Cement Raw Materials (MAT): Complete Guide

Mathematics and Control Engineering of Grinding Technology is the systematic mathematical volume of the grinding process: the book by L. Keviczky, M. Hilger and J. Kolostori, published in the East European applied mathematics series: the volume applies the full machinery of the modern control theory to the grinding circuits: the comminution kinetics, the mill models, the state equations, the identification methods, the optimal control, the simulation and the practical implementation of the adaptive controllers in the industrial plants: the honest note: the catalog of the package lists this file under the truncated label “Mat” because the original title was cut in the recording; the verified content of the PDF is the mathematics and the control of the grinding technology of exactly this monograph.

The Complete Cement Technical Package (931 files including the books, the Excel tools, the courses and the presentations: $249.99 one-time: instant download via the PayPal payment) includes this mathematical monograph PDF with its equations, its block diagrams and its experimental results: the process engineer, the control engineer and the graduate student of the automation find in it the complete treatment of the grinding circuit as a dynamical system: this article walks the book: the grinding models first, the state space representation second, the identification third, the control fourth, the practical implementation last: the reader finishes with the mathematical map of the closed-loop grinding mill.

Why the mathematics of the grinding matter: the finish grinding of the cement consumes 30-40% of the plant electric energy, the raw grinding another 25-30: the improvement of a few percent in the efficiency of the separator, the recirculation and the control of the mill appears directly in the plant’s power bill: and the grinding circuit is also the most dynamic system of the plant: the feed arrives with the varying hardness of the clinker, the mill degrades with the ball charge, the separator drifts with the wear, and the human operator reacts after the fact: the mathematical control of the mill closes the loop at the speed of the machine instead of the speed of the man: the monograph of Keviczky and the colleagues is the classical foundation of the mill automation that the modern DCS and the AI controllers implement directly.

1. The Grinding Process as a Dynamical System

The first principle of the book: the grinding mill is not a static map from the feed to the product: it is a dynamical system with the internal state, the delays and the disturbances:

• The state of the mill: the material inside the mill (the holdup) and its size distribution, the media charge, the moisture, the temperature: the state variables the model must track: the inventory of the powder inside the mill is the heart of the dynamics: the classical mills carry 15-30% of the feed flow as the internal holdup, which gives the mill the response time of the minutes;
• The time constants: the response of the product fineness to the feed rate change: the 5-15 minutes in the closed circuit ball mill: the response of the separator speed usually the fast, the response of the feed the medium, the response of the classifier the slower: the control engineer must know the time scales before the controller design;
• The disturbances: the moisture swings of the feed, the hardness of the rock, the wear of the media, the blinding of the separator, the seasonal temperature: the disturbances are the reason the static setpoint does not work and the adaptive and the robust control appear: the book classifies the disturbances with the design responses;

The dynamical view changes the engineering: the mill is not treated as the black box of the “load cells and the electricity” but as the system with the internal state, the memory and the response: the control of the mill is the control of the dynamic system, and the mathematical framework of the book is built on the first-principles and the empirical models of this first chapter.

The time scales matter even more than the chemistry here: the mill responds over the minutes, the separator over the seconds and the operator over the quarter hour: the model of the book works on the time grid that captures all three: the sampling of the data at the minute, the simulation step of the second: the honesty of the book: it tells the reader that the mill time scale is the reason the simple manual control drifts and the reason the measured-state control works: the engineer who internalizes the time constants reads the trend graphs of the DCS with completely different eyes: he sees the state of the mill, not just the last line of the product.

