(130c) Combating Corrosion Under Insulation By Industrial Iot Monitoring

Problem description

Corrosion under insulation, often known as CUI, has been one of the largest unsolved industry challenges for decades. Vast resources are spent on finding and combating the issue, while the threat of catastrophic failures remains for corroded areas that go undetected.

Production plants may contain hundreds of kilometers of piping systems to transport fluids or gases under pressure. Leakages must be avoided at all costs, due to the risk of disastrous explosions and severe pollution. Corrosion is a major contributor to this risk, as it tears down the wall thickness and reduces the strength of pipes. Several factors affect the risk and rate of corrosion, such as coating quality, material and surface temperature. However, wetting is the most important factor – without water, there is no corrosion.

A large portion of these pipes are insulated. The insulation is protected by metal cladding to keep water from intruding into the insulation. However, water tends to find its way through the cladding, for instance as rain water intrudes small openings in the cladding. As the cladding surrounds the insulation and process piping, it may hide the indications of any wetting or corrosion, appearing pristine from the outside. The insulation might also retain water. This increases the risk of corrosion going undetected and drives a tremendous effort in inspection and maintenance of insulated piping to combat CUI.

Current state of maintenance planning

To maintain the integrity of a process plant, necessary maintenance must be performed. As CUI slowly and covertly develops and eventually causes a leakage, one must have a strategy to detect corrosion or factors leading to it in time to mitigate the threat. Risk based inspection (RBI) methods assess the probability of factors that may lead to CUI and determine the priority and order of inspection based on an overall risk evaluation. Since CUI is a hidden threat, inspection often involves removal and replacement of cladding and insulation.

Operators have different RBI strategies and methods, but typical factors considered are process temperature, asset age, pipe material, coating type, and degree of wetting. Most of these factors are well known and can be obtained from documentation. However, while being one of the most important causes of CUI, wetting is highly unpredictable and variable, and its risk is often classified based on assumptions.

The uncertainty of the results of RBI assessments lead to overwhelming resources being spent on inspection and maintenance, and tons of replaced cladding and insulation wasted causing negative environmental impact. Inspection campaigns typically find that over 90% of the inspected areas are dry and unaffected by corrosion. At the same time, there is always a risk of corrosion appearing in the areas not prioritized for inspection as the location of wetting is highly unpredictable. All in all, the current methods of maintenance planning lead to high costs and remaining risks that could be improved by sufficient knowledge about the wetting.

Several attempts have been made to determine the degree of wetting to enable more precise and accurate risk assessments. Some methods rely on personnel operating large and expensive equipment in the process plants and have not shown satisfactory results. Other methods provide continuous monitoring of moisture but are costly and inefficient to install and operate. These experiences have shown that to safely and efficiently predict the risk of CUI, broad coverage of continuous wetting monitoring is required. This leads to the need for cost efficient and smart industrial internet of things (IIoT) sensors, and software solutions that can transform the vast amounts of raw measurements to actionable insights that enable optimized maintenance plans.

Technology

To enable cost efficient IIoT monitoring with dense coverage in large process plants, the sensor production cost is merely one factor. The total lifetime cost also consists of costs related to installation, provisioning, integration, and operation. The battery life determines how many years the costs can be split by.

Installation can be very costly, as piping systems often need scaffolding to access. Furthermore, wireless technology usually requires installation of gateways with power and network cabling throughout the site. To keep the installation costs down, the Fusion sensors are equipped with a fixation device allowing them to be plugged into the cladding in seconds. An optional self-drilling tool enables future robotic installation. More importantly, the solution eliminates the need for gateways as each sensor communicates directly with the CirruSense cloud through cellular 5G base stations that are usually already available in industrial areas.

The smart sensors are configured and connected to the cloud when they leave the factory. The only provisioning needed is to associate the sensors to the point they are measuring, which is done in seconds by selecting the point in an app and tapping the sensor with NFC or scanning a QR code from the phone. Identifier Links enable Industry 4.0 standardized unique identification of every sensor.

Operation and support are handled by Trisense as part of the service. Every sensor reports health metrics and connectivity status to our operations team and is remotely configured and updated to monitor and optimize its performance. Sensors, connectivity, operations, data storage, analysis and dashboards are all delivered as a subscription service, enabling the customer to focus on utilizing the insights provided by CirruSense.

Finally, battery life is crucial to the business case, as replacing sensors is an expensive operation. By turning every stone in electronics, firmware, communication protocol and cloud platform, the sensors have a battery life of 15 years with one measurement per hour and one transmission per day.

From raw data to insights

With dense coverage of sensors continuously monitoring the conditions of piping systems comes vast amounts of data that must be securely and robustly transferred and stored. However, measurements by themselves do not yield much value.

First, the data must be structured in such a way that they can be processed and refined. As sensors might be moved or replaced, a system that enables mapping between different sensors and logical measurement points to different time periods is needed. This allows focusing on analyzing continuous data sets without concerning the history of physical sensors. Furthermore, measurement points are organized in a structure that reflects the customer’s assets and their properties, making it possible to efficiently analyze data and build insights specific to the CUI domain.

Next, measurements must be normalized and sanitized to ensure the quality of data. As an example based on field experience, we know that sensors in the same plant are differently affected by sun radiation, causing undesired variations in measurements irrelevant to the wetting state. By compensating for this, the interesting variations in the data appear. This is performed in the cloud together with other calibrations and corrections.

States of the monitored assets are then classified, so multiple measurements of continuous ranges are distilled down to classifications indicating whether the monitored point is dry, humid or wet for every day. Furthermore, the classifications are aggregated over time and over multiple measurement points, to create top-level reporting that ultimately helps optimize maintenance plans.

Finally, as the value addition described above happens in the cloud, it is possible to improve results as refinements are made to algorithms and new features are developed.

From insights to action

As measurements are collected, organized, analyzed and aggregated, top-level reports reveal the risk levels of different monitored assets. The reports can be viewed in interactive dashboards or presented in PDFs. In any case, the main purpose of the insights is to enable optimization of maintenance plans. By knowing the degree and location of wetting for every asset, it is now possible to spend the maintenance resources where it matters most.

Another valuable possibility is to perform spot repairs when water intrusion occurs. The event detection system of CirruSense notifies the operators by e-mail, allowing immediate action before the water degrades the coating and eventually leads to corrosion. This helps maintain the integrity of the assets and is especially helpful on new or newly refurbished assets.

Finally, continuous monitoring documents the integrity of the assets, as well as the effectiveness of repairs performed.

Future possibilities

Data collected over several years can increase our understanding of how different factors affect corrosion of the pipes. By utilizing tools such as machine learning together with feedback from inspection results, continuous wetting data, as well as ongoing research into other factors such as coating degradation mechanisms and effects of temperature cycling, we will be able to provide highly precise predictions of CUI risk.