JP3519539B2 - Flow control device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for adjusting a flow rate by adjusting a valve opening degree in a throttle body or the like, and more specifically, a case where a fuzzy control method is applied to manually adjust the flow rate. The present invention relates to a flow rate adjusting device capable of adjusting the flow rate with equal accuracy and required processing time.

[0002]

2. Description of the Related Art A throttle body used in an automobile engine or the like supplies air at an appropriate flow rate according to the driving situation. However, since there is individual difference in the relationship between valve opening and flow rate, it is manufactured. At times, the valves have to be adjusted so that a predetermined flow rate flows at a certain reference condition for each one. This is called the adjustment of the main flow rate.

Conventionally, the main flow rate is adjusted as shown in FIG. That is, the device includes, as a detection system, an exhaust pump 74 that causes an air flow to the throttle bore 71 of the throttle body 70, and the flow path 7 thereof.
5 and a negative pressure sensor 76 attached thereto. And as an operation system, a valve 72 of the throttle body 70
Lever 78 provided coaxially with, an adjusting screw (hereinafter referred to as “TAS”) 79 for pushing the lever 78, and a TAS 79
And a motor 80 for driving. The motor 80 is
It is a step motor. A pulse generation circuit 82 is provided as a control system.

The pulse generation circuit 82 includes a negative pressure sensor 76.
There is provided a V / F converter 83 that generates pulses at a rate proportional to the difference between the output value (voltage signal) and the target value, and a sequencer 84 that generates pulses at a constant rate. The pulse rate of the sequencer 84 is a rate at which the motor 80 rotates at an extremely low speed that can be stopped quickly. Then, one of the pulse generated by the V / F converter 83 and the pulse generated by the sequencer 84 is selected by the selector 85 according to the output value of the negative pressure sensor 76 and applied to the motor 80. There is.

Then, the valve 7 of the throttle body 70
2 is fully closed and the exhaust pump 74 is driven to drive the air flow F.
The negative pressure sensor 76 is about -670 mmHg.
Indicates a negative pressure of about (gauge pressure, the same applies hereinafter). From this state, the motor 80 is driven to gradually open the valve 72, and the measured value of the negative pressure sensor 76 is set to the target value (for example, -450 mmH).
When g) is reached, the valve 72 is stopped.

This procedure will be described with reference to the graph of FIG. In this graph, the horizontal axis indicates the elapsed time from the start of driving the motor 80, the upper side indicates the pulse rate for driving the motor 80, and the lower side indicates the reading value of the negative pressure sensor 76. At the start of driving the motor 80, the measured value of the negative pressure sensor 76 shows about -670 mmHg, and is constant until timing A about 3 seconds later.
This is because the tip of TAS79 is still lever 7 in this section.
This is because the valve 72 does not move even when the motor 80 is driven because the valve 72 is not in contact with 8. TAS7 at timing A
When the tip of 9 reaches the lever 78, the motor 80
The movement of the TAS 79 due to the drive of is transmitted to the valve 72 via the lever 78, and the valve 72 is gradually opened. Therefore, the measurement value of the negative pressure sensor 76 gradually increases as the opening degree of the valve 72 increases. Although not shown in FIG. 5, the throttle body 70 is provided with a return spring that biases the valve 72 in the closing direction, so that the valve 72 opens while resisting the bias.

Focusing on the pulse rate here, it is 6 at a timing B about 4 seconds after the start of driving the motor 80.
It is constant at 0 pulse per second. Since the measured value of the negative pressure sensor 76 is low (the negative pressure is high) in this section, the selector 85 in the pulse generating circuit 82 applies the pulse generated by the V / F converter 83 to the motor 80, and the V / F
The converter 83 generates pulses at a rate proportional to the difference between the output value and the target value as described above, but the V / F converter 83 generates because the difference between the output value and the target value is still large. This is because pulses are generated at the maximum possible rate.

After the timing B, the V / F converter 83 generates a pulse having a rate proportional to the difference between the output value and the target value, so that the pulse rate decreases as the output value increases, that is, the rotation of the motor 80. The speed gradually slows down. The measured value of the negative pressure sensor 76 is -5.
When it reaches 00 mmHg (G) (timing C), the selector 85 determines the source of the pulse applied to the motor 80 as
The V / F converter 83 is switched to the sequencer 84. Therefore, thereafter, the motor 80 is driven at an extremely low speed at a constant pulse rate of 15 pulses per second. This is because the target value of −450 mmHg (I) is very close to the target value so that the vehicle can be stopped quickly. Then, when the measured value reaches -460 mmHg (H) (timing D), the sequencer 84 stops the generation of the pulse.
About 0.5 seconds later, the motor 80 stops at the timing E, and the measured value at this time is the target value of -450 mmHg (I).
Is consistent with.

