JPS62248917A - Device for measuring response error for boiler - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a response error measuring device in a boiler that measures dead time of a controlled object.

[Prior Art] For example, even if the fuel flow rate is increased, the boiler steam temperature does not rise immediately, and there is a response delay (hereinafter referred to as dead time). However, if the fuel flow rate is increased needlessly even though the steam temperature does not rise within this dead time, the steam temperature will rise too much after the dead time has passed, making it difficult to control the steam temperature to a predetermined temperature. As described above, in a control system, the dead time of a controlled object is an important element related to control stability.

Conventionally, for example, the dead time in boiler steam temperature control is
This was determined by inputting the fuel flow rate as a step signal and conducting a dynamic characteristic test to measure the time it takes for the steam temperature of the boiler to rise.

[Problems to be Solved by the Invention] However, the dead time measuring means described above is a method in which all operating terminals are set to manual mode and the control I device is disconnected), and a step input is applied to only one operating terminal. Therefore, it was impossible to measure when the dead time changes continuously or when using an actual control system.

In view of the above-mentioned circumstances, the present invention has been made to enable measurement of dead time in real time (time when the device is actually operating), to enable tracking of continuous changes in dead time, and to a control device. This was done with the purpose of making it possible to measure dead time when using a

[Means for Solving the Problems] The present invention provides a random noise generator that outputs random noise, a signal from the random noise generator, and a signal provided from the random noise generator to a controlled device. A parameter calculation device that calculates coefficients of an autoregressive model based on a measurement target signal output from the controlled device; and a parameter comparison device that calculates a magnitude relationship between the coefficients from the calculated coefficients and derives a dead time from the result. I appreciate the structure that provides this.

[Function] The noise signal from the random noise generator is applied to the controlled device and the parameter calculation device, and the controlled device outputs a signal to be controlled according to the noise signal and is given to the parameter calculation device, and the parameter calculation device outputs a signal to be controlled according to the noise signal. The calculation device calculates the coefficients of the autoregressive model based on the signal to be controlled and the noise signal, the coefficients of the autoregressive model are compared in magnitude by the parameter comparison device, and the dead time is calculated from the results.

[Example] Hereinafter, an example of the present invention will be described with reference to the accompanying drawings.

FIG. 1 shows an embodiment of the present invention, in which numeral 1 represents a controlled device having an object to be measured, such as boiler steam temperature, and numeral 2 represents a dead time measuring device. The dead time measuring device 2 is composed of a random noise generating device 3 that generates a disturbance, a parameter calculating device 4 for an autoregressive model, a parameter comparing device 5, and a dead time display device 6. A minute input signal u(kl
(see Figure 2) in the controlled device 1 and the parameter calculation device 4.
The parameter calculation device 4 calculates an autoregressive model based on the minute input signal u (k) and the output signal y (k) from the measurement object 1 (see Figure 3). The parameter comparison device 5 is used to calculate the dead time by comparing the magnitude relationship of the parameters from each coefficient of the autoregressive model obtained by the parameter calculation device 4. The time display device 6 is for displaying the determined dead time.

During actual operation of the device, the random noise generator 3 sends the controlled device 1 (for example, if the measurement target of the controlled device 1 is the boiler steam temperature, a control valve that controls the flow rate of fuel supplied to the boiler burner) and the parameter calculation device. 4 is given a minute disturbance as an input signal u (k), and the controlled device 1 outputs, for example, boiler steam temperature as an output signal @ y (k
) to the parameter calculation device 4. here,
The output signal corresponding to the input signal u (k) at the time “!
The measurement target of /(k) (one person's output per person) can be expressed as follows using an autoregressive model.Here, m
; order of autoregressive model; discrete time e(k); noise or error al, bl: If the sampling period of each m coefficients or data is At, the elapsed time is expressed as t = R - At. . Therefore, the parameter calculation device 4 converts the coefficients ai and bl of equation (1), and calculates the coefficients at e bi using the recursive least squares method expressed as such that the noise or error e (k) is minimized. It is determined.

Here, J is an evaluation function. However, the order m of the autoregressive model is set in the parameter calculation device 4 in advance to a value sufficiently larger than the dead time (m>>LZAt), which is large enough for the number of measurers.

Once the coefficients ai and bi of the autoregressive model are calculated,
Each coefficient ai*bl is added to a parameter comparison device 5, which compares the magnitude of each coefficient, and calculates the dead time from this magnitude relationship. That is, if there is a dead time, the coefficients of bl, b2 . ... is small, and bl becomes sufficiently larger than bl from the time when the dead time is exceeded.

Therefore, l b + l kuku I b i l
...Find i for which σ holds, and calculate the dead time from L = (i-1) * A t - (v). The determined dead time is displayed on the dead time display section 6.

However, the controlled device 1 can measure the object to be measured even if the control device is not connected as shown in FIG. 4 or even if the control device 7 is connected as shown in FIG. The reason is that the feedback loop by the control device 7 does not affect the dead time.

In the embodiments of the present invention, the case where the controlled device is a boiler has been described, but it is also applicable to a denitrification device, the measurement target can be measured not only temperature but also pressure, and other aspects that deviate from the gist of the present invention. Of course, various changes may be made within the scope.

[Effect of the invention 1] According to the response error measuring device in a boiler of the present invention, (D) Since minute random noise is used as an input signal to the controlled device and the parameter calculation device, the dead time measurement target is It is possible to measure dead time even under controlled conditions. (I [> If the dead time changes continuously, the change can be tracked in real time.) Since the method of least squares can be used, it is possible to achieve various excellent effects such as being able to measure dead time while actually driving.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram of an embodiment of the response error measuring device in a boiler according to the present invention, and FIG. 2 is an explanatory diagram of an example of the response error measuring device in a boiler according to the present invention, and FIG. FIG. 3 is a graph showing the relationship between the output signal applied from the controlled device to the parameter calculation device in the device shown in FIG. 1 and time, and FIGS. FIG. 2 is an explanatory diagram of a configuration example of a controlled device to which a response error measuring device is applied.

Claims (1)

[Claims] 1) A random noise generator that outputs random noise, and a measurement target output from the controlled device in response to a signal from the random noise generator and a signal given to the controlled device from the random noise generator. A boiler characterized by being provided with a parameter calculation device that calculates coefficients of an autoregressive model based on a signal, and a parameter comparison device that calculates a magnitude relationship between the coefficients from the calculated coefficients and derives dead time from the result. Response error measuring device.