Process engineers use MATLAB® and Simulink® to analyze real-time sensor data, implement control strategies, and create predictive maintenance systems based on big data and machine learning.
MATLAB and Simulink help process engineers:
- Develop predictive maintenance systems by applying numerical techniques on high-speed sensor data
- Use machine learning with historical data to troubleshoot process problems
- Use data modeling to improve process performance
- Develop and implement advanced predictive control (APC) strategies
- Adopt digitization without depending on data scientists or IT personnel
"As a manufacturing company we don’t have data scientists with machine learning expertise, but MathWorks provided the tools and technical knowhow that enabled us to develop a production preventative maintenance system in a matter of months."Dr. Michael Kohlert, Mondi Gronau
Watch an Example
Beyond Process Simulators
Traditional process simulators are generally sufficient for steady-state conditions, but they cannot handle the dynamic nature of the inputs that real plants deal with. With MATLAB, you can write your own equations and algorithms, giving you full control over the entire model.
You can also integrate MATLAB with process simulators like Aspen Plus and gPROMS for custom unit operations, advanced analytics, designing control schemes, and exploiting optimization routines such as genetic algorithms.
Digital Twins for Oil and Gas Production (48:37)
Watch how (12:17) Dow uses MATLAB with Aspen Plus for process optimization
Optimize Assets with Predictive Maintenance and Signal Processing
MATLAB can help you develop predictive maintenance algorithms customized to the specific operational and architectural profile of your equipment. Use Predictive Maintenance Toolbox™ to design condition indicators and estimate the remaining useful life of your rotary equipment.
You can use Signal Processing Toolbox™ to automate the monitoring of performance of your control loops, remotely determine the extent of corrosion or pitting in your pipelines, and detect the location and quantity of pipeline leaks.
Read how Baker Hughes used MATLAB to implement a predictive maintenance platform for gas and oil extraction equipment and reduced overall costs by 30-40%.
Watch how (39:51) Tupras implemented an automated control loop performance assessment system at their refineries.
Machine Learning and Big Data
Interactive apps in Statistics and Machine Learning Toolbox™ let you apply machine learning techniques without having to be an expert in data science. MATLAB also provides a single, high-performance environment for working with big data – be it structured or unstructured. This enables you to perform fault detection and diagnosis faster and better monitor your processes.
Read how I2C2 researchers analyzed millions of rows of process data and developed machine learning models for predicting milk powder’s functional properties.
Deep Learning and Image Processing
With just a few lines of MATLAB code, you can build deep learning models that use your process data to predict abnormal conditions. Use Image Processing Toolbox™ apps to automate common processes like segmenting image data and batch processing large image data sets. You can use MATLAB in image processing applications such as flame characterization, thermal imaging of equipment, and plastic film quality inspection. With deep learning in MATLAB, you can learn feature representations directly from image and video data.
Read how Dexerials Corporation uses AI for detecting defects in real time for film manufacturing.
Process Improvement with Data Modeling
Use multivariate analysis tools in MATLAB to determine the independent driving variables affecting process performance. System Identification Toolbox™ lets you create and use models of dynamic systems that are not easily modeled from first principles or specifications. The toolbox also lets you interactively perform online parameter and state estimation.
Watch how Shell used MATLAB (3:35) to develop models and perform real-time optimization on a batch process.
Develop and Implement APC Strategies
You can use MATLAB controls products to design controls schemes and perform dynamic simulations for better analysis of plant behavior. Design, simulate, and deploy linear and nonlinear model predictive controllers for your plant using Model Predictive Control Toolbox™.
Controls engineers can also embed process models from Aspen Plus and gPROMS into Simulink. In this way, you can redeploy existing models to design a control strategy in your preferred environment.
Read how Tata Steel saved 40% energy on their industrial cooling towers by optimizing the control strategy via a digital twin.
MathWorks can help you adopt and implement big data strategies specific to the needs of your organization. You can use prebuilt MATLAB toolboxes and reference architectures to simplify a wide range of applications: from integrating with enterprise IT systems, the cloud, and production data infrastructure to scaling your computation to clusters or deploying your models as applications to share with non-MATLAB users.
See how you can connect directly with AVEVA™ PI systems and enable real-time operational intelligence.
Watch how Shell embraced digitization (29:14) using MATLAB Production Server™. Shell engineers automated their processes for integrating data from different sources, building models, and deploying their analytics onto cloud and enterprise systems.
Closing the Loop with DCS Deployment
MATLAB algorithms can be integrated with a variety of DCS systems with the Industrial Communication Toolbox™. The toolbox provides access to live and historical OPC data directly from MATLAB and Simulink. You can read, write, and log OPC data from devices, such as DCS, supervisory control and data acquisition systems, and PLCs. Industrial Communication Toolbox lets you work with data from live servers and data historians that conform to the OPC DA, HDA, and UA standards.
Read how Genentech uses MATLAB and Industrial Communication Toolbox to build a supervisory control algorithm development platform for bioreactors.
“Another advantage of developing our own system in MATLAB and Simulink is that we can capture the organizational knowledge and expertise accumulated by Johnson Matthey engineers rather than relying on another company’s one-size-fits-all solution.”Tim Watling, Johnson Matthey
Let us know how we can help you.
Contact us for a custom evaluation, and get pricing based on your needs.