This flow rate adjustment is performed by using the V / F converter 83 when the measured value at the initial stage of adjustment is far from the target value.
While the adjustment time is shortened by rotating S79 at a high speed by proportional control, the sequencer 84
Is used to rotate the TAS 79 at an extremely low speed to complete the adjustment without overshooting the target value. This is because the return spring included in the throttle body 70 cannot be adjusted in the opposite direction due to the characteristics of the return spring, so that the target value cannot be exceeded. If it should be exceeded, the valve 72 must be returned to the fully closed state and the adjustment must be restarted from the beginning.

[0010]

However, the conventional adjusting method has the following problems.

The first problem is the time required for adjustment. As described above, the adjustment time is shortened by using the proportional control at the initial stage of the adjustment, but it still takes about 10 seconds from the adjustment start to the adjustment end timing E. This is much longer than the time required for manual adjustment (about 3 seconds for a normal worker and about 1 second for a skilled worker), which is a major obstacle to labor saving in the adjustment process. Of course, the adjustment time can be shortened by increasing the proportional coefficient or the maximum rate of the V / F converter 83, but the sequencer 84
There is a risk of over-adjustment occurring when switching to (timing C). Further, if so-called PID control in which integral control or derivative control is added to proportional control is performed, the time can be shortened without fear of overshooting, but the control formula is complicated and expert knowledge is required for its determination.

The second problem is how to deal with diversification of models. There are many models of the throttle body 70, and the bore diameter differs depending on the model. The larger the bore diameter, the larger the change in the negative pressure with respect to the rotation amount of the TAS 79. Therefore, the timings A to C in the lower part of the graph in FIG.
The steep slope to. Therefore, the throttle body 70
It is necessary to change the proportional coefficient of the V / F converter 83 according to the bore diameter of. Throttle body 7 with small bore diameter
This is because if the throttle body 70 having a large bore diameter is adjusted with the proportional coefficient for adjusting 0 being set, there is a strong possibility that overshoot will occur. Due to these problems, even if the adjustment work is automated, there is almost no merit.

The present invention has been made in order to solve the above-mentioned problems, and adjustment can be performed in a time equivalent to manual adjustment, there is no fear of overshooting, and the same control is performed regardless of the model to be adjusted. An object of the present invention is to provide a flow rate adjusting device that can be adjusted by.

[0014]

To achieve this object, the invention of claim 1 is directed to a motor for driving a valve, a pressure sensor for measuring a pressure at a position downstream of the valve, and a control means for controlling the motor. a flow regulating device provided with the door, the said control means, and the difference obtained by subtracting the target value of <br/> Karaso indicated value of the pressure sensor, a plurality of difference condition expressed by the degree of their Difference membership function that defines the relationship with the assigned probability distribution, and the previous reading of the pressure sensor
Seki and deviation obtained by subtracting the current indicator value, the probability distribution assigned to the plurality of deviation condition expressed by the degree of their from
The specified deviation membership function, the rotation speed of the motor is specified for each combination in which the difference condition and the deviation condition are both non-negative, and the instruction of the pressure sensor is given.
The combination including the difference condition when the value is equal to the target value
"Rotation stop" is specified, and the previous value indicated by the pressure sensor
Deviation when the deviation obtained by subtracting the indicated value from this time is the maximum
Specify "Stop rotation" for the combination that includes the conditions, and
For the combination of,
The faster the rotation speed is specified, the more significant the deviation is.
An output rule matrix designating a rotation speed slower as a deviation condition is stored, and at least one of the plurality of difference conditions and the plurality of deviation conditions has an unequal interval on the difference membership function or the deviation membership function. The control means determines the probability distribution to each of the difference conditions according to the instruction value of the pressure sensor, determines the probability distribution to each of the deviation conditions according to the value of the time change, and determines the difference condition and the difference condition. For each combination with the deviation condition, the lower of the probability allocated to the difference condition and the probability allocated to the deviation condition is selected as the allocation probability of that combination, and the rotation determined in the output rule matrix is determined by this allocation probability. It is characterized in that a weighted average of speeds is calculated and used as an output to the motor.

