JPH07325606A - Process control unit - Google Patents

Abstract

PURPOSE:To perform process control more strictly by performing on-line, real- time regressive analytic arithmetic. CONSTITUTION:The regressive analytic arithmetic part 7 incorporated in a human interface 6 performs regressive analytic processing on the basis of a measured value of a specific state quantity obtained from a process 1 through a process controller 2 and calculates and supplies such a state quantity set value that a final command becomes an optimum value to the processor controller 2, which performs process control on the basis of the set value as a command, so that while the on-line, real-time regressive analytic arithmetic processing is performed, the process control is performed. Consequently, while set values of various state quantities of temperature, humidity, a raw material amount, a flow rate, a pressure, and others are automatically corrected so that final commands of the process, e.g. production, quality, operation efficiency, etc., become optimum at all times, the process control is performed to perform optimum control over the process.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a process control device, and more particularly to a process control device having an online and real-time regression analysis processing function.

[0002]

2. Description of the Related Art Conventionally, as a process control device for executing automatic, online, and real-time control of a controlled process, a PID control device or an optimum regulator which is a multi-variable target value regulator is used for each feedback loop such as temperature, pressure and flow rate. The one provided for each control device is widely used.

[0003]

However, as described above, the conventional process control device incorporates the PID control device or the optimum regulator, but they have an optimum control target value set in advance. Since the whole process is controlled by controlling each part of the process so as to match it as much as possible, there are the following problems.

That is, the ultimate goal of process control is usually to optimize the production yield, quality, operating efficiency, etc., but these are dependent on the individual characteristics and characteristics of each process, and the length of operating time. They are subtly different and cannot be uniformly determined in advance. Therefore, conventionally, it is not the process control device side that determines the optimum values such as the production amount, the quality, and the operation efficiency.
Accumulate trend data for each process control device, perform off-line regression analysis processing on the upper computer side,
The obtained result is reflected again in the control variable of the process control device.

Therefore, there is a problem that it takes time to set the optimum control conditions, a time delay occurs, and the control is not always executed under the optimum control conditions.

The present invention has been made in view of the above-mentioned problems of the prior art. By executing regression analysis calculation processing online and in real time on the side of the process control device, optimum process control can always be executed. An object of the present invention is to provide a process control device having the above structure.

[0007]

A process controller according to the present invention includes a process controller for controlling various state quantities obtained from a process to be controlled to match a process set value, a process controller and a LAN for a control system.
By using the human interface connected by, and the measured value of the predetermined state quantity obtained from the process by the process controller built in the human interface,
Calculate the set value of the state quantity based on the regression analysis calculation,
It is provided with a regression analysis calculation unit to be given to the process controller.

According to a second aspect of the present invention, in the process control apparatus according to the first aspect, the regression analysis calculation unit executes the regression analysis calculation for each preset cycle.

[0009]

In the process control apparatus according to the first aspect of the present invention, the regression analysis operation unit incorporated in the human interface executes the regression analysis process based on the measured value of the predetermined state quantity obtained from the process through the process controller, Calculate the state quantity set value so that the final target value will be the optimum value and give it to the process controller. The process controller executes the process control with this set value as the target value, and performs regression analysis calculation processing online and in real time. By performing process control while executing, the temperature, humidity, raw material amount, flow rate, pressure, and other various state variables are always adjusted so that the final target value of the process, for example, production amount, quality, operating efficiency, etc., is optimized. The process control can be performed while automatically correcting the set value of, and the optimum process control can be realized.

According to the process control apparatus of the invention of claim 2,
The regression analysis calculation unit executes the regression analysis calculation at regular intervals, so that the final target value of the process can be automatically corrected online and in real time.

[0011]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 1 shows an embodiment of the present invention. The process control apparatus of this embodiment inputs various process values of a control target process 1, for example, temperature, pressure, flow rate, level, weight, etc. Process controller 2, which executes control, control by an optimum regulator, and control for matching various state quantities (process values) of a process with set values, the process controller 2 and the control system L
It is composed of a human interface 6 connected by an AN 3 and having a CRT 4, a keyboard, a mouse, or an input operation unit 5 such as a touch screen.

