JPH0566603B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for controlling the internal temperature of a reactor. Specifically, the present invention relates to a method of controlling the internal temperature to an optimum value using a learning function when increasing the temperature of a batch reactor in which the internal temperature is controlled by the jacket temperature.

[Background technology and its problems]

The internal temperature of the polymerization reactor etc. is controlled by cascade type feedback control which is controlled by the jacket temperature.

In such a process, there is a time delay of several tens of minutes after the jacket temperature is changed until the internal temperature responds, and over-control is likely to occur due to feedback control. As a result, the jacket temperature fluctuates widely and the internal temperature of the reactor tends to become unstable.

In addition, in this type of control, operators determined parameters such as the amount of jacket temperature adjustment, temperature change rate, and setting time for temperature switching based on their experience, and operated with manual settings. However, there were large individual differences in temperature control due to differences in operator skill.

[Purpose of the invention]

The purpose of the present invention is to eliminate such conventional drawbacks, suppress large fluctuations in jacket temperature, and eliminate differences due to operator skill.
An object of the present invention is to provide a method for controlling the internal temperature of a reactor, which can accurately control the internal temperature of the reactor to an optimum value.

[Means and actions for solving problems]

In the present invention, the jacket temperature is raised to a preset temperature and then maintained at that temperature, and when the internal temperature of the reactor reaches the exothermic point, it is lowered by a preset temperature adjustment amount, and then the After the internal temperature of the reactor reaches the exothermic point, the internal temperature of the reactor is controlled to a preset internal temperature reference pattern by lowering it at a set temperature change rate after a set time has elapsed, and then gradually increasing it. In the method for controlling the internal temperature of a reactor, the temperature increase region after the internal temperature of the reactor reaches the exothermic point,
divided into a first period from when the internal temperature of the reactor reaches the exothermic point until a set time has elapsed, and a second period from after the elapse of the set time until the internal temperature reaches the saturation point, detecting the internal temperature of the reactor for the first section and the second section;
An evaluation function is obtained by integrating the difference between the internal temperature detection pattern and a preset internal temperature reference pattern, and the temperature adjustment amount is corrected according to the evaluation function of the first section. The above objective is achieved by modifying the temperature change rate according to the evaluation function of the interval.

Therefore, the method of the present invention will be specifically described using FIG. FIG. 1 shows a jacket temperature control curve (jacket temperature control pattern) with respect to a target internal temperature characteristic curve (internal temperature reference pattern) of the reactor.

That is, in order to control the actual internal temperature of the reactor to the internal temperature reference pattern, the jacket temperature control pattern is first raised rapidly, then maintained at the temperature PV _B after reaching the temperature PV _B , and then the internal temperature gradually decreases. A pattern may be used in which the temperature is lowered by the temperature adjustment amount ΔSV at the time t ₀ when the temperature reaches the heating point PV _A , and after the set time t ₁ has elapsed, the temperature is rapidly lowered by the set amount dSV/dt, and then gradually raised. I understand that.

When controlling the internal temperature of the reactor, the temperature range below the exothermic point PV _A (temperature increase period A from the start of temperature increase until the jacket temperature reaches temperature PV _B , and constant temperature zone _{B to maintain temperature PV B} ) ) is not so important, but the temperature rise area above the heating point PV _A (section C from when the internal temperature reaches the heating point PV _A to the set time t ₁ and from t ₁ the internal temperature reaches the saturation point) Control over the period D) up to time _t2 is extremely important from the standpoint of preventing abnormal reactions and the like.

Temperature control in section C is based on the temperature adjustment amount ΔSV, and temperature control in section D is based on the temperature change rate.
According to dSV/dt, temperature control over sections C and D is performed for a set time _t1 . However, in the past, these parameters were manually set by operators based on their experience, resulting in individual differences in temperature control.

The present invention calculates an evaluation function obtained by integrating the difference between the actual internal temperature of the reactor and a reference pattern for each section, and corrects the control parameters for each section according to the evaluation function for each section. This is what I am trying to do.

That is, if the internal temperature detection pattern detected from the reactor is T _R and the internal temperature reference pattern in section C (t ₀ - _{t 1} ) is T _R ', then the evaluation function J ₁ in section C is
can be found as J ₁ =∫ ^t _1t0 (T _R −T _R ′)dt. If the absolute value |J ₁ | of the evaluation function J ₁ determined in this manner exceeds the preset threshold value ε ₁ , the temperature adjustment amount ΔSV is corrected, and the threshold value ε ₁ In the following cases, the temperature adjustment amount ΔSV is kept unchanged.