In this flow rate adjusting device, the pressure at the position downstream of the valve is measured by the pressure sensor, and the indicated value is input to the control means. The control means performs the following processing for this input. First, regarding the indicated value,
And previously been the target value is determined, and the difference obtained by subtracting the target value of Karaso indicated value, a difference membership function that defines the relationship between the probability distribution assigned to the plurality of difference condition expressed by the degree of its definition Has been done. The probability distribution to each difference condition is determined by the input instruction value. on the other hand,
For the time variation of the indicated value, the last instruction of the pressure sensor
Of the deviation obtained by subtracting the indicated value this time from the value, and the probability distribution assigned to the multiple deviation conditions expressed by the degree
A deviation membership function that defines the relationship is defined. Then, the probability distribution to each deviation condition is determined by the time change of the input instruction value . Thus, when the probability distribution to each difference condition and each deviation condition is determined, the lower one of the probabilities allocated to both conditions is selected for each combination of the difference condition and the deviation condition. As a result, one probability is assigned to each combination of both conditions. Then, the number of rotations of the motor is specified in advance by the output rule matrix for each combination of both conditions . Ie pressure
Includes the difference condition when the force sensor reading is equal to the target value
"Stop rotation" is specified for the combination, and the pressure sensor
The maximum deviation is the difference between the previous reading and the current reading.
If the combination includes the deviation condition when
For the rest of the combinations specified, a positively large difference
The faster the rotation speed is specified, the higher the difference condition
The slower the rotation speed is designated the deviation condition corresponding to the threshold deviation.
Has been. Therefore , the weighted average of the rotation speeds defined in the output rule matrix is calculated by the allocation probability, and the motor is driven by this calculated value. Here, at least one of the plurality of difference conditions and the plurality of deviation conditions is arranged at unequal intervals on the corresponding membership function so that the indicated value quickly becomes the target value.

[0016]

[0017]

Specifically, in the present invention, for example, the difference membership function is provided with a difference condition such that the indicated value is “approximately equal” to the target value, “slightly larger”, “larger”, and the like. The deviation membership function defines deviation conditions such as “nearly zero”, “slightly large”, and “large” with respect to time change, and at least one of these is arranged at unequal intervals on the corresponding membership function. ing. In the output rule matrix, the larger the instruction value is from the target value, the larger the rotation speed is specified, and the smaller the change with time is, the larger the rotation speed is specified.
Therefore, the two inputs of the indicated value and its change over time specify the optimum rotation speed according to the situation to drive the motor and adjust the valve, which is almost the same as the manual adjustment by a skilled worker. The adjustment can be performed smoothly with the required processing time and accuracy, and the adjustment can be performed without changing the membership function or rule matrix regardless of the type of valve.

The invention according to claim 2 is based on claim 1.
In the flow rate adjusting device described in the paragraph 1, the plurality of difference conditions are arranged at a narrower interval as they are farther from a target value in the difference membership function.

In this flow rate adjusting device, the plurality of difference conditions are arranged at unequal intervals in the difference membership function, and are set to be narrower as they are farther from the target value. Therefore, the rotation speed when the indicated value is far from the target value. Even if a fast rule matrix is adopted, the motor control by the slow rotation speed rule works efficiently when the indicated value approaches the target value, and it stops at the target value without overshooting, so the flow rate can be adjusted quickly. You can

[0021] The invention according to claim 3, claim 1 or
Alternatively, the flow rate adjusting device according to any one of claims 2 to 5, wherein the plurality of deviation conditions are arranged at wider intervals as they are farther from zero on the deviation membership function.

In this flow rate adjusting device, a plurality of deviation conditions are arranged at unequal intervals on the deviation membership function, and are set wider as they are farther from zero. Therefore, the rule that the rotation speed is fast when the deviation is close to zero is set. Even if the matrix is adopted, when the deviation becomes large, the motor control according to the slow rotation speed rule works efficiently, and the flow rate can be adjusted quickly because the adjustment is stopped at the target value without overshooting.

Further, the invention according to claim 4 is the invention according to claim 1.
The flow rate adjusting device according to any one of claims 1 to 3 , wherein the same rotation speed is designated in a combination of the difference condition and the deviation condition to which the allocation probability is given. In this case, the weighted average is calculated with the highest allocation probability among those allocation probabilities as the allocation probability to the rotation speed.

In this flow rate adjusting device, if a combination of the difference condition and the deviation condition, to which the allocation probabilities are given, includes the one in which the same rotation speed is specified in the rule matrix, With respect to the rotation speed, the highest allocation probability is selected, and the appropriate probability is distributed. The weighted average of the rotation speeds is calculated based on the probability distribution thus determined, and the motor is driven to adjust the flow rate.

[0025]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention will be described in detail below with reference to the embodiments shown in the drawings. This embodiment embodies the present invention as an apparatus and method for adjusting the main flow rate of a throttle body.