2. The Models of the Comminution

The second chapter builds the grinding models: the comminution equations that describe the size reduction:

• The size-discretized models: the batch grinding equation (the selection and the breakage functions), the population balance: the continuous version with the residence time distribution: the S(x) selection and the B(x) breakage functions are measured on the mill: the sampling of the mill contents and the sieving: the book gives the experimental procedure of the determination of the breakage rates;
• The Rosin-Rammler and the Gaudin-Schuhmann distributions: the classical distribution functions of the ground product, the parameters n and the size constants: the modeling of the mill product with the two-parameter distributions: the use in the separator model:
• The energy laws: the Kick, the Rittinger, the Bond work index: the comminution energy as the integral of the force-displacement: the selection of the law by the size range: the Bond index as the industrial standard:
• The breakage and the grinding kinetics: the matrix formulation of the grinding: the breakage matrix the selection matrix and the states: the linear approximation of the milling for the control purposes: the book develops the linear and the nonlinear variants with the credibility of the validation on the real mills:

The comminution models of the chapter are the vocabulary of the whole monograph: every later control scheme is built on the equations of the breakage, and every simulation of the plant starts with the S and B functions of the specific mill: the engineer of the chapter understands exactly what the mill does per revolution of the media.

The models of the chapter have one purpose: the prediction: the model that predicts the mill response to the feed change, the media change and the separator change is the model that the control uses: the book grades the models by the input that was measured for the calibration: the bond models are the coarse, the specificoric models the usable, and the digital twins of the modern plants are built exactly on the same architecture that the monograph establishes: the reader of the chapter leaves with the understanding that the model is not the art on the slide: it is the earned approximation of the physics of the mill, verified by the grinding tests.

3. The Structure of the Grinding Circuit and Its Identification

The third chapter connects the model to the circuit: the closed loop of the mill with the separator and the recirculation:

• The circuit forms: the open loop, the closed loop with the bucket elevator and the classifier, the closed loop with the dynamic separator: the recirculation ratio: typically 100-300% of the fresh feed: the internal and the external recirculation: the mathematics of the steady-state of the recycle:
• The variables of the circuit: the fresh feed, the recirculation, the fines product, the coarse reject: the three-way stream splitting and the mass balance of the circuit: the residence time distribution of the recycle loop: the book derives the transfer function of the loop: the loop gain and the stability:
• The identification methods: the step tests and the PRBS excitation (the pseudo-random binary sequences), the cross-correlation, the least squares estimation, the linear regression of the mill parameters: the identifications on the real data: the book presents the measured identification of the industrial cement mill: the time constants and the gains from the data: the modern plant repeats the same tests with the DCS trending system:

The identification chapter is the bridge between the theory and the plant: the model of the specific mill cannot be bought: it must be measured: the book gives the sequence: the test design, the data collection, the estimation, the validation: the plant engineer of the package reads the chapter and runs the step test on his own mill: this is the chapter that converts the mathematics into the plant data.

The plant test procedure of the book deserves the detail: the mill is isolated from the disturbances, the feed is stepped by the 5-10% and held for the full residence time, the power, the fineness and the recirculation are recorded at the minute intervals, and the identification software fits the transfer function: the test lasts the 2-4 hours of the quiet shift: the danger of the short test: the model captures the transient and misses the steady state: the book warns against the “quick identification” and demonstrates the difference on the data: the discipline of the identification is the discipline of the honest model, and the plant engineer who follows the book’s protocol earns the trust of the controller he builds on the model.

4. The State Space, the Transfer Functions and the Simulation

The fourth chapter builds the complete dynamic representation:

• The state space model: the A, B, C, D matrices of the circuit: the states as the hoppers, the separator dynamics, the mill inventory: the observability and the controllability of the mill circuit: the book demonstrates that the mill circuit is both observable and controllable with the standard instrumentation:
• The transfer functions: the loop from the feed setpoint to the product fineness, the disturbance rejection chains: the pole at the origin of the integrations, the time lags, the dead time of the transport: the Bode analysis of the circuit:
• The Kalman filter: the state estimation of the mill from the noisy measurements: the filter of the cement mill with the online fineness: the estimated state feeds the controller: the modern plant implements the state estimators in the mill control software:
• The simulation: the discrete time simulation of the circuit model: the step responses, the closed loop runs: the book’s FORTRAN era algorithms reconstructed in the modern tools: the simulation as the design laboratory of the controllers:

The fourth chapter is the toolbox chapter: the heavy mathematics of the package: the plant engineer who endures it comes out with the ability to put his mill into the phase space of the equation: the engineers of the modern cement plant do the same in the process simulator software, and the mathematics of the book explains what the software computes.

