JPH06327990A - Control method of roller mill - Google Patents

Abstract

PURPOSE:To improve a control method of roller mills to rapidly and accurately control a roller mill by obtaining a model from a linear equation in which product particle data, as an example of operating state data is represented by the gas flow volume and separator revolution as examples of operating set values as variables. CONSTITUTION:An optimum operation control device for roller mills 1 is provided with a memory 2 for storing a mill property model representing the previously set relationship between operating state data showing the operating state of a roller mill 10 and an operating set value of a control factor of the roller mill 10 corresponding to the data. A calculating part 3 for calculating a new operating set value using the mill property model stored in the memory 2 so that an evaluation function for evaluating the operating state data and/or operating set value may be optimum under conditions which meet the previously set constraint conditions is provided. An optimum operating set value output part 4 for outputting the optimum operating set value calculated by the calculating part 3 to a control part contained in the roller mill 10 is provided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to a cement raw material,
The present invention relates to a roller mill control method for suitably controlling a vertical roller mill that grinds cement clinker, granulated slag, and the like.

[0002]

2. Description of the Related Art The above-described roller mill has a great merit that it has a good crushing effect (that is, the basic unit showing the crushing amount per mill power is low) as compared with a ball mill known as a production facility. ing. However, the control of the roller mill is very sensitive to disturbances such as the pulverization characteristics of the raw material, so that the roller mill is a system that is difficult to operate. Therefore, at present, the roller mill is operated in a state of relying on the qualitative know-how of the skilled operator as shown in FIG. 4, but it is difficult to perform suitable control, and there is a demand for automation. Therefore, the present inventors have filed Japanese Patent Application No. 4-20131.
No. 2 roller mill control method was developed. In this method, basically, the correlation between the operation state data of the roller mill and the operation set value of the control factor is modeled in advance, and a certain operation set value is applied to the model to obtain the above operation set value. Was obtained, and the operation set value was adjusted so that the operation state data had a desired value.

[0003]

In the conventional roller mill control method as described above, a mill characteristic model is used, and the variables of this model are calculated by using various methods. However, as the variables of the roller mill, there are many operation setting values such as pressurizing force, crushing amount, separator rotation speed, air volume, etc., and it is necessary to comprehensively correct them. The calculation for is likely to be large scale. Further, the coefficient of the mill characteristic model obtained in this way is expected to change due to factors such as the grindability of the raw material, which are difficult to measure in advance, and there is room for improvement in terms of accuracy. SUMMARY OF THE INVENTION It is an object of the present invention to improve a roller mill control method and to provide a roller mill control method capable of performing quick and accurate control of a roller mill in order to solve the problems in the conventional art.

[0004]

In order to achieve the above object, the present invention is a roller mill control method for controlling the operating state of the roller mill by adjusting the operation set value of the control factor of the roller mill. The correlation between the operation state data indicating the state and the operation setting value is modeled in advance, and a certain operation setting value is applied to the model to obtain certain operation state data corresponding to the operation setting value. In a roller mill control method for adjusting a certain operation set value such that a certain operation state data has a desired value, the model uses product particle data, which is an example of the operation state data, and air volume which is an example of the operation set value. And a linear expression expressing the separator rotation speed as a variable, which is configured as a roller mill control method. Furthermore, there is provided a roller mill control method for controlling the operation state of the roller mill by adjusting the operation set value of the roller mill control factor, wherein the correlation between the operation state data indicating the operation state and the operation set value is previously set. A certain operation setting data corresponding to the certain operation setting value is obtained by applying a certain operation setting value to the model in advance, and the certain operation setting value is set so that the certain operation state data becomes a desired value. In the roller mill control method for adjusting, the model has a linear expression in which the product particle data, which is an example of the operation state data, is represented by variables of the air flow rate and the separator rotation speed, which are an example of the operation set values, and
A roller mill control method including a model by multiple regression analysis representing operating state data other than the product particle data.

[0005]

According to the present invention, the correlation between the operation state data indicating the operation state of the roller mill and the operation set value of the control factor of the roller mill is modeled in advance, and a certain operation set value is applied to the model. In order to control the operating state of the roller mill by obtaining certain operating state data corresponding to the certain operating set value by adjusting the certain operating set value so that the certain operating state data becomes a desired value, The model includes a linear expression in which product particle data, which is an example of the operation state data, is represented by variables of the air flow rate and the separator rotation number, which are an example of the operation set values. The above linear equation is obtained by linearly approximating the classification point calculation formula, which is well known as a model formula for expressing product particle data, and this classification point calculation formula is analyzed to have a reliability sufficient for practical use. It has been confirmed that there is. Therefore, reliability in controlling the operating state of the roller mill can be ensured. Further, the model includes a model by multiple regression analysis that represents operating state data other than the product particle data. By reducing the symmetry of the multiple regression analysis that requires repeated calculation in this way, it is possible to improve the calculation speed and reduce the calculation error.
As a result, it is possible to obtain a roller mill control method capable of controlling the roller mill quickly and accurately.

