16 Oct Grinding on Autopilot
The integration of artificial intelligence and advanced process control in the grinding processes of the modern cement plant is leading to optimised grinding performance, higher output rates, lower specific electrical consumption and fewer fluctuations in fineness.
The cement industry is undergoing a significant transformation, driven by a range of challenges including climate engagement initiatives, digital transformation, energy cost fluctuations and market volatilities.
At the forefront of this revolution, equipment suppliers are pioneering process innovation to assist their customers. Leveraging breakthrough technologies, they enable the complete autopiloting of cement manufacturing processes, boosting production volumes, curbing energy consumption, and consistently enhancing product quality and uniformity. The aim is to meet the diverse demands of end-users by offering environmentally friendly cement options whilst simultaneously lowering operational costs.
For over a decade, ES Processing has adopted a proactive strategy, leading the development of the first APC-AI based solution to autopilot ball mill operations, introducing artificial intelligence (AI) and advanced process control (APC) as invaluable tools for optimising the grinding process in cement manufacturing.
Since 2012 the company’s optimization solutions for ball and vertical mills have demonstrated their resilience by continually enhancing cement quality, improving process performance and reducing specific energy consumption.
APC-AI integration for improved cement grinding
In conventional cement plant operations, the central control room (CCR) operator relies on cement quality analysis and other process indicators to make decisions regarding mill actions, drawing on experience to anticipate process reactions and make calculated adjustments. However, to achieve enhanced performance and stable operation, a computational tool is indispensable. This tool must be able to utilise current process indicators and historical plant conditions to accurately predict future plant behaviour and compute a series of control actions.
Introducing these autopiloting solutions as the cornerstone of manufacturing excellence, ES Processing integrates AI into the core of the cement industry. The company’s CMO/VMO solutions use AI, machine learning (ML) and APC to optimize the operation of ball mills and vertical mills, respectively.
The model predictive control (MPC) module integrated within the CMO/VMO solutions leverages advanced numerical algorithms derived from plant data. It employs a strategy analogous to a game of chess, continuously and simultaneously adjusting different manipulated variables of the mill, such as fresh feed, separator speed and hydraulic pressure, every 30s while predicting process reactions for the subsequent 120 moves (equivalent to 1h of operation) – a significant improvement over traditional hourly sampling routines. Variables are continuously adapted to changing process conditions such as material quality, feed rates, rejects and pressure. Ultimately, the MPC module dynamically controls and stabilises the mill in real-time, ensuring the production of a consistently uniform cement at the targeted quality. This optimized performance results in higher production volumes at lower costs.
Optimising the grinding process
Cement grinding processes are highly nonlinear and unstable, directly impacting cement quality, plant performance and operational costs. Hence, the primary objective of the CMO/VMO optimisers is to maximise the production of consistent cement according to the desired quality targets. These optimisers play a pivotal role in refining product quality, significantly minimising variations, whilst concurrently increasing output and stabilising operations.
They bring about substantial improvements in cement fineness, ultimately resulting in improved overall production rates and reduced specific energy consumption. By continuously assessing the optimal setpoints for mill manipulated variables, CMO/VMO ensure seamless operations while guaranteeing consistent cement fineness and optimal performance.
The optimisers include a complex combination of:
Soft Sensor technology: very advanced but stable models formed by combining multiple data-based algorithms adopted from ML, based on linear and non-linear identification techniques, and genetic algorithms that determine the best correlation between different process parameters and product quality results. Therefore, they are able to regularly predict the main cement quality indicators with high accuracy, with a precision of three sigma. These Soft Sensors rely on engineering correlations between historical process and quality data to infer key parameters such as cement fineness from a defined set of inputs exhibiting strong correlations with the outputs. The Soft Sensors provide continuous feedback to control systems in the absence of laboratory data. As laboratory measurements become available, the Soft Sensors dynamically enhance their learning curve and finetune their predictions accordingly.
Model predictive control (MPC): a highly complex multivariable model based on transfer functions built according to impulse test results performed on each piece of equipment. These are able to handle complex plant dynamics, including long-dead times and non-minimum phase behaviour, constraint handling, hierarchical and weighted optimisation and predictive control, thus able to adjust the grinding manipulated variables every 30s.
With this combination, the plant will consistently strive towards operational objectives by leveraging continuous prediction of cement fineness and process reactions, along with optimal adjustments to operating conditions.
This resembles having a multitude of top-tier, highly-skilled operators continuously monitoring all process parameters, including online product analysis, working in unison to regularly fine-tune process parameters and optimize the entire operation.
As a true autopilot system, the CMO/VMO solution has led to transformative shifts in efficiency, sustainability and profitability for a range of clients across the globe.
Ball mill performance in terms of product fineness before (left) and after (right) the introduction of the CMO system.
Benefits in cement plants
CMO and VMO autopilot systems have been integrated into diverse cement grinding processes across the globe, where they have provided significant benefits.
Portugal-based SECIL installed a CMO/VMO autopilot system to improve the cement mill performance at its Sibline plant, Lebanon. Following installation, benefits included:
- >10 per cent increase in production rate
- >10 per cent reduction in specific electrical energy consumption
- >50 per cent reduction in standard deviation of Blaine.
The installation of a CMO/VMO autopilot system at one of Titan’s cement plants resulted in:
- >15 per cent increase in production rate
- >8 per cent reduction in specific electrical energy consumption
- >50 per cent reduction in standard deviation of Blaine.
The CMO/VMO installation for the cement mill at Saman’s plant in Iraq led to:
- >10 per cent increase in production rate
- >8 per cent reduction in specific electrical energy consum