The chapter 4 also gives the stability discipline: the engineering judgment of the mill circuit in the frequency domain: the phase margins and the gain margins of the open loop, the delay budget of the controller: the book’s design rules: the gain margin above 6 dB and the phase margin between 30-60 degrees keep the circuit comfortable with the parameter drift: the stability margins are the bridge between the abstract matrices and the real plant: the same margins that the modern controller software displays on the tuning screen, the book computes by hand and the reader keeps in the head.

5. The Control of the Grinding Circuit: The Loops

The fifth chapter crosses to the control: the architecture of the grinding control:

• The inner loops: the feed flow loop, the mill power loop, the separator speed loop: the PID of each: the cascade of the feed to the power: the inner loop of the fan and the draft:
• The outer loop: the fineness control: the measurement of the fineness online (the sampling, the laser diffraction) or the inferred from the power: the setpoint of the fineness target: the outer loop the sample:
• The model based control: the internal model control, the feed-forward of the hardness disturbances: the adaptive: the gain scheduling of the hardness: the multi-variable control of the feed and the separator: the book presents the classical and the modern structures head to head:
• The constraint handling: the mill power limits, the blower: the constraints of the mill motor, the max recirculation: the control with the anti-windup: the machine protection and the automation coexist:

The chapter emphasizes the hierarchy: the inner loops must be fast and robust, the outer loops slow and accurate, and the disturbances must be measured and preview where they can: the grinding control design in this chapter is the same as the modern Success of the mill: the same at 2 in the morning.

The tuning of the loops receives its own treatment in the book: the classical reaction-curve tuning and the lambda tuning of the closed loop and the model-based internal model control for the dead-time-dominant circuits: the book compares the tunings on the same simulated mill and shows that the aggressive tuner destabilizes the loop at the moment the separator separation changes: the advice of the book: the tuning must respect the process lag, not the operator’s impatience: the plants that pass through the saturation are the plants with the anti-windup problems: the chapter closes with the tuning constants that the process engineers of the cement industry still use on the DCS console.

6. The Nonlinearities and Difficulties of the Real Mill

The book does not pretend the mill is easy: the sixth chapter of the nonlinearities:

• The saturations: the classifier saturation, the motor limits, the flow saturations: the controller must respect the physical limits: the windup and the reset:
• The time-varying behavior: the mill charge changes with the media wear, the hardness of the seam change daily: the seasonal humidity, the aging of the lining: the adaptive controllers and the self-tuning regulators: the gain the controller themselves:
• The instrument problems: the weighing errors, the sampler dead time, the sensor drift: the fault detection and the data validation: the book’s treatment of the “hard data” in the loop: the estimator that knows the sensor quality:
• The manual-automatic interface: the operator trust of the regulator: the bumpless manual/auto transfer, the operator overrides and the alarm limits: the book acknowledges the human inside the process: the controller is a partner of the operator, not the replacement of the hourly vigilance:

The chapter of the practical difficulties is the honest chapter: the mathematical elegance hits the plant dirt, and the designer needs the saturation models and the anti-windup just to keep the loop running: the book: the measurements of the real mill under the control: the improvement of the plant supply, the reduction of the deviation: the numbers of the real plants in the appendix.

7. The Real Plant Results of the Control

The application chapters of the book bring the measured results:

• The field trials: the implementation of the adaptive controller on the real industrial mills: the results of the book: the one-third reduced deviation of the fineness, the increased average, the mill power savings: the book’s reported 2-5% energy: the productivity 3-10%: the discipline of the measurement: the before/after:
• The “row” plants: the plants of the monograph: the coal mills and the cement mills of the Hungarian industry: the trials reported in the engineering detail: the actual data: the time series of the operation, the controller performance and the week-long field runs:
• The simulation tests: the designed controllers exhaustively tested in the simulation before the plant: the case: the model the simulation the plant trial the production: the methodology of the chapter teaches the same discipline to the reader: never the plant test the untested controller:
• The lessons of the implementations: the identification time, the staff, the loop tuning, the fail: the practical list of the “the controllers did not work” times: the conditions for the success of the advanced control in the plant:

The results chapters are the “evidence” of the book: the mathematical framework, the nonlinearities and the identification come together into the improved production: the numbers of the field are the reason the industry of the advanced process control (the expert systems, the APC of the cement plants today) grew from the school of the 1970s: the book is both the root and the contemporary of the modern mill control.