The human interface 6 is a human interface console, a workstation,
An operator, such as a general-purpose computer, operates the input operation unit 5 while viewing the display on the CRT 4, accesses the process controller 2, and intervenes in the process control.

A program for normal operation is registered in the human interface 6, and a regression analysis calculation program is also registered as a feature of the present invention. In addition, in FIG. 1, the part that executes this regression calculation program is shown as a regression analysis calculation part 7 in order to make it easy to understand.

Next, the operation of the process control device having the above configuration will be described. Normally, the process controller 2
Performs mutual transmission of data to the process 1 and outputs control values such as measured values of various state quantities such as temperature, pressure, flow rate, weight, level, current value, voltage value, power value of each part of the process. Input them, compare them with the set values that are built in themselves, and calculate the control amount necessary to match the measured values with the set values. For example, increase or decrease the fuel supply amount, supply the mixed air for combustion. If it is necessary to increase / decrease the amount, increase / decrease the opening / closing amount of the fuel valve or mixed air valve, and similarly increase / decrease the process value such as pressure, flow rate, weight, or level. Adjust the increase and decrease of blower and motor power, and further increase and decrease the current value, voltage value, power value, etc. by adjusting the volume.
The amount of control is calculated and calculated by the optimum regulator, and this control amount is output to the process 1 to control each part of the process 1 so that the state amount always matches the set value.

However, the set values of various state quantities set in the process controller 2 are uniformly determined, and are caused by the individuality of the control target process 1, the characteristic change over time, the change of the external environment, and the like. There are cases where the setting value of this place is not necessarily the optimum value, for example, it is uniform to set a slightly lower temperature to obtain a good quality product or to slightly change the mixing ratio of raw materials. Process characteristics such as obtaining a quality product or obtaining a product with a predetermined quality even if a little more impurities are mixed will appear.

In such a case, conventionally, the set value is set uniformly, and the change needs to be input by the operator from the human interface. Moreover, the extent to which the change is made is determined by various data offline. Was collected, the regression analysis operation was performed, and the obtained results were observed.

However, in the process control device of the present invention, the regression analysis calculation unit 7 of the human interface 6 executes the regression analysis calculation based on the process data accumulated until then according to the instruction of the operator, and is obtained. The result is displayed on the CRT 4 and is output to the process controller 2 to change the set value. That is, as shown in FIG. 2, of the process data (state quantity) accumulated in the process controller 2 or the process data that the human interface 6 has fetched from the process controller 2 by that time, it is necessary for the regression analysis calculation. Data, that is, data of the selected state quantity (input variable data) and objective variable data (final target value) of the various state quantities are selected and fetched (step S1), and regression analysis calculation is executed (step S2). , The optimum predicted value for the final target value of the objective variable is calculated (step S3).

The optimum predicted value thus obtained is output to the process controller 2, and the process controller 2 controls the process based on this output (step S4).

In this case, if the activation cycle of the regression analysis calculation is set on the human interface 6 side without depending on the operator's command, the regression analysis calculation can be automatically and repeatedly executed for each activation cycle. become.

Next, the regression analysis calculation will be described. For example, in the case of a batch treatment reaction process, the final target value, product quality, depends on various state variables such as the rise time to reaction temperature, reaction time, reaction temperature, temperature, humidity, and raw material amount. Therefore, first, a regression analysis is performed by performing regression analysis based on the quality vs. temperature rise time, reaction time, temperature or humidity, and the like.

Next, a quality target value is given to this regression analysis equation, and the independent variable that causes the above, that is, the above rise time,
Among the reaction time, reaction temperature, temperature, humidity, and raw material amount, variables other than one variable that can be set are given the current environmental value or the current set value, and the value of the remaining one variable is determined.

Next, this set value is given as a target value of the process controller 2, and the operation is performed based on this set value to obtain operation result data.

Next, this operation result data is added to the previous regression analysis data, and the regression analysis is performed again together with the previous data to calculate the regression calculation formula.

In this way, if the quality target value is obtained by adjusting one independent variable, only that independent variable is adjusted thereafter, the other variables are continued as before, and the process control is executed.