[0006]

The embodiments of the present invention will be described below with reference to the accompanying drawings to provide an understanding of the present invention. The following embodiments are examples of embodying the present invention and are not intended to limit the technical scope of the present invention. here,
FIG. 1 is a block diagram showing a schematic configuration around a roller mill optimum operation control device 1 according to an embodiment of the present invention, FIG. 2 is a schematic configuration diagram of a general roller mill, and FIG. 3 is a fractional point calculated value and an actually measured value. It is a graph which shows a comparison result. In the roller mill 10 which is an example of a general roller mill, as shown in FIG. 2, the raw material charged on the rotating table 12 from the raw material charging port 14 on the side of the mill casing 11 cooperates with this table 12. The rotating roller 13 and the table 12
Crushed and crushed between and. Then, the particles that have been made finer than a predetermined particle size from the above raw material are blown upward by the hot air blown from the hot air generating furnace 16 into the mill casing 11 through the intake port 21. The blown-up particles are classified into coarse particles and fine particles by the action of the separator 19 which is rotated by the drive of the motor 20, and the fine particles are hot air and the mill outlet 1 at the upper part of the mill casing 11.
5 is guided to the bag filter 22 through the feed duct 17 and collected as a product. On the other hand, the separator 19
The coarse particles classified by are dropped in the mill casing 11, returned to the table 12 again, and pulverized.
Then, the particles that have not been blown up from the vicinity of the table 12 by the hot air fall to the bottom of the mill casing 11, and are returned from the bottom to the raw material side by a basket elevator or the like (not shown). In addition, the product particle size distribution is obtained for the fine-grained product collected by the bag filter 22. When the particle size of the product is too fine in view of this product particle size distribution, the motor 20 is driven so that the rotation speed of the separator 19 is decreased, and conversely, when the product particle size is too coarse, the rotation speed of the separator 19 is increased. Controlled. The table 12 and the motor 1 that drives the table 12 to rotate.
If the value of the mill vibration detected by the vibration detection sensor (not shown) attached to the speed reducer (not shown) between 8 and 8 becomes excessive, it will cause mechanical damage to each part of the roller mill 10. When the value of the mill vibration exceeds 6 mm / sec, the operation of the roller mill 10 is urgently stopped by the action of a trip circuit (not shown).

The present invention is a control method applicable to such a roller mill 10 and is configured as follows. In this method, the correlation between the operation state data indicating the operation state of the roller mill 10 and the operation set value of the control factor of the roller mill 10 is modeled in advance, and a certain operation set value is applied to the model. The configuration is the same as that of the conventional example in that certain operation state data corresponding to the certain operation set value is obtained and the certain operation set value is adjusted so that the certain operation state data has a desired value. However, in the present method, the model includes a linear expression in which the product particle data, which is an example of the operation state data, is represented by variables of the air flow rate and the separator rotation speed, which are an example of the operation set values. In addition, a model by multiple regression analysis showing operation data other than the product particle data is also included. Furthermore, the configuration is different from the conventional example in that each variable of the above model is made dimensionless. This method is a roller mill optimum operation control device 1 for outputting an optimum set value to a control device portion (hereinafter abbreviated as a control unit) according to the conventional method included in the roller mill 10.
Is embodied by. That is, as shown in FIG. 1, the apparatus 1 has operating state data indicating the operating state of the roller mill 10 (mill differential pressure, product particle size distribution, mill power, mill vibration, etc.).
And the operation set values of the control factors of the roller mill 10 corresponding to these (roller pressing force, raw material input amount, separator rotation speed,
(A hot air flow rate), a memory 2 that stores a mill characteristic model that represents a preset relationship with the blowing hot air volume, and a mill characteristic model that is stored in the memory 2 The calculation unit 3 that calculates a new operation set value so as to optimize the state data and / or the evaluation function for evaluating the operation set value, and the optimum operation set value calculated by the calculation unit 3 are given to the roller mill 10. And an optimum operation set value output unit 4 for outputting to the control unit included in the. However, the mill characteristic model is a linear expression showing the relationship between the product particle size distribution (90% pass particle size), which is the product particle size data, and the separator rotation speed and the blowing hot air volume, which are the operation setting values corresponding to this data. The model represented,
It includes operation state data other than product grain size data and a model by multiple regression analysis showing the relationship between operation setting values corresponding to this data, and each of the above elements 2 to 4 is a mill including these multiple models. It can handle characteristic models.

