Digital transformation is the integration of digital technologies to enhance productivity, efficiency and decision making in all areas and sectors. The industrial sector is currently joining the wave and undergoing a true revolution towards a digitalized sector at a slow but consistent pace. This has proven to be very efficient in optimizing industrial processes and improving its efficiency. In this blog, we will define the industrial digitalization and some steps and guidelines to get through the digital transformation journey.

The industrial digitalization is the use of cutting-edge technologies to refine the way that industries operate and use the power of data to improve the decision making. The key technology for this revolution is the integration of Internet of Things or what is called industrial IoT (IIoT) and artificial intelligence (AI). The convergence of these two technologies offers enterprises novel prospects and distinctive instruments to collect massive data, analyse it and integrate AI to maximize the benefits from it.

The digital transformation could be seen as 4 different forms or stages depending on the data and the maturity of this process or production line specifically.

1. Data surveillance and analysis: Real data collection

The old PLC generations were designed to collect data from the sensors and present it in basic HMI screens for operators. With the advances of its technology, the control systems are now able to collect and store a vast amount of data from tons of sensors. At this stage of digital transformation, you only need this data real time streaming to improve your process. The goal will be to build a strong data representation to get insights from the data and to monitor your real time reading in the process. At Algorex, we make use of this technology to build a well-designed digital platform, so the client has access to: real time sensors reading and their historical values in the kilns through nice time series, calculation of tons of KPIs to assess the productivity of drying and wood quality, personalized reports and many more features.

 2.  Anomaly detection: data with no history

At this stage of digitization, the data is collected and monitored following the first step. Next stage will be automating the anomaly detection using algorithms. The process experts’ presence is mandatory at this stage to define the rules and thresholds that characterize the anomalies in your system and/or process. Then the AI will interfere using system experts to build all the logic rules to define your algorithms. This combination is powerful in the sense that the problems, breakdowns, anomalies will be detected automatically and in a real time basis. At Algorex, we difference ourselves from the competitions by building a powerful, flexible, and robust algorithms defined by a worldwides drying experts which allows us to avoid false alerts.

 3.  Predictive analysis: real time and historical data

Arriving at this stage of digitizing your process, your data pipeline has gathered enough data to build a good history of your process. Moving forward with real-time anomaly detection, it’s time to anticipate and predict anomalies and parameters. The power of IA will jump in this time with more sophisticated and powerful tools through predictive models. Depending on your process and available data, this prediction enables a good new features and perspectives to anticipate your anomalies before they happen. It could be also used to predict your critical/objective parameters with a lag of time. It gives more time to react and take the adequate decisions to optimize the productivity of your process and product’s quality.

 4.  Prescriptive analysis: data, history, and predictions

All the previous steps are useless if you don’t integrate it into your decision-making process. The prescriptive analysis is the use of advanced processes and tools to recommend the optimal course of action or strategy moving forward. We can see this from two perspectives, first the process experts will have more information from the past, current and future state of the process and can take more adequate decisions. Second, we can use operational research (RO) algorithms to optimize our KPIs. The predicted KPI is controlled by several parameters in the process. So, the role of RO algorithms is to calculate the best combination of those lasts to optimize the predicted KPI. Using the advantage of the time lag, those parameters have enough time to be adjusted and optimize the system in the next hours.

The power of data and AI are massive, depending on your process and the technologies implemented, a lot of amazing things could be done to strengthen your decision making. But the complexity and variability of production processes make the integration challenging. The AI is not a magic wand, you should be caution and perseverant until you get the maximum insights you can.