Integrated product approach from cement to concrete
CO2 - opportunity and challenge at the same time
Cement production has the highest carbon footprint per euro of sales. In addition, the price of CO2 is increasing rapidly in Europe and is already a determining economic factor. This trend will accelerate significantly as the free allocation of certificates expires by 2030. In the end, only those cement and concrete manufacturers who manage to decarbonize quickly and efficiently will remain competitive. The coming years will decide who will prevail in the market with innovative and low-CO2 portfolios.
Efforts of cement and concrete manufacturers
The greenhouse gas CO2 is released when the limestone is burned into cement clinker through a chemical reaction. The path is clearly marked: over the next few years, cement manufacturers will gradually have to include more and more clinker-reduced products in their portfolios.
At the same time, this development means that concrete manufacturers have to deal more intensively with the handling of these clinker-reduced cements.
In the concrete application, these are much more complex than conventional cements due to the required precision and standard specifications. For example, tipping points are encountered much more quickly, which means that the required strength is not achieved, the concrete separates or stiffens too much.
A decarbonized industry therefore means a joint effort for cement and concrete manufacturers, in which they have to work hand in hand to bring low-CO2 products into mass use.
An integrated product approach made by alcemy
alcemy's product portfolio follows the value chain from the cradle in cement grinding and mixing to the processing of the concrete on the construction site. In this way, we support cement and concrete manufacturers alike in the challenges of this new age. Our intelligent quality monitoring and automation solution is based on the same approach in both products: data collected during production in the laboratory or via sensors in the factory or truck is sent to alcemy via interfaces.
alcemy’s technology…
- monitors whether problems or deviations occur in production and draws the staff’s attention to them.
- predicts quality parameters of the current production based on the data.
- gives factory personnel recommendations for settings to ensure optimal quality.
alcemy's products
alcemy for cement:
From reactive to proactive quality control with the help of machine learning.
Our AI software for intelligent quality control is the tool for producing cements of the highest quality. With this new approach, we support manufacturing companies in mastering increasingly complex cement recipes for a decarbonized future.
The challenge
In the final cement manufacturing step, the raw materials, including the clinker, are ground into cement. In addition to other parameters, the strength is decisive for the quality. The determination of strength is measured by creating a standard concrete after curing for one, two, seven and 28 days. During production, there is no insight into the future strength of the cement currently being produced and the strength can only be controlled on the basis of the experience of the control room drivers and quality and production employees.
Our approach
Our technology enables a continuous analysis of the quality-relevant data from chemistry, mineralogy and grain size distribution. With our intelligent algorithms, target values for optimizing the current cement production are continuously passed on to the control center.
0. Model training thanks to impact: The alcemy model is trained based on values that already exist in the cement plant. In addition to the chemical and mineralogical properties, this also includes the results achieved, such as the strength after curing for one to 28 days.
1. Strength prediction: Just as the properties and qualities of natural resources fluctuate, all fluctuations must also be taken into account and compensated for by correct control when producing homogeneous and qualitatively stable cement. Our models predict quality properties such as compressive strength and Blaine of the current grind or blend. This means that the fluctuations in resources are not only visible in the raw data, but also in the target values.
2. Optimization: However, it does not just remain with the pure strength prediction, but the technology uses the knowledge gained to create control specifications for the cement mill for the current production and to find optimal target values for fineness or the right recipe to achieve the desired target properties. These are then transmitted to the control room operator or imported directly into the process control system.
Benefits for our customers
All alcemy customers were able to improve their quality by 20% to 50% up to an average standard deviation of 1.5 MPa after six months and have thus become quality leaders in a very short time.
35% quality improvements (hrs) across all cement families
⌀ €400,000 in cost savings by reducing the clinker factor and grinding energy
⌀ 15,000 tons less CO2 emissions per customer
Increase in efficiency & gain of information through a central database for all parties involved
Our application in detail
All information on cement control at a glance.
X-ray view of the cement qualities.
Manage recipes and connect them with the findings from the analyses.
