Many clients view their production chemical program as an annoying cost, but to me it seems that this is driven by an overall level of dissatisfaction with their chemical programs, as reported on market surveys. They seem to be viewed as bothersome costs, instead of being treated as a critical part of ensuring production and preventing problems within the production system.  A fitting analogy would be like ordering lawn service only after you have more weeds than grass, instead of taking a proactive approach.  So why is there such discrepancy between how chemical programs are perceived now vs how they should be perceived if they are effective? Furthermore, how can many digital transformation programs which are underway today help improve chemical treatment efficacy and therefore improve profitability?

As I see more and more chemical treatment programs, it seems that perhaps chemical treatment programs are not ineffective in general. Rather, it’s the monitoring of the effectiveness of the chemical treatment programs that is ineffective, this monitoring influences the recommended injection rates and overall effectiveness.

When we look at how monitoring of chemical performance is done today, it is by in large a very manual process. If we use monitoring how effectively your chemical program is at protecting your wells from scale deposition as an example.  Let’s look at some of the key flow conditions that impact the risk of scale deposition and how frequently changes in those are typically measured and monitored today.

• Produced Fluid Rates- generally acquired via well tests performed every 2-4 weeks
• Temperature - generally manually observed/measured at surface when collecting a sample and less frequently for downhole
• Pressure – generally manually observed/measured at surface when collecting a sample and less frequently for downhole
• Water Quality – generally based on physical water samples analyzed in a laboratory on a monthly or quarterly basis

Given the above typical monitoring regimen, it is safe to assume that new data used to provide input into a scale prediction exercise is available best case monthly, more likely less frequently.  Once the result of this exercise are available, an adjustment to the chemical dosage may be recommended and the chemical injection would be manually adjusted to meet the recommendation.   While it may be safe to assume that some of these conditions remain relatively stable over a period of time, it would be naive to assume that all them remain in a steady state for that period of time.

In exploring if there is a better way to do things, we would recommend a simple methodology consisting of series of steps.  In order to test this methodology, we have worked with a Latin American Customer that are experiencing a high number of ESP failures, many of those related to issues stemming from scale precipitating in their wells.  The steps we took were as follows:

1. Understand which data we need at higher frequency in order to proactively monitor the effectiveness of the chemical treatment program.  We determined that if possible, we would want to monitor at a higher frequency: changes in produced fluid rates, temperatures, pressures, and water quality.
2. Determine which of that data we can get by digitally connecting existing equipment/systems. We determined that both pressures and temperatures could be acquired by connecting to the existing ESP controller (downhole conditions), and existing wellhead temperature and pressure sensors (surface conditions). 
3. Understand where data driven, and physics-based models can be employed to predict in real-time information where we were unable to acquire from equipment.  In our case we deployed a virtual flow model (VFM) to acquire a real-time production rate, and raw temperatures and pressures as inputs into a continuous scale prediction model.  This was augmented by the updates to the water quality from laboratory test results.
4. Strategically install additional digitally enabled equipment to fill data gaps.  One gap that we observed was the chemical injection rate.   To close this data gap, we installed chemical injection pumps that were digitally enabled so we could monitor the chemical injection rate.
5. Implement close loop control.  This is where we use the data we are now acquiring in real-time. This is simply taking a process that had historically been manual and based on outdated data and allowing the chemical injection to be automatically adjusted based on the real-time changes in the production system.

Results

• The results we found were very positive (and in some cases quite surprising). Since the autonomous system has been running, the chemical injection is being adjusted between 250-300 times per day.  This is an astounding number and really does support the thought that the production system is in fact very far from steady-state, as traditional methods would have us believe.
• When looking at maintaining compliance between the actual injection rate and the target injection rate, the autonomous system is compliant more than 99% of the time, versus the 60% historical compliance.  
• The models have been remarkably accurate, as the VFM has been within 3% of all bi-weekly well tests and has highlighted some invalid well tests along the way. The scale prediction model has been continually validated against commercial scale simulation methods and has been within 2%.  
• With the system running autonomously, required trips to the field have been reduced by more than 75%.  This has a very positive impact limiting both the exposure of field people to safety risks, and the overall carbon footprint of the operations.

With these favorable results, why do we continue to do things manually?  The answer is that there is a perception that moving from a manual monitoring method to a more real-time system would be too expensive, and I agree, on the surface it looks like it would be.  However, we wanted to test this out by doing a very thorough economic analysis, using real numbers provided by the Customer we are working with.  In our preliminary economic feasibility study, we have considered the following factors:

• Historical Failure Rates – we wanted to understand these numbers for all the wells in the asset, because it would help us in our “what if” scenarios.  For example, what impact would be had if our system reduced the failure rate by x% YOY?
• Historical Production and Decline Rates – this helps determine the impact on the go forward production revenue if our solution could reduce downtime.
• Chemical spend (historical and future) – this needs to be understood as accurately as possible for each well.  Wells that were historically under-treated will need more chemical, and obviously wells historically over-treated will need less.
• Typical Downtime Duration – we need this to factor in the production revenue that is lost when a failure occurs.
• Work-over costs per well – to understand how much failures are costing now and what impact failure reduction could have on overall costs.
• Additional equipment (CAPEX) – this needs to be factored into the overall cost vs benefit analysis.
• Goals (failure reductions, ROI targets, etc.) – this is critical in defining the key performance indicators and acceptance criteria to determine the overall success of the solution.

Once we incorporated all of the numbers into the analysis, and assumed a 15% reduction in failures YOY, the results showed that the Customer would see an ROI in excess of 100% after just one year, and over 500% after 5 years.  These are numbers that make sense to anyone in our industry and are very aligned with some of the key drivers for all Customers: reducing costs and maximizing production.

Obviously, the results speak for themselves, but we know that it can be very overwhelming these days where it seems like everyone has a revolutionary digital innovation meant to solve all the issues faced by Operators today. We have learned some things that we believe are worth sharing.

• When it comes to digitization projects, they should be focused on reducing specific problems present in your operations.
• Walk before you run, start with a subset of your worst wells and field trial the solution.
• Give it time to prove its value. Any solution that will have a material impact on your business will take some time to display true value.  Change doesn’t happen overnight.
• Be invested as a partner.  Success will come when the effort is collaborative with a joint desire to succeed.
• Start by assessing what is possible with existing equipment and systems.  Use the data you already have access to before you start installing new equipment.
• Economic feasibility analysis needs to be thorough.  Digitization may very well not make sense on all your wells or assets.
• With technology and innovation available today, traditional methods may not always be the most cost-effective way to ensure positive results.

As companies work to digitize their assets and workflows, it is important that the production chemical programs are not excluded because there is so much room to improve from the status quo and it can impact many areas within the production system.  We’ve proven value in downhole applications and improving ESP run life but having access to more data will also drive more value creation in surface equipment and production facilities.

Disclaimer: All opinions expressed by the blog contributors are solely their current opinions and do not reflect the opinions of SLB or its affiliates. The blog's opinions are based upon information they consider reliable, but neither SLB nor its affiliates warrant its completeness or accuracy, and it should not be relied upon as such.