The refrigeration advanced process control (APC) solution is intended to be used by engineers, refrigeration technicians, and refrigeration consultants whose role is to improve the performance of refrigeration systems. For a refrigeration APC project to succeed, there are three groups of stakeholders that should be involved in the process: the clients of the machines, the maintenance teams that work on the machines, and the operators who are using the machines.
This approach is not aiming to optimize the performance of an individual machine. It is instead optimizing the performance of the entire refrigeration plant. A multitude of machines including heat exchangers, pumps, compressors, and cooling towers are included in the APC solution. For this optimization, we calculate a coefficient of performance (COP) that is defined by the amount of refrigeration power produced over the amount of electrical energy consumed. This is calculated in real time for each line in the refrigeration plant along with the real time process variables for the various machines that make up the line in order to serve as variables for our optimization model. This solution also includes an asset health approach as well with regards to the heat exchangers. We use a predictive maintenance model to identify dirty heat exchangers; dirty heat exchangers can increase energy consumption by over 25% in some cases and as such an improved maintenance strategy for heat exchangers can provide large benefits towards reducing energy consumption. Ultimately, following this approach of optimizing COP and improving asset health leads towards the goal of obtaining a more stable refrigeration plant.
In order to collect the data needed for this application, we are using the IOT Unified components of SORBA. SORBA has over 20 unique industrial drivers which cover most common industrial communication protocols. Typically, data for these applications is stored in an OPCUA server which we can connect to with SORBA, or we are pulling data directly from the PLC. The refrigeration APC solution can sometimes include up to 600 collected tags for larger refrigeration systems, but the measured variables are relatively standard from line to line. We are measuring suction pressure, discharge pressure, temperatures of the water basins and cooling tower, compressor energy, fan energy or electrical consumption, wet bulb temperature and the temperature of the loads.
Once the data tag structure is configured, we wait to collect a sufficient amount of data; the presence of historical data can accelerate this process. During this time, we are calculating and recording the COP in order to create a baseline of what normal behavior is for the system. Once data has been collected, we can feed that data into an optimization analysis. This is a nonlinear regression that generates a digital twin based off the historical data. The model looks at what setpoints led to good behavior and can then recommend the optimal setpoint in real time for the control variables with relation to an independent variable. For this application, the optimization variable is COP, with the optimization goal being to maximize COP. The independent variable is DTMax, the difference in temperature between the wet bulb temperature and the coldest load. This independent variable gives us a frame of reference for the different operating conditions that the refrigeration plant experiences. We are then taking control over the suction pressure, discharge pressure, and the temperature of the cooling tower water basin.
Once the model has been created and deployed, the model must be verified in an open loop test prior to closing the loop. During open loop testing, we track the optimal band, the optimal setpoint, the predicted COP, and the calculated COP. It is important to verify that the predicted and calculated COP values are like one another as this is an indicator of how well your model is matching up with the reality of the refrigeration plant. At this stage, before moving to closed loop testing, it is important to add in any safety checks to the scripting logic if some of the controllable variables need to stay within certain ranges.
For closed loop testing, the controllable variables are written back into the machine, allowing SORBA to fully take control over those variables. Once this is done, you can conduct a series of APC ON and APC OFF tests in order to compare the performance of the refrigeration plant when the APC is running and when it is not. When comparing these trials, it is important to not compare the COP’s directly to one another as a multitude of variables such as the load temperatures required during that testing period as well as factors such as ambient temperature can impact the COP beyond the performance of the APC and the performance of the plant without the APC solution. To account for this, we calculate the theoretical maximum COP based on the Carnot cycle for the same DTMax. We then take the ratio of the calculated COP during the APC ON or APC OFF test to the theoretical maximum COP from that testing period. It is then these ratios that can be compared to one another in order to see how much energy is being saved when the APC solution is on compared to when it is off.
Depending on how efficiently the plant was running prior to implementing SORBA’s APC solution, we have seen anywhere from a minimum of 3-5% increase of COP with some plants achieving up to an 8-15% increase of COP. SORBA provides a platform of tools that allow engineers to improve processes and see real improvements in machine performance. There is a large gap between PLC code and analytic code that can be used to improve performance; SORBA helps to bridge that gap. Additionally, SORBA provides a software solution that can provide massive benefits at a low cost. Most solutions for improving refrigeration systems come with the addition of new hardware that can be quite expensive. SORBA offers a purely software solution that works with the data that is already being collected. Finally, it is important to note that the refrigeration APC solution is a template. SORBA is completely customizable to fit any specific refrigeration system and can be applied to a multitude of other industrial systems beyond refrigeration.