“alcemy simplifies my everyday work, because all target values are always up-to-date in one place, which eliminates any misunderstandings. I’m a big fan of alcemy’s software.”
Markus Bollich, control room operator at the SPENNER cement plant in Erwitte
alcemy for concrete:
Real-time monitoring and quality control for ready-mixed concrete using machine learning.
Our technology uses the data in the ready-mixed concrete plant and truck mixer, continuously predicts the quality properties of the concrete in real time and translates these into KW target values for the plant. This means that complaints on the construction site and inefficiencies in the factory are a thing of the past.
The challenge
At the moment, quality control in the ready-mixed concrete plant is still largely based on the instinct and many years of experience of the employees. The tools for quality control, such as slump tests or consistency comparison curves, are still mostly created manually and are therefore very time-consuming.
Our approach
Our AI-supported software, which runs in the cloud and collects data in the ready-mixed concrete plant as well as in the truck mixer, gives laboratory technicians and mixing supervisors the right digital tools to make the manual work steps much easier, more automated and more efficient.
The alcemy concrete product consists of two modules:
The monitoring module uses data generated during production and travel to the construction site and uses this to continuously forecast the expected quality properties of the concrete. This is possible because the algorithms are trained with historical data for each plant. The constant forecasts can be viewed by employees from production and quality assurance at any time and create a completely new level of transparency and information.
The control module is responsible for the recipe control. Forecasts are used to identify deviations in advance and to transmit them to the plant quickly and easily. For example, concrete that is otherwise too stiff can be prevented from the outset by adding more water. Our software calculates supporting KW target values for the active power and water addition and continuously transmits them to the mixing supervisors.
Benefits for our customers
During our three-month series of tests, the water balance was manually intervened in 96% of all batches. With the help of alcemy technology, manual intervention was no longer necessary. In addition, the fluctuations in the slump have been reduced by up to 50%.
Material cost reduction
Reduction of quality fluctuations by up to 40%
80% less manual intervention regarding sand moisture
Raising the revenue level
reduction in complaints
Our application in detail
Seamless quality monitoring from production to handover.
Oil pressure, water addition and drum performance can be easily monitored by our comprehensive remote support of the truck mixer. In this way, the plant is informed at all times about changes while the truck mixer is in motion and can track all changes in real time until the material is handed over to the customer. This enables seamless quality assurance from production to delivery of the concrete.
Every construction site at a glance.
Thanks to the real-time display, measured values and predicted concrete properties can be viewed transparently – and clearly for all truck mixers that are currently on the road. Alcemy supports you in completing values that are usually difficult to measure with detailed forecasts. Outliers in the values are highlighted in real time for a quick overview of possible problems.
Parallel optimization of batches.
The seamless monitoring of the truck mixers that have already started enables a parallel optimization of the ready-mixed concrete for the next batches in the plant. With the help of reliable data from the truck mixer and the plant, alcemy can be used to make predictions about the required target values for the next batch. This guarantees optimal quality when it is handed over to the customer.
“We believe that the concrete plant of the future will resemble a factory that can flexibly combine many raw materials. In order to still guarantee high quality, adaptive and autonomous quality control systems will be indispensable.”
dr Olaf Assbrock, General Manager of the Federal Association of the German Ready-Mix Concrete Industry
Basically, the same technology is behind our products for cement and concrete production; we use machine learning for predictive quality control in cement and ready-mixed concrete plants.
data collection
During cement and concrete production, vast amounts of data are generated that have so far remained unused. We collect this data, merge it and make it usable. To do this, we integrate our software with laboratory information systems and install our own sensors in concrete truck mixers. This enables us to collect data along the entire value chain.
training and evaluation
We analyze the collected data and identify use cases in which we can support our customers in their daily processes with statistical analyzes and machine learning. This can be, for example, the prediction of quality properties of cement and concrete, such as compressive strength, Blaine or slump. We have high quality standards for our models because the experts in cement and concrete production have to rely on them. We regularly retrain our models, continuously evaluate them and develop them further.