The models included in the mill characteristic model will be derived below. The following classification point calculation formula is well known as a formula representing the relationship between the operation setting amount (separator rotation speed, air volume) of the roller mill 10 and the product particle data (product particle size distribution). According to this formula, the classified particle diameter (90% pass particle diameter) can be obtained from the air flow rate and the separator rotation speed under other constant mechanical conditions. d = k × Q / H × R ^−5/3 × N ₀ ^−4/3 (1) where d: Classified particle size (μm) Q: Actual air flow into separator (m ³ / sec) H: Rotor Height (m) R: Rotor radius (m) N ₀ : Rotor speed (rpm) (= separator speed N)
× 9.07) K: constant In order to use the above equation (1) in the computing unit 3, the equation is transformed into a linear equation having the amount of change (air volume Q and separator rotation speed N) as variables. The constant terms are collectively set to k'and the natural logarithm of both sides is taken. At this time, the constant k'also includes the conversion factor between the rotor speed N ₀ and the separator speed N. lnd = ln (k ′ · Q · N ^−4/3 ) (2) Since the above equation (2) should be established even if a change occurs in each variable, the following equation is also established. ln (d + Δd) = ln (k ′ · (Q + ΔQ) · (N + ΔN) ^−4/3 ) (3) Taking the difference between both sides of the above equations (1) and (2), the logarithmic formula gives the following equation. Can be transformed into ln (1 + Δd / d) = ln (1 + ΔQ / Q)-(4/3) ln (1 + ΔN / N) (4) The above equation (4) is expanded into a series and only the first-order term is considered, assuming that the variables are sufficiently small. By adopting, the following linear equation can be obtained. Δd / d = ΔQ / Q- (4/3) (ΔN / N) (5) The classified particle diameter d obtained from the air volume Q and the separator rotation speed N by using the above formula (5) and the actual collection were performed. The obtained 90% pass particle diameter was compared with the value of the Cirrus analysis of the product.

Specifically, as a result of the unit regression analysis, a good correlation coefficient of 0.796670 was obtained (see FIG. 3). Next, using the relational expression between the classified particle size d obtained by the above-mentioned single regression analysis by the above formula (5) and the classified particle size obtained by the Cirrus analysis which is a well-known analysis method, the above (5) Consider making a formula correction. In order to generalize, the simple regression equation is as follows. d = ad ′ + b (6) where d: actual measured value of classification point (by Cirrus analysis) d ′: calculated value of classification point (by above formula (5)) a, b: coefficient What is substituted into the equation (5) is finally used as the above model in the mill characteristic model. Δd / (d−b) = ΔQ / Q− (4/3) (ΔN / N) (7) The variables in equation (7) above should be dimensionless by dividing the amount of change by the current value. Therefore, 90% when using this formula
The current value of pass particle size is required. However, as the product particle size distribution, the well-known Rosin-Rammler distribution rule is used from the two points of 32 μpass% and 8μpass%, which have the highest correlation with the AJS 30μ residue value and the Blaine value, not in the form of 90% pass particle diameter. It is obtained by extrapolation. Subsequently, the models included in the mill characteristic model are derived. As for the model, operation state data y _i other than the above product particle size distribution
The partial regression coefficient can be obtained by expressing the relationship between the value and the corresponding operation setting value x _j as a function and performing multiple regression analysis. When performing this multiple regression analysis, the variables (y _i ,
x _j ) is made dimensionless as follows in combination with the above model. Δ′x _j = Δx _j / x _j = (x _{j + 1} −x _j ) / x _j (8) Δ′y _i = Δy _i / y _i = (y _{i + 1} −y _i ) / y _i (9) The model is expressed as follows using the dimensionless variables Δx _j and Δy _i . Δ'y _i = a _ij Δ'x _j (10) (a _ij : partial regression coefficient) The roller mill 10 was controlled by the device 1 using the mill characteristic model including the model thus derived. The operation example in this case is shown below.