Now, as shown in Figure 2, the evaluation function J ₁ is +,
If the absolute value |J ₁ | exceeds the threshold value ε ₁ , the temperature adjustment amount ΔSV is too small, so it is corrected to increase it. For example, n
If the evaluation function and temperature adjustment amount at the time of batch processing are J _1(o) , ΔSV _(o) and the constant is β, then the temperature adjustment amount ΔSV _(o+1) at the time of next batch processing is ΔSV _{(o +1)} =ΔSV _(o) +J _1(o)・β is corrected and set. Therefore, in the next batch process, the internal temperature detection pattern in section C decreases and approaches the reference pattern T _R ', so that the internal temperature can be controlled to the target value.

Furthermore, if the internal temperature reference pattern in section D (t ₁ to t ₂ ) is T _R '', then the evaluation function J ₂ in section D is
can be obtained by J ₂ =∫ ^t2 _t1 (T _R −T _R ″)dt. The absolute value |J ₂ | of the evaluation function J ₂ obtained by this is the threshold value set in advance. If ε exceeds ₂ , the temperature change rate
dSV/dt is modified, and if it is less than or equal to the threshold value ε ₂ , the temperature change rate dSV/dt is kept unchanged.

Now, as shown in Figure 3, the evaluation function J ₂ is +,
If the absolute value |J ₂ | exceeds the threshold value ε ₂ , the temperature change rate dSV/dt is too small and is corrected to increase it. For example, if the evaluation function and temperature change rate during the nth batch process are J2 _2(o) , dSV _(o) /dt, and the constant is γ, then the temperature change rate during the next batch process is
dSV _(o+1) /dt is corrected and set to dSV _(o+1) /dt=dsv _(o) /dt+J _2(o) ·γ. Therefore, in the next batch process, the internal temperature detection pattern in section D decreases and approaches the reference pattern T _R '', so that the internal temperature can be controlled to the target value.

Furthermore, the absolute values of the evaluation functions J ₁ and J ₂ |J ₁ |, |J ₂ |
If both exceed the threshold values ε ₁ and ε ₂ , when the number of corrections K of both parameters reaches the set number M, an evaluation function J ₃ covering both sections C and D is determined. The internal temperature reference pattern for both sections C and D (t ₀ to _{t 2} ) is
If T _R , the evaluation function in both intervals C and D is
J ₃ can be determined by J ₃ =∫ ^t2 _t0 (T _R −T _R )dt. If the absolute value |J ₃ | of the evaluation function J ₃ found in this way exceeds the preset threshold ε ₃ , the set time t ₁ is corrected, and the threshold ε ₃ In the following cases, the set time _t1 is kept unchanged.

Now, as shown in Figure 4, the evaluation function J ₃ is +,
If the absolute value |J ₃ | exceeds the threshold value ε ₃ , the set time t ₁ is too large and is corrected to make it smaller. For example, the evaluation function and setting time for the n-th batch processing are J _3(o) ,
t _1(o) and the constant is δ, the set time t _1(o+1) for the next batch processing is t _1(o+1) = t _1(o) + J _3(o)・δ Modified and set. When the set time _t1 is corrected, in the next batch process, the internal temperature detection pattern over sections C and D decreases and approaches the reference pattern. Therefore, from now on, the frequency of correction of the parameters in section C and section D, that is, the temperature adjustment amount ΔSV and the temperature change rate dSV/dt, can be reduced, and the number of corrections can also be kept within a certain range.

〔Example〕

FIG. 5 shows an example in which the method of the present invention is applied to a reactor. In the figure, the internal temperature of the reactor 1 is determined by the jacket 2 attached around it.
It is controlled by the ratio of steam and cold water supplied into the tank.

A circulation path 4 having a circulation pump 3 in the middle is connected to the jacket 2, and steam and cold water are supplied to the circulation path 4 through valves 5 and 6 which are opened and closed by a jacket temperature controller 7. It's summery. The jacket temperature controller 7 adjusts the opening degrees of the valves 5 and 6 so that the value from the jacket temperature detection end 8 that detects the temperature inside the jacket 2 matches a preset jacket temperature control pattern. .

The internal temperature of the reactor 1, for example the polymer temperature within the reactor 1, is detected by the internal temperature detection end 11 and then sent to the calculation section 12, bias control section 20 and switch switch 31, respectively. The calculation unit 12 calculates the difference between the detected value from the internal temperature detection end 11 and the internal temperature reference pattern in the storage unit 13. This difference is given to the pattern control section 40 and, after being instructed to the temperature controller 14, is given to the sequence control section 18. The difference given to the sequence control section 18 is combined with the jacket temperature bias value given from the bias control section 20 in the arithmetic section 15, and then the detected internal temperature value is determined by the signal from the switch 16 (switch switch 31). is turned on at the time t ₀ when the temperature reaches the heat generating point PV _A ), and is combined with the jacket temperature control pattern given from the pattern control section 40 in the arithmetic section 17 and then given to the jacket temperature controller 7. . As a result, the jacket temperature controller 7 uses this set value to set the valves 5 and 6.
Adjust the opening.