Optimization of production processes
Based on our models, we determine optimal target values for various production steps in cement and concrete production in order to achieve the desired quality goals. The target values are fed back to the control centers in the plants or transmitted directly to the expert systems that control the systems.
Data-driven development
In addition to the automatically running processes, we offer the experts in cement and concrete insights into their data from a new data-driven perspective with our software. This combination of expert knowledge in cement and concrete on the one hand and data science on the other results in unprecedented opportunities and insights.
machine learning infrastructure
Our model training and hyperparameter optimization needs a reliable and scalable infrastructure. That’s why our software runs on the cloud-based solution Kubernetes. Furthermore, we use “Infrastructure as Code” with terraform to implement changes to our productive system quickly and securely. We also have the ability to create test environments that copy our entire application – from the database to the web app. This enables us to develop software quickly, flexibly and securely.
alcemy for cement:
From reactive to proactive quality control with the help of machine learning.
Our AI software for intelligent quality control is the tool for producing cements of the highest quality. With this new approach, we support manufacturing companies in mastering increasingly complex cement recipes for a decarbonized future.
The challenge
In the final cement manufacturing step, the raw materials, including the clinker, are ground into cement. In addition to other parameters, the strength is decisive for the quality. The determination of strength is measured by creating a standard concrete after curing for one, two, seven and 28 days. During production, there is no insight into the future strength of the cement currently being produced and the strength can only be controlled on the basis of the experience of the control room drivers and quality and production employees.
Our approach
Our technology enables a continuous analysis of the quality-relevant data from chemistry, mineralogy and grain size distribution. With our intelligent algorithms, target values for optimizing the current cement production are continuously passed on to the control center.
0. Model training thanks to impact: The alcemy model is trained based on values that already exist in the cement plant. In addition to the chemical and mineralogical properties, this also includes the results achieved, such as the strength after curing for one to 28 days.
1. Strength prediction: Just as the properties and qualities of natural resources fluctuate, all fluctuations must also be taken into account and compensated for by correct control when producing homogeneous and qualitatively stable cement. Our models predict quality properties such as compressive strength and Blaine of the current grind or blend. This means that the fluctuations in resources are not only visible in the raw data, but also in the target values.
2. Optimization: However, it does not just remain with the pure strength prediction, but the technology uses the knowledge gained to create control specifications for the cement mill for the current production and to find optimal target values for fineness or the right recipe to achieve the desired target properties. These are then transmitted to the control room operator or imported directly into the process control system.
Benefits for our customers
All alcemy customers were able to improve their quality by 20% to 50% up to an average standard deviation of 1.5 MPa after six months and have thus become quality leaders in a very short time.
35% quality improvements (hrs) across all cement families
⌀ €400,000 in cost savings by reducing the clinker factor and grinding energy
⌀ 15,000 tons less CO2 emissions per customer
Increase in efficiency & gain of information through a central database for all parties involved
Our application in detail
All information on cement control at a glance.
X-ray view of the cement qualities.
Manage recipes and connect them with the findings from the analyses.
"alcemy simplifies my everyday work, because all target values are always up-to-date in one place, which eliminates any misunderstandings. I'm a big fan of alcemy's software."
Markus Bollich, control room operator at the SPENNER cement plant in Erwitte
alcemy for concrete:
Real-time monitoring and quality control for ready-mixed concrete using machine learning.
Our technology uses the data in the ready-mixed concrete plant and truck mixer, continuously predicts the quality properties of the concrete in real time and translates these into KW target values for the plant. This means that complaints on the construction site and inefficiencies in the factory are a thing of the past.
The challenge
At the moment, quality control in the ready-mixed concrete plant is still largely based on the instinct and many years of experience of the employees. The tools for quality control, such as slump tests or consistency comparison curves, are still mostly created manually and are therefore very time-consuming.
