Companies use artificial intelligence targeting productivity and reducing environmental impact

Artificial intelligence has become a new ally for large companies in their quest to increase productivity, learn about the environmental impact of their production and improve sustainability indicators.

Information collected by sensors on the machines and analysis of databases with production history and other variables, such as weather and market, have optimized the functioning of the machines and the use of resources in the industry and in the field.

Examples include the solutions adopted by Braskem, in the petrochemical sector. using algorithms to define the best formulation when producing recycled plastics.

Guilherme Baeta, chief technology officer (technical director) of the company, says that the work combines virgin resin —produced in the company itself— with the plastic consumed and collected for reintroduction in the market.

The difference in quality control that it is possible to have in these two types of materials makes it more complex to define which final product to create and the best formulation to reach it, explains the executive.

According to Baeta, these decisions are being supported by artificial intelligence, which points out the five best alternatives according to the scenario. He says that the production of recycled materials started from an incipient volume in 2020. Now the company seeks to accelerate production to reach 300 thousand tons of products of this type sold by 2025.

The analysis of information also allowed Braskem to reduce by 20% to 40% the volume of products that were out of specification at 20 of its units, this time considering items made with new raw materials.

Baeta claims that the result was obtained by incorporating intelligence to offer suggestions at a time when the industry is changing the type of resin it is producing.

This became possible because the software started to indicate the formulation and method of preparation for each type of plastic according to the situation.

Previously, it was necessary to wait for the reactor to run for each new configuration and then take a sample to check whether the product was adequate. The process, which could take one or two hours to reach an adequate result, has become almost automatic, says the executive.

The petrochemical company is also accelerating a startup created by its own employees that involves data and sustainability.

Called S-Index, the new company aims to speed up the process called Life Cycle Analysis, a technique adopted to understand the environmental footprint of a product and evaluate options to reduce it.

Baeta says that, based on Braskem’s own information, parameterization made by specialists and market data, it is possible that analyzes with good precision will be carried out in hours, instead of taking three or four months.

The startup tests its products on four large customers.

In order to centralize the data collected by sensors in its factories, including water and energy consumption, Nestlé inaugurated in December a technical center in the city of Araras (inland São Paulo).

Gustavo Moura, the company’s digital transformation manager, says that the greater volume of information makes it possible to create alerts and carry out quick interventions, avoiding waste and loss of productivity.

It is possible, for example, to anticipate a pipe blockage and avoid hours of downtime due to the problem. Artificial intelligence can also indicate the best time to clean the machines.

In the production of biscuits, variables such as temperature and humidity are analyzed to define the best configuration of the industrial oven, in which the least possible amount of gas is consumed without losing efficiency.

At the Gerdau plant in Ouro Branco (MG), the melt shop’s production is mirrored by a system known in the market as the digital twin.

“We virtualized the entire physical environment and created a parallel environment that allows us to analyze the best type of steel to be produced for each moment and the best sequencing of products”, says Gustavo França, the company’s global digital technology leader.

The technology also appears in three other factories in Minas Gerais. In them, the objective is to optimize the use of charcoal produced by Gerdau itself. The software is used by the company to define the strategy for using and storing the material.

“As you touch on the issue of productivity, there are gains in sustainability”, says França.

In agribusiness, artificial intelligence has been tested as a tool to improve productivity and, therefore, reduce the consumption of water, seeds and agricultural pesticides, in addition to ensuring better use of the land.

At Bayer, tools based on artificial intelligence are offered from the Climate Field View unit.

Abdalah Novaes, leader of digital solutions for Bayer’s agricultural division for Latin America, explains that the platform gathers data from satellites with information sent by the producers’ equipment used in planting, fertilizing, spraying and harvesting.

“At the end of a cycle, the producer has information about productivity in each area and receives guidance on how to have better results in the next harvest.

Novaes says that the system can suggest changes in parameters such as the speed of the machines when planting and the amount of seeds and water to be used in each region of the property.

“Today, and more and more, it will be necessary to expand the production that can be carried out in the same space of land”, he says.