In the bias control section 20, first, the rate of change calculation section 21 calculates the detected internal temperature value from the internal temperature detection terminal 11 at a constant sampling period (for example, 1
The internal temperature change rate α (=dPV/dt) at each point in time is determined. To find the rate of change α, for example, as shown in Fig. 6, the internal temperature detection values at each sample time point are PV ₁ and PV ₂ , the values of the internal temperature reference pattern are SV ₁ and SV ₂ , and E ₁ = When PV ₁ −SV ₁ E ₂ =PV ₂ −SV ₂ , the internal temperature change rate α is determined by α=E ₂ −E ₁ /t [° C./min or second].

When the rate of internal temperature change α ₁ to α _o at each sample time determined in this way exceeds a preset threshold value ε, the jacket temperature is bias-controlled by the bias adjustment section 22.

For example, in FIG. 7A, the internal temperature change rate α ₁
.about.α ₃ exceeds the threshold ε, the jacket temperature is bias-controlled by a value obtained by multiplying each internal temperature change rate α ₁ -α ₃ by a constant x, as shown in FIG. 7B.

In addition, as shown in FIG. 8A, among the internal temperature change rates at each sample time (for example, 30 second period),
The internal temperature change rate α _(k) , α _(k+1) , α _(k+2) is the threshold value ε
, the jacket temperature control pattern is bias controlled by a value obtained by multiplying these internal temperature change rates by a constant x, as shown in FIG. 8B. Furthermore, if a cascade output is added to this, the jacket temperature control pattern becomes as shown in FIG. 8C.

On the other hand, in the pattern control section 40, according to the flowchart of FIG _. In addition, the set number M of corrections is set as an initial value, and then the evaluation function
Find J _1(o) and J _2(o) . The evaluation function found here
J _1(o) and J _2(o) are determined by the evaluation unit 42 to determine whether their absolute values |J _1(o) |, |J _2(o) | are larger than the threshold values ε ₁ , ε ₂ or not. is judged.

If the absolute value of the evaluation function J _1(o) |J _1(o) | is larger than the threshold ε ₁ , the temperature adjustment amount for the next batch processing is ΔSV _(o+1) = ΔSV _(o) +J _{1 (o)}・β, while if the absolute value |J _1(o) | is less than the threshold ε ₁ , the temperature adjustment amount for the next batch processing remains unchanged, that is, ΔSV _(o+1) = ΔSV _(o) .

Furthermore, if the absolute value |J _2(o) | of the evaluation function J _2(o ) is larger than the threshold ε ₂ , the temperature change rate during the next batch processing is dSV _(o+1) /dt=dSV _{( o)} /dt＋J _2(o)・γ
while the absolute value |J _2(o) | is the threshold value ε ₂
In the following cases, the temperature change rate during the next batch process is maintained as is, that is, dSV _(o+1) /dt = dSV _(o) /dt.

Furthermore, if the absolute value |J _1(o) | is less than or equal to the threshold value ε ₁ , and the absolute value |J _2(o) | is less than or equal to the threshold value ε ₁ , the batch processing number n is set to 1. After adding, return to calculating the evaluation function.

On the other hand, if the absolute value |J _1(o) | is larger than the threshold value ε ₁ and the absolute value |J _2(o) | is larger than the threshold value ε ₂ , then the temperature adjustment amount ΔSV and It is determined whether the number of corrections K for correcting the temperature change rate dSV/dt has reached a preset number M.

If the number of modifications K has not reached the set number M, 1 is added to the number of modifications K, then 1 is added to the batch processing number n, and the process returns to calculating the evaluation function.

When the number of modifications K reaches the set number M, the number of modifications K is reset to 0, and then the evaluation function J3 _(o) is determined. The evaluation function J _3(o) obtained here is judged by the evaluation unit 42 as to whether or not its absolute value |J _3(o) | is larger than the threshold value ε ₃ .

If the absolute value |J _3(o) | is larger than the threshold ε ₃ , the set time for the next batch processing is t _1(o+1) = t _1(o)
+J _3(o)・δ. On the other hand, if the absolute value |J _3(o) | is less than the threshold ε ₃ , the set time for the next batch process remains unchanged, that is, t _{1(o+ 1)} is maintained at =t _1(o) . After that, 1 is added to the batch processing number n, and then the process returns to calculating the evaluation function.