The evidence chapter also carries the economics: the control of the grinding returns the money: the fewer the stops and the less the deviation at the quality limit, the more cement at the same energy: the book’s reported field values: the increase of the mill utilisation, the reduction of the specific electric energy and the reduction of the reagent waste: the amortization of the automation projects in the case plants: the payback in the months, not in the years: the engineer of the plant plans the automation project with these numbers on the feasibility sheet: the control is the investment that pays in the operation.

8. The Control of the Raw Mill and the Coal Mill

The book extends beyond the finish mill: the control of the other mills of the plant:

• The raw mill: the drying and the grinding combined: the mill gas temperature, the moisture of the product, the fineness: the content of the drying gas and the feed rate: the raw mill is the closed loop of the counter-inside process: the constraints of the gas flow:
• The coal mill: the coal drying and the explosion safety: the mill outlet temperature limits, the CO monitoring, the inert zones: the control of the coal supply to the kiln and the calciner: the safety interlocks of the mill once:
• The common architecture: the cascade loops of all the mills: the material feed, the separator speed, the gas flow, the temperature: the parameter estimation on all: the book generalizes the methods: the same framework the three mills:

The generalization of the chapter: the grinding sections of the plant are the same dynamical family: the mathematics trained on the cement mill apply to the raw mill with the gas state included and to the coal mill with the safety: the plant engineer with the book does not need the separate theory for every sprinkled unit: the one framework: the adaptations.

The coordination of the three mills inside the plant: the raw mill is coupled to the kiln gas (the drying duty and the draft), the coal mill is coupled to the kiln fuel demand, and the cement mills are coupled to the dispatch plan: the book presents the supervisory hierarchy that integrates the mills as the members of the plant-wide closed loop: the modern cement plant operates this hierarchy in the plant DCS with the grade scheduling, and the mathematics of the chapter is the description of that integration: the control engineer who reads the book sees the whole plant as the system, with the mills as its actuators.

9. The Simulation Environment of the Book

The numerical side of the monograph: the simulation experiments:

• The discrete simulation models: the block-diagrams of the circuit in the computational environment: the integration of the differentials: the typical runs: the step tests of the simulator: the behavior of the loop under the model: the same environment delivers the design trials of the controllers:
• The identification experiments: the simulation-driven testing: the PRBS input, the noise: the estimation of the same parameters: the accuracy: the comparison of the estimators: the book’s laboratory is a virtual mill:
• The optimal control runs: the time of the controller on the model: the LQ, the adaptive: the comparison of the performance of the run: the criterion of the min the variance of the product quality:

The simulation: the authors develop the software of the experiments in the conventional numbers: the modern reader rebuilds the identical experiments in the Python and the MATLAB: the methodology of the book: the virtual: the model-before-plant: the chapter teaches the discipline of the simulation as the tool of the control engineering: today’s operator does the same with the graph in real time.

The simulation-explicit discipline of the chapter: the virtual commissioning: the controller is connected to the simulated mill before it touches the real mill: the plant staff trains on the simulation, the dangerous scenarios are rehearsed in the safe space, and the transfer to the plant is the formality the day the readiness is complete: the book’s simulation chapter is the ancestor of the modern virtual commissioning, and the cement plants of today run the identical workflow with the plant simulators: the methodology does not age, the environment does.

10. The Frequently Asked Questions

Is this book only for the mathematicians, or does the plant engineer need it?

The mathematics is required to follow the chapters, but the plant applications are the heart: the identification tests, the simulation and the field results are all described with the industrial data: the process engineer with the basic control-the signal processing and the motivation: the book: the fundamental of the cement mill: the formulas of the chapters are the reference of the mill control projects of the plant: the reader can take the formulas and the procedures without deriving every line.

Does the book apply to the modern vertical roller mills too?

The mewall mechanics of the vertical roller mills differ (the grinding bed, the pneumatic class), but the dynamical structure (the inventory, the recirculation, the classifier, the disturbances) is the same family: the state-space and the identification methods of the book transfer directly: the ball mill monograph the vertical: the modern APC of the VRM implements the same control loops of the book in the new machines: the dynamics of the mill the quantitative?

What really is the result of the closed loop control of a cement mill?

The measured results of the book: the better steady fineness (the lower standard deviation), the higher average through the energy: the reported energy 2-5% and the productivity 3-10%: the results vary with the plant, the instrumentation and the discipline: the monet: the typical: the reduction of the deviations: the product: the quality: the operator of the console: the manual: the chapter 6 of the real difficulties is the most honest to read before the project.

Does the book include the software code or the mechanism the reader can reproduce?

The book includes the detailed mathematical descriptions, the block diagrams, the difference equations and the specification of the experiments: the reader reconstructs the simulation and the estimators from the formulas: the modern reader implements the same in the Python/MATLAB: the current copy of the package is a PDF of the monograph, and the companion Excel and the control files of the package complement it: the reproduction discipline of the chapter 9 supports the independent validation.

Is the “Mat” file of the package a MATLAB m-file?

Here the honest clarification: the catalog entry of the package shows “Mat” because the original title was truncated: the file is not a MATLAB script: it is the PDF of the monograph “Mathematics and Control Engineering of Grinding Technology” by the Hungarian school of Keviczky, Hilger and Kolostori: the packaged file is the mathematical book taken: the extension .pdf: the article of this page describes the content of that PDF: the MATLAB-like calculations the reader does in his own environment.

What makes this monograph unique within the package?

It is the only volume of the package dedicated to the full dynamical mathematics of the grinding: the complete framework from the size distribution to the adaptive control: the other books of the package cover the machinery and the operation (the handbook, the kiln practice): this one covers the mathematics of the closed loop: together they give the engineer the model, the machine and the control: nothing of the loop is missing in the library.

How does the book relate to the artificial intelligence and the machine learning of the modern plants?

Honestly: the current AI projects of the cement industry reuse exactly the framework of the book: the models, the identification, the SoftSensors and the predictive control: the modern option adds the data-driven trainers (the neural estimators of the fineness, the reinforcement policies) on top of the same loop architecture: the engineer with the monograph reads the machine learning papers without the confusion about the fundamentals: the innovation of today is the parameter-estimation chapter of the book accelerated by the big data: the book remains the skeleton, the ML the flesh.

Is the mathematics of the book still the mathematics of the modern DCS?

Yes, practically the same: the modern distributed control systems of the cement plant implement the PID cascades, the feedforward, the gain scheduling and the adaptive loops exactly as the book designs them: the difference is the computing hardware and the data storage, not the control logic: the process controls of the modern plants are the field applications of the mathematics of this monograph, and the engineers of the plant meet the formulas of the book on every tuning screen of the control room.

11. Conclusion

Mathematics and Control Engineering of Grinding Technology: the monograph of the close control: the comminution models, the circuit structure, the identification, the state space and the adaptive control: the field results of the industrial cement and the coal mills: the book is the classic of the grinding control school, and its mathematics remains the basis of the modern mill automation: the reader of the monograph learns to see the mill not as the machinery of the iron but as the dynamical system of the states, the transfer functions and the loops: the vision of the control.

The Complete Cement Technical Package (931 files including the books, the tools, the courses: $249.99 one-time: instant download via the PayPal) includes this monograph PDF with the process control courses, the kiln control materials, the Excel calculators of the mill: the whole left the plant: the mathematical control practiced next to the industrial handbooks: the library of the cement engineer: the complete: the one-time purchase: the instant: the click of the “Buy”: the knowledge of the plant: the control of the mill: the mathematics and the practice at the same package.

This file is part of the Complete Cement Technical Package (931 files) available from cementequipment.org. Respective rights holders; library copy for the licensed single user.