Editor's note: This article is part of the State of Smart Manufacturing theme week.

The digital age has made its way into industrial processing, enabling organizations to increase productivity, cut costs and remain competitive in creative ways.

Implementing sensors, networks and data acquisition systems allows processors to optimize equipment performance and throughput. This data can be harnessed by process engineers. who are granted immediate visibility into machinery health and operations. This allows a proactive approach to resource management, and swift responses to process defects or abnormal machinery behavior.

Monitoring for chemical reactors

Reactors transform raw materials into desired solutions and substrates by accurately mixing chemicals and reactants, while also applying heat and pressure. This is a careful procedure that requires close monitoring and confidence in equipment. There are some key parameters that should be monitored.

Level sensors are used to track the rate at which the raw material is removed from the storage tank, as well as the rate at which the raw material begins to fill up the reactor. Level sensors can be differential-pressure (DP) design, which uses changes in pressure at the bottom of the vessel versus the top of the vessel to calculate level. Magnetic level sensors measure the changes in the magnetic field as the level rises to calculate the level inside the vessel. Ultrasonic level sensors detect the presence of liquid by the change to the sound waves produced inside the vessel. Magnetic and ultrasonic level sensors do not need to be in contact with the chemical inside the vessel. This is particularly useful for corrosive substances or those that might foul a differential pressure sensor.

A flow sensor monitors how fast this transfer process occurs. Reactants are fed to the reactor using flow sensors. The most common type uses an orifice plate to constrict the flow and the difference in the pressure created before and after the orifice plate is used to calculate flow rate. For systems with corrosive fluids or fluids with particulates, flow sensors with less invasive technologies such as ultrasound flow meters can be used. It’s important to carefully calibrate ultrasound flow meters since they tend to not be as accurate as orifice-plate based flow meters.

Inside the reactor, temperature, pressure and level needs to be continuously monitored as the reaction starts to take place. The temperature, pressure and level sensors inside a reactor need to be able to withstand extreme changes in process conditions. A change in temperature is the strongest indication that the reaction has started. A thermocouple is a common detection mechanism used inside a temperature transmitter. A thermocouple, however, needs direct contact with the fluid in order to give accurate measurements. For systems where this is not practical, such as with thick slurries that could plug the sensing mechanism, a laser sensor might be a better option.

As the raw material begins to change into the final product, the pressure inside the reactor will rise as byproducts and waste gases are evolved. A pressure sensor should be selected that is designed to detect the upper end of the pressure range of the reactor. Otherwise, the sensor will be “pegged out” once the reaction gets going. The chemical mass of the raw material will be different from that of the final product. A change in level indicates that the raw material has been consumed during the process, and that the reaction is now complete.

Each reaction will proceed in the same way if the same quantity and type of raw materials and reactants are used. A reactor profile can be developed based on the data collected from the temperature, pressure and level sensors.

This reactor profile will be the baseline for what is normal for this reaction, and these reactor profiles inform maintenance activities around the reactor. Sensors that operate under high stress conditions, such as the excessive pressures and temperatures of a reactor, will need more frequent maintenance to ensure data is accurate. Uptime and downtime charts should be developed that track the state of the process in real time. Deviations away from normal operating conditions should be investigated right away to address the underlying issue, and to prevent simple issues from compounding further.

Monitoring for bioreactors

Cells or other microorganisms such as bacteria are used to produce a specific protein. This protein is then separated from the cells and purified into the final drug product. The reaction is the activity of the cells and as they grow, more of the target protein is produced. The more protein available, the higher the concentration of the final product.

With a bioreactor, temperature, pressure and level are important. Cells need to be kept at a target temperature, or else they will die. Thermocouple based temperature sensors are common in bioreactors. It’s important to select a thermocouple that is long enough to project into the interior of the bioreactor without hitting the impeller blades that mixes the contents.

The pressure that is exerted onto the cells also needs to be properly maintained to avoid disrupting the fragile cell walls. Cells produce waste as part of their cell growth. The level inside the bioreactor needs to be monitored to ensure the ratio of cells to waste is optimal for adequate cell growth. DP level transmitters are typically sufficient due to the low operating pressures used for bioreactors.

Just like in a chemical reactor, a bioreactor profile needs to be developed that correlates to normal cell growth. Deviations away from the normal bioreactor profile need to be quickly addressed to identify issues with the reactor, reaction or sensors. Many of the sensors used in bioreactors are continuously submerged in solutions, which makes maintenance challenging.

One unique consideration is that stainless steel bioreactors need to be cleaned in between batches using cleaning-in-place (CIP) and steaming-in-place (SIP) cycles. These cycles have unique temperature and flow profiles that need to be monitored to ensure the inside of the bioreactors are thoroughly cleaned.

Monitoring for storage tanks

For many types of storage vessels, similar parameters are monitored, including temperature, pressure and level. As a result, basic thermocouple-based temperature sensors, DP pressure sensors and DP level sensors are typically sufficient to monitor the internal conditions within a storage vessel.

In chemical plants, long-term storage vessels are typically installed outside. Sensors need to be rated for weather conditions in extreme temperatures. In frozen climes, external heating mechanisms such as heat tracing and heating blankets can be used in lieu of pricier all-weather sensor models. Here, temperature sensing can ensure that media is always ready to flow — and that expensive electrical energy is used conservatively in heating. Additionally, in warm climes, temperature monitoring can help prevent sensitive materials from exceeding safe storage temperatures. In many cases, hazardous materials may be required to have leak detection systems.

For inert or safe materials, monitoring may not be as critical. In those cases, instantaneous data monitoring may not be necessary. Right-sized data supply to data needs can help process engineers make informed decisions about feedstock supply and equipment performance.

Managing data from process networks

All of the data collected by the sensors in a processing facility is fed into a database. By hosting this data in the cloud, process engineers can have transparency into equipment health and operations status from a number of devices. Alarm setpoints can identify deviations away from normal operations or behavior and help employees solve issues before they turn into broken equipment or a runaway reaction.

This can be a lot of data and can be more effectively managed with the help of artificial intelligence (AI) or machine learning algorithms which can offer succinct analysis from large datasets. For companies that operate an on-premises computer server system, data storage is a concern. One solution is to right-size data collection and transmission from sensor points. Alternatively, the sensors can be programmed to only transmit data once a threshold is reached.

Data security and safeguards against network intrusions are key. There have been several notable occurrences of process facilities being hacked and operations disrupted by black hats installing ransomware on industrial networks. Considering the dangerous chemicals and materials that might be present, it is a serious concern.


Sensing networks in process engineering aren’t new, although they are arguably table stakes for any manufacturer focused on productivity, safety and efficiency — which should be all of them.

Sensors can then be used to transmit critical process data in real time that can be integrated into the overall process and data acquisition system. In this way, the entire process can be visualized from a single point of view, allowing for a better, faster decision-making process. Making full use of the data will require new approaches and skillsets, but will ultimately yield benefits that outweigh any challenge.

About the author

Simone Ammons has led a 15-plus year career as a process engineer spanning the oil and gas, chemical, and pharmaceutical industries. She now leverages her extensive engineering background as a technical writer to concisely articulate complex concepts into a format that is easy to understand.

To contact the author of this article, email GlobalSpeceditors@globalspec.com