Therefore, according to this embodiment, an evaluation function is obtained by integrating the difference between the internal temperature detection pattern of the reactor 1 and the internal temperature reference pattern, and the jacket temperature control parameter is corrected according to this evaluation function. Because of this, the internal temperature of the reactor can be easily and accurately controlled to the optimum value without relying on the operator's experience as in the past.As a result, abnormal reactions in the reactor can be prevented and the temperature of the reactor can be controlled. Labor savings can be achieved by automating control.

In addition, the temperature range in which the internal temperature of the reactor increases is divided, and an evaluation function is obtained for each section. For example, the absolute value of the evaluation function J ₁ in section C in Fig. 1 exceeds the threshold ε ₁ . When, the temperature adjustment amount ΔSV is corrected,
Since the temperature change rate dSV/dt is corrected when the evaluation function J ₂ in the D interval exceeds the threshold value ε ₂ , the temperature increase curve of the internal temperature can be controlled finely and accurately.

Furthermore, if the absolute values of both evaluation functions J ₁ and J ₂ exceed their respective thresholds ε ₁ and ε ₂ and the number of corrections K has reached the set number M, the evaluation function J ₃ is calculated, Set time t ₁ according to the absolute value of this evaluation function J ₃
Even if the temperature adjustment amount ΔSV and temperature change speed dSV/dt are frequently corrected, the set time _t1 is corrected each time the set number of times M is reached, so frequent corrections will be avoided. There is no need to perform any processing, and the amount of these corrections can be kept within a certain range.

〔Effect of the invention〕

As described above, according to the present invention, the internal temperature of the reactor can be controlled while suppressing large fluctuations in jacket temperature.
It can be accurately controlled to the optimum value without any individual differences between operators.

[Brief explanation of the drawing]

Figures 1 to 4 are diagrams for explaining the present invention. Figure 1 is a diagram showing the relationship between the internal temperature reference pattern and the jacket temperature control pattern, and Figures 2 to 4 are diagrams showing the evaluation in each section. FIG. 3 is a diagram showing the relationship between functions and parameters. Figures 5 to 9 show an embodiment of the present invention. Figure 5 is a block diagram showing the entire system, Figure 6 is an explanatory diagram for determining the rate of temperature change, and Figure 7 is bias control. FIG. 8 is an explanatory diagram showing another example of bias control, and FIG. 9 is a flowchart showing pattern control of jacket temperature. 1... Reactor, 2... Jacket, 40... Pattern control section.

Claims (1)

[Claims] 1. The jacket temperature is raised to a preset temperature, maintained at that temperature, and then lowered by a preset temperature adjustment amount when the internal temperature of the reactor reaches the exothermic point, and then After the internal temperature of the reactor reaches the exothermic point, the internal temperature of the reactor is lowered at a set temperature change rate after a set time has elapsed, and then gradually raised, thereby bringing the internal temperature of the reactor to a preset internal temperature standard. In a method for controlling the internal temperature of a reactor in a pattern, the temperature increase region after the internal temperature of the reactor reaches the exothermic point is controlled until a set time elapses after the internal temperature of the reactor reaches the exothermic point. The internal temperature of the reactor is divided into a first period and a second period from after the set time has elapsed until the internal temperature reaches the saturation point, and for the first period and the second period, the internal temperature of the reactor is detecting the internal temperature, and integrating the difference between the internal temperature detection pattern and a preset internal temperature reference pattern to obtain an evaluation function, and correcting the temperature adjustment amount according to the evaluation function of the first section, A method for controlling an internal temperature of a reactor, comprising: modifying the temperature change rate according to the evaluation function of the second section. 2. In the method for controlling the internal temperature of a reactor according to claim 1, when the number of corrections of the temperature adjustment amount and temperature change rate reaches a preset number of times, the internal temperature reaches the exothermic point. A method for controlling an internal temperature of a reactor, characterized in that an evaluation function is obtained for a third interval from the time the temperature reaches the saturation point until the saturation point is reached, and the set time is corrected according to the evaluation function. 3. In the method for controlling the internal temperature of a reactor according to claim 2, the temperature adjustment amount, temperature change rate, and set time are corrected when each evaluation function exceeds a preset threshold. A method for controlling the internal temperature of a reactor, characterized in that: 4. In the method for controlling the internal temperature of a reactor according to any one of claims 1 to 3,
A method for controlling an internal temperature of a reactor, characterized in that, in the correction, a product of the evaluation function and a constant is added to the previous value.