Our approach
Our AI-supported software, which runs in the cloud and collects data in the ready-mixed concrete plant as well as in the truck mixer, gives laboratory technicians and mixing supervisors the right digital tools to make the manual work steps much easier, more automated and more efficient.
The alcemy concrete product consists of two modules:
The monitoring module uses data generated during production and travel to the construction site and uses this to continuously forecast the expected quality properties of the concrete. This is possible because the algorithms are trained with historical data for each plant. The constant forecasts can be viewed by employees from production and quality assurance at any time and create a completely new level of transparency and information.
The control module is responsible for the recipe control. Forecasts are used to identify deviations in advance and to transmit them to the plant quickly and easily. For example, concrete that is otherwise too stiff can be prevented from the outset by adding more water. Our software calculates supporting KW target values for the active power and water addition and continuously transmits them to the mixing supervisors.
Benefits for our customers
During our three-month series of tests, the water balance was manually intervened in 96% of all batches. With the help of alcemy technology, manual intervention was no longer necessary. In addition, the fluctuations in the slump have been reduced by up to 50%.
Material cost reduction
Reduction of quality fluctuations by up to 40%
80% less manual intervention regarding sand moisture
Raising the revenue level
reduction in complaints
Our application in detail
Seamless quality monitoring from production to handover.
Oil pressure, water addition and drum performance can be easily monitored by our comprehensive remote support of the truck mixer. In this way, the plant is informed at all times about changes while the truck mixer is in motion and can track all changes in real time until the material is handed over to the customer. This enables seamless quality assurance from production to delivery of the concrete.
Every construction site at a glance.
Thanks to the real-time display, measured values and predicted concrete properties can be viewed transparently – and clearly for all truck mixers that are currently on the road. Alcemy supports you in completing values that are usually difficult to measure with detailed forecasts. Outliers in the values are highlighted in real time for a quick overview of possible problems.
Parallel optimization of batches.
The seamless monitoring of the truck mixers that have already started enables a parallel optimization of the ready-mixed concrete for the next batches in the plant. With the help of reliable data from the truck mixer and the plant, alcemy can be used to make predictions about the required target values for the next batch. This guarantees optimal quality when it is handed over to the customer.
“We believe that the concrete plant of the future will resemble a factory that can flexibly combine many raw materials. In order to still guarantee high quality, adaptive and autonomous quality control systems will be indispensable.”
dr Olaf Assbrock, General Manager of the Federal Association of the German Ready-Mix Concrete Industry
Basically, the same technology is behind our products for cement and concrete production; we use machine learning for predictive quality control in cement and ready-mixed concrete plants.
data collection
During cement and concrete production, vast amounts of data are generated that have so far remained unused. We collect this data, merge it and make it usable. To do this, we integrate our software with laboratory information systems and install our own sensors in concrete truck mixers. This enables us to collect data along the entire value chain.
training and evaluation
We analyze the collected data and identify use cases in which we can support our customers in their daily processes with statistical analyzes and machine learning. This can be, for example, the prediction of quality properties of cement and concrete, such as compressive strength, Blaine or slump. We have high quality standards for our models because the experts in cement and concrete production have to rely on them. We regularly retrain our models, continuously evaluate them and develop them further.
Optimization of production processes
Based on our models, we determine optimal target values for various production steps in cement and concrete production in order to achieve the desired quality goals. The target values are fed back to the control centers in the plants or transmitted directly to the expert systems that control the systems.
Data-driven development
In addition to the automatically running processes, we offer the experts in cement and concrete insights into their data from a new data-driven perspective with our software. This combination of expert knowledge in cement and concrete on the one hand and data science on the other results in unprecedented opportunities and insights.
machine learning infrastructure
Our model training and hyperparameter optimization needs a reliable and scalable infrastructure. That’s why our software runs on the cloud-based solution Kubernetes. Furthermore, we use “Infrastructure as Code” with terraform to implement changes to our productive system quickly and securely. We also have the ability to create test environments that copy our entire application – from the database to the web app. This enables us to develop software quickly, flexibly and securely.
