In the oil and gas (O&G) industry, downstream operations have consistently outperformed the automotive, chemicals, and integrated O&G industries over the last decade. It is at the beginning of this decade that things took a different turn.
The pandemic caused a severe drop in demand which is only beginning to recover, and crude oil prices have been under fluctuations owing to geopolitical turbulence. Following this turn of events, organizations in this industry are now looking for ways to make their operations more agile and responsive to market changes, including managing operating costs in the chemical industry.
Following the success stories of AI in upstream and midstream operations, time is ripe for exploiting AI in oil and gas within downstream operations, where opportunities lie abundant. Take a look at the most promising use cases of AI in downstream oil and gas in the following paragraphs.
Top use cases of AI in downstream oil and gas
In downstream operations, AI finds a diverse set of applications, ranging from employee safety to operations optimization and equipment maintenance. Applications like optimization of crude distillation units (CDUs) and fired heaters, and predictive maintenance of critical equipment are aimed at achieving cost savings – especially as operations and maintenance represent the major chunk of costs of downstream operations.
Take a look at these use cases in detail below.
#1. Intelligent fired heater operation to optimize CDU performance
In the operation of downstream O&G facilities like refineries, fired heaters account for a significant amount of energy consumption – i.e., nearly 60% of all energy used in the refining process. Because they are operated using a fixed set of parameters that are seldom altered, fired heaters typically operate with some room for energy efficiency improvement. Despite running them on the basis of manufacturers’ data sheets, there is some room for improvement. In fact, a 1% efficiency gain can save 900 tons of fuel annually.
These gains are now possible to achieve with AI algorithms, which take into account numerous parameters, and fine tune the operating parameters of fired heaters to improve CDU performance. One of the key factors that degrade efficiency, is the calorific value of the fuel used in the heat exchanger, where sulfur content is a key variable in efficiency calculation. In addition, factors like air requirement must also be taken into account for fuel firing.
These factors can now be taken into account by ingesting real-time crude oil composition data and fuel sulfur content, and feeding it to AI algorithms for controlling firing heater operations. For this application, the use of soft-sensor technology and multi-output artificial neural networks (ANNs) have shown promising results., Studies suggest an energy saving potential of over 208tn BTUs per year.
#2. Predictive maintenance of critical equipment in refineries
Downstream operations consist of complex thermodynamic processes that occur at high temperature within intricately engineered equipment. It is therefore not surprising that maintenance accounts for a crucial percentage of operating costs for downstream businesses. That’s why early interventions leveraged approaches like Analytical Hierarchy Processes (AHP), the American Petroleum Institute (API) guide, and Failure Modes and Effect Analysis (FMEA) to guide their maintenance operations.
These, however, now represent legacy approaches to running maintenance operations in downstream oil and gas. With AI-powered predictive maintenance now at hands, downstream players can unlock new baselines of maintenance costs with high precision. In downstream operations, predictive maintenance can be applied for critical equipment like heat exchangers, reactors, distillation columns, compressors, and piping systems.
Heat exchanger predictive maintenance typically makes use of statistical models to predict the rate of fouling. Such approaches promise a potential cost reduction of 30%. For the maintenance of compressors, distillation columns, and reactors, a different approach is used. Here, an AI-based classifier is taught to differentiate normal operating conditions against abnormal ones, or data pertaining to operational states is divided into clusters to identify situations pertaining to a high degree of abnormality. These approaches can be used to predict significant corrosion or clogging in piping, and reciprocating compressor failure.
#3. Preventing health, safety and environmental (HSE) hazards with AI
The downstream O&G sector has experienced some of the most hazardous accidents over the last decades. While safety and environmental risks are a part of the O&G industry, downstream operations are prone to a specific set of risks. These include explosions and fires, leakages due to piping corrosion, chemical exposure in confined spaces, vehicle collisions, and trapped technicians in confined spaces.
The root causes of these accidents are varied. They range from negligence and human error to lack of timely maintenance and repair on critical equipment. AI finds numerous applications in the mitigation of HSE risks. A few of these include:
- Use of aerial surveillance drones to capture real-time images, and analyzing them with image processing algorithms to detect fires and safety hazards in time.
- Using AI to monitor health data from wearable devices on technicians to identify safety and health hazards and make timely interventions.
- Applying AI algorithms on operational systems to ensure all the equipment is properly calibrated. This can also help with regulatory compliance.
AI in downstream oil and gas: where to start?
With multiple use cases of AI being prototyped within a span of weeks, it is crucial for downstream players to prioritize highest value use cases. In other words, investments in AI must be made in accordance with a strategically identified roadmap.
From the purview of costs, predictive maintenance represents the low-hanging fruit, promising immediate reduction in operating costs, and enhancing the ROI on existing assets in use. Moreover, predictive maintenance implementations can also have a significant impact on worker safety and operational efficiency. As a result, it presents a holistic improvement opportunity for downstream players.
Ready-to-deploy predictive maintenance solution from UptimeAI has lowered the entry barrier even further, by eliminating the costs of prototyping, model training and optimization, use case development, and software creation.
Start your AI journey to transform your downstream oil and gas operations to unleash new ceilings of profitability in your market. Contact UptimeAI to get started.