We believe that digitalization will have a significant impact on many areas of the chemical industry, among which the improvement in manufacturing performance may be the greatest. Chemical manufacturers have invested in IT systems and infrastructure, which can generate large amounts of data about chemical companies, but so far, many manufacturers have failed to take advantage of this mountain of potential intelligence. With cheaper computing power and more advanced analytical tools, professional and commodity chemical companies can put this data to work, collect information from multiple sources, and use machine learning and visualization platforms to discover ways to optimize factory operations.
Advanced analysis can greatly improve the level of understanding of chemical plant production operations; this can help chemical companies solve previously unsolvable problems and reveal those they never knew existed, such as hidden bottlenecks or unprofitable production lines.
In this article, we will show three main advanced analysis-based tools—predictive maintenance; yield, energy, and throughput analysis; and value maximization models that can help improve the performance of chemical producers’ assets and supply chains.
Predictive maintenance analyzes historical performance data of production equipment and its machinery to predict when the equipment may fail, limit the time the equipment is out of service, and identify the root cause of the problem. Yield, energy, and throughput analysis (YET for short) can be used to ensure that a single production unit is as efficient as possible during operation, helping to increase its output and throughput or reduce its energy consumption. At the same time, the value maximization model carefully examines thousands of parameters and conditions that have an impact on the overall profitability of the overall supply chain, from raw material procurement to complex and often interrelated chemical production steps, to final sales, and then provides How to best use intelligence on given market conditions.
In short, these advanced analysis methods can increase EBITDA by 5 to 10 percentage points. In addition to increased profit margins, companies that quickly adopt these methods can use these methods to build a competitive advantage, even for companies that are struggling with overcapacity. They can use these tools to continuously improve the way they manage the production system and real-time reallocate resources throughout the production process in the most effective and value-creating way.
Reduce downtime through analysis
Chemical companies can use big data analysis to find ways to predict failure, thereby extending the operating time of key assets. This predictive maintenance system collects historical data1 to generate insights that cannot be observed with traditional technology. By applying advanced analytics, companies can determine the environment that may cause machine failures. They can then monitor all relevant parameters to intervene before a failure or prepare to replace components in the event of a failure, thereby reducing downtime. Predictive maintenance can usually reduce machine downtime by 30-50% and extend machine life by 20-40%.
Chemical companies have begun to see considerable gains in this area. A major surfactant manufacturer has continuously encountered problems with recirculation and discharge pumps in its largest plant. When one of the pumps broke down, the factory had to stop production for 10 hours while installing a new one; besides the cost impact of production losses, these pumps were also very expensive. Engineers tested several hypotheses to determine possible causes of failure; they also tried alternative materials in pumps and seals, as well as different process conditions, but none of them could solve the problem.
An advanced analytical method has changed all of this. It combines detailed analysis of data from hundreds of sensors and the expertise of plant engineers, and re-examines process variables and other data sources; this allows the company to develop a method to predict when failures will occur. The problem only appears in some surfactant formulations, not all batches, indicating that the key lies in the specific process conditions of the equipment. The team developed a model based on the "random forest" algorithm, which takes into account specific parameter settings in production, such as extreme temperatures and temperature changes, as well as information on surfactant product types and formulations.