Here, the following specifications are used as the operation set values of the control factor. x _1: raw material input (t / h) x _2: tensioning hydraulic (roller rolling ^{force) (kgf / cm 2) x} 3: Separator rpm (rpm) x _4: blow hot air volume (m ³ / min) The As the operation state data of the roller mill 10, the following specifications are used. y ₁ : 90% pass particle size (μ) y ₂ : Power (kW) y ₃ : Vibration (mm / s) Of the mill characteristic models, the model is the above operation set value x
_{1 to} x ₄ and the operation state data y _{1 to} y ₃ are described in (7) to
It is described as follows by substituting it into the equation (10). Model: Δ′y ₁ = Δ′x ₄ − (4/3) Δ′x ₃ (11)

[Equation 1] The models represented by the above equations (11) and (12) are preset and stored in the memory 2. The arithmetic unit 3 mainly comprises a central processing unit (CPU) (not shown), and stores arithmetic algorithms such as the simplex method in a ROM (not shown). Then, the calculation is performed using the mill characteristic model (model) stored in the memory 2. The optimum operation set value, which is the result of this calculation, is output to the control unit included in the roller mill 10, whereby the operating state of the roller mill 10 is controlled. As described above, the mill characteristic model of the roller mill 10 of this embodiment is as follows. Product particle data (product particle size distribution), which is an example of operating state data,
A model described by a linear expression in which the air flow rate and the separator rotation speed, which are examples of operation set values, are represented as variables is included. This linear equation is obtained by approximating the classification point calculation formula, which is a well-known model for expressing the product particle size distribution, and it has been confirmed by the above analysis that it is sufficiently reliable for practical use. Therefore, it is possible to secure reliability in controlling the operating state of the roller mill.

Further, the mill characteristic model includes a model for multiple regression analysis representing operating state data other than the particle data. In this way, by reducing the number of objects of the multiple regression analysis that requires repeated calculation, it is possible to improve the calculation speed and reduce the calculation error. Furthermore,
By making each variable of the above model, dimensionless, it becomes relatively easy to handle each variable in operation. As a result, it is possible to obtain a roller mill control method capable of controlling the roller mill quickly and accurately. Although the mill characteristic model is represented by the model in the above embodiment, it is possible to further add the model in actual use. In the above embodiment, the control method of the roller mill 10 which is relatively easy to operate has been described, but there is no problem even if this method is applied to other types of mills.

[0012]

Since the roller mill control method according to the present invention is configured as described above, part of the mill characteristic model is product particle data, which is an example of operating state data, and air volume is an example, which is an operation set value. And the separator rotation speed can be described by a linear expression that is expressed as a variable. This linear equation is obtained by linearly approximating a classification point calculation formula, which is a well-known model for expressing product particle data, and it has been confirmed by the above analysis that it is sufficiently reliable for practical use. Therefore, reliable control can be secured in controlling the operating state of the roller mill. Further, a part of the mill characteristic model may be a model by multiple regression analysis that represents operating state data other than the particle data. In this way, by reducing the number of multiple regression analysis targets that require repeated calculation, it is possible to improve the calculation speed and reduce the number of calculation points. As a result, it is possible to obtain a roller mill control method capable of performing quick and accurate control of the roller mill.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration around a roller mill optimum operation control device 1 according to an embodiment of the present invention.

FIG. 2 is a schematic configuration diagram of a general roller mill.

FIG. 3 is a graph showing the results of comparison between calculated values of classification points and measured values.

FIG. 4 is a state explanatory view showing a control state of a conventional roller mill by a skilled operator.

[Explanation of symbols]

1 ... roller mill optimum operation control device 2 ... memory 3 ... arithmetic unit 4 ... optimum operation setting value output unit 10 ... roller mill x ₁ ~x ₄ ... operation setting value y ₁ ~y ₃ ... operational state data

Claims (2)

[Claims] 1. A roller mill control method for controlling an operating state of the roller mill by adjusting an operating set value of a control factor of the roller mill, wherein a correlation between the operating state data indicating the operating state and the operating set value. Is modeled in advance, certain operation setting data corresponding to the certain operation setting value is obtained by applying the certain operation setting value to the model, and the certain operation is performed so that the certain operation state data has a desired value. In a roller mill control method for adjusting a set value, the model includes a linear expression in which product particle data, which is an example of the operation state data, is represented by variables of an air flow rate and a separator rotation number, which are an example of the operation set values. Characteristic roller mill control method. 2. A roller mill control method for controlling an operating state of the roller mill by adjusting an operating set value of a control factor of the roller mill, wherein a correlation between the operating state data indicating the operating state and the operating set value. Is modeled in advance, certain operation setting data corresponding to the certain operation setting value is obtained by applying the certain operation setting value to the model, and the certain operation is performed so that the certain operation state data has a desired value. In the roller mill control method of adjusting the set value, the model is a linear expression that represents product particle data that is an example of the operating state data as variables of the air flow rate and the separator rotation speed that are an example of the operation set values, and
A roller mill control method including a model by multiple regression analysis that represents operating state data other than the product particle data.