you can add in comments from the rig contractors. Maybe the companyman comes in at the end of the day and adds some context that is perceived to be important, but the data is already inserted.

I think something like this is important because the foundations of LLMs and text mining algorithms that we can use will be based, in part, on these time-log entries. If you have the performance data entered there automatically, that puts you on the cusp of breaking so many silos.

How close do you think this industry is to really breaking down those silos?

I don’t know the answer to that, but I would say that it’s Corva’s business to break down silos. I feel we are at the forefront of that. But I’m an overly optimistic person. A service company might come to me and say, we can do something, and then a year later it’s a different story.

We’re partnering with a lot of service providers across frac, wireline and geology, along with the rig contractors. That’s one of the best things we have going. The integration of all of these companies and all this information is what everyone has been asking about for the past 20 years. That’s where we see the future going.

There always seems to be this balance between innovation and competitive advantage in this industry. At Corva, you’re developing new technologies that can benefit the industry at large, but you’re still a business trying to maintain market share. Where do you see your work as it sits within that balance? How much do you worry about competitive advantage?

The longer we have offered Predictive Drilling, the more we realize that there’s always something new that we can work on. You find out the stuff you don’t know. There are so many different components to a system like that.

One is rig control. We need this rig to drill as smooth as butter, and we need to avoid dysfunctions as they happen. Those problems are quite difficult on their own, and they require a lot of experience and a lot of iteration. You have to operate within certain limits to avoid dysfunctions, and if you can automate the limits to be tighter and tighter, you can eke out more differential pressure and drill more smoothly. This is a complicated thing that I don’t think anyone else is doing, and it takes a lot of effort and iteration, but it’s a function we’ve built into the system.

All of this is to say that there are a lot of hard problems to solve in this industry, and we’re always going to find something.

As you mentioned earlier, Corva has been involved in a partnership with Nabors over the past couple of years, with the companies announcing an expansion of the partnership into Nabors’ RigCLOUD platform in 2025. Were there any challenges that you initially found with incorporating your system with their technologies?

I think the initial challenges with integration were really centered on aligning the architectures and ensuring low-latency data flow.

What are some of the challenges you find with getting data as close to real time as possible?

I think the main problem is that companies have their own home-built systems, including us, and the integration of those systems isn’t always smooth. With certain WITSML providers, our system would overload their servers and cause them to crash because we’re built on a cloud platform. When we need extra server space in the cloud, we can just add that. That’s harder to do when you’re integrating with a system on a physical server. Now, we have DevOps engineers who are focused on watching every single blip and investigating it. These real-time mission-critical systems are systems where you can’t have any gaps in time.

For instance, if you make a request to an API (application programming interface) and then you get the data back, you can close the request right away and that’s fine. But if you wanted to keep the connection open for whatever reason – maybe you wanted to do something based on that request and come back to the interface later – 90% to 95% of the time spent when you do come back is just on making sure the connection is still secure. That adds up.

Unfortunately, that’s just something we have to deal with.

It sounds like you’re a fan of open-source APIs. Are you encouraged by the increasing industry discussion we’ve seen around interoperability?

I am encouraged. I’m optimistic that we’re all moving in that direction. Some people are just moving faster than others.

Do you think the industry is moving as fast on this front as it should be?

No, I think it needs to be faster. When I first entered this industry, I’d look at WITSML and think, wow, this is such a crappy format, just because of my software background. But now, I realize that it is what it is. It works. It’s an open standard, and everyone knows it. I think it’s one thing that has really propelled this industry forward. It’s one thing I’ve noticed people getting behind on the frac side, for instance.

The key here is working with the standards body. If you’re a big company like Microsoft or Google, you can take a standard and add your own customization. We’re not seeing enough of that – meaning that we’re not seeing companies take WITSML 1.4 and add these things to make the world better. We need to push harder.

You mentioned using WITSML for frac data. That brings me to another question: Corva offers Predictive Drilling, but the company has also touted the usefulness of that system in completions and how it can help companies shift from reactive to predictive completions. How would you define predictive completions?

Predictive completions – or as we call it here, Predictive Frac – is a longer-term strategy that includes connecting to the equipment directly. We don’t have that capability yet, but we’re working on something we call Guided Frac. In real time, we want to bring in all of the pump schedules and actively monitor the pumps, which is not dissimilar from other service companies.

What we’re doing that’s different is we want to overlay past stages in real time, so you can see, for instance, how the pressure response is compared with past stages.

We’re also working to predict the pressure response, as well as building a lot of things around anomaly detection. As soon as we see these pressure spikes, we will be able to respond very quickly, and the system will have this ability to create alerts and actually tag and put comments in them so that the frac crew can review it later.

We’ve talked a bit about data already, and you’ve talked about the industry’s need to better maximize the data it already has. Is there a specific area where you think a company like Corva is closest to helping the industry understand data in a way that it hasn’t before?

We provide all of these analytics, but from our perspective, we’re not the company that goes and finds the one thing that you need to know. If you just ran a well, what are the three things you need to know from that well? What’s the lesson learned? Those are things the customer determines.

But what I would say is that we’ve turned a corner from just giving them all the analysis and then moving on to the next phase. With Corva, we’ve built somewhere around 200 apps that can help you distill that data down as much as you want. But that still goes back to figuring out what’s really important and what’s driving value.

When you talk to anyone in this industry about automation, the conversation always seems to pivot back to the role of the human on the rig. No one wants to take the human off the rig completely. From a tech company standpoint, how much does the role of the human factor into your thinking?

The human is central to the pace of innovation. From a technology standpoint, we can build very quickly, but real progress happens when what we build fits naturally into how people work. That’s where the user experience becomes critical. The interface matters a lot. If it’s intuitive and fits the workflow on the rig, everything moves faster.

Right now, we’re doing a lot around directional and geo integration, things like landing the curve directly within our interface. The technology is there, and the demos are strong. The focus now is making sure it’s delivered in a way that feels seamless and natural for the people using it day to day. We’re constantly pushing to improve how the technology shows up for the user.

More broadly, we see automation and AI changing the role of the human on the rig, not removing it. The human shifts more toward supervision, strategy and decision making.

The goal with automation is to augment human expertise, not replace it, so people can spend more time applying judgment, experience and creativity where it matters most. DC