However, when the adjustment is not sufficient with one independent variable, that is, when the target quality cannot be reached within a certain range, the value at which the quality target value is optimized for that independent variable. Is set as a set value, and this is given to the process controller 2. Then, another independent variable is selected and the regression analysis operation is repeated in the same manner as above.

In FIG. 3, the target quality value is point A, on the other hand, the regression analysis calculation is performed using temperature data as an independent variable.
The case where the control is executed is illustrated, and it is assumed that the first quality-versus-temperature data is the (A1, a1) point. In this case, α1 is obtained as a regression calculation formula, that is, a regression line. Therefore, since the optimum value prediction point of the temperature corresponding to the target quality value A is P1, this P1 is set in the process controller 2 as the first temperature setting value, and the process is controlled using this P1 as the temperature target value.

As a result, if the obtained quality-versus-temperature data is (A2, P1) point, a regression calculation formula, that is, a regression line α2 is obtained again using this data, and the target quality is obtained on this regression line α2. The optimum value prediction point of the temperature corresponding to the value A is P2.

Next, this temperature P2 is given to the process controller 2 as a set value, and the process is operated again,
Assuming that the quality-versus-temperature data obtained as a result is the point (A3, a3), this point is included in the allowable range B of the target quality A. In this case, the temperature condition a
By controlling the process with 3 as the target value, a product of target quality can be obtained.

Thus, although the target quality was not obtained because the product was operated at the target temperature a1 until then, the product of the target quality is obtained by operating at the target temperature a3 as a result of the quality-temperature regression analysis calculation. Will be obtained. Moreover, in this case, the regression analysis calculation is not performed offline as in the past, but can be executed online and in real time in the human interface 6 and reflected in the process control, so that more strict process control can be realized. .

If the target quality cannot be obtained by the above-described quality vs. temperature regression analysis / control, or if the target quality cannot be obtained by a preset number of times, for example, three times of regression analysis / control. , The temperature that gives the highest quality in the regression analysis calculation result is fixed as the temperature set value, and then the regression analysis / control is continued in the same manner as above for other independent variables, for example, quality vs. humidity.

Further, in the above embodiment, the regression analysis calculation processing of quality vs. independent variable is exemplified, but it is necessary to obtain the optimum set value of the independent variable with respect to other final target values such as production amount and operation efficiency as the characteristics of the process. In some cases, a regression analysis calculation such as (production amount, operation efficiency) vs. (rise time, reaction time, reaction temperature, temperature, humidity, amount of raw material) is performed.

[0032]

As described above, according to the invention of claim 1,
The regression analysis operation unit built into the human interface executes regression analysis processing based on the actual measured values of various state quantities obtained from the process through the process controller, and the various state quantities of the process such that the final target value becomes the optimum value Since the set value is calculated and given to the process controller, and the process controller uses this set value as the state quantity target value to execute the process control, the online,
By performing process control while executing regression analysis calculation processing in real time, the final target value of the process is always
For example, the process control can be performed while automatically correcting the set values of various state quantities so that the production amount, the quality, the operation efficiency, etc. are optimized, and the optimum process control can be realized.

According to the second aspect of the present invention, since the regression analysis calculation unit executes the regression analysis calculation at a constant cycle, in addition to the above effects, the final target value can be automatically corrected online and in real time. The value can be kept optimal.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention.

FIG. 2 is a flowchart of regression analysis calculation processing according to the above embodiment.

FIG. 3 is an explanatory diagram of a regression analysis calculation process of the above embodiment.

[Explanation of symbols]

 1 Process 2 Process Controller 3 LAN for Control System 4 CRT 5 Input Operation Unit 6 Human Interface 7 Regression Analysis Operation Unit

Claims (2)

[Claims] 1. A process controller for controlling various state quantities obtained from a process to be controlled to match process set values, a human interface connected to the process controller via a control system LAN, and a built-in human interface. The process controller uses a measured value of a predetermined state amount obtained from the process, calculates a set value of the state amount based on a regression analysis calculation, and includes a regression analysis calculation unit to give to the process controller. Comprising process control equipment. 2. The process control device according to claim 1, wherein the regression analysis calculation unit executes the regression analysis calculation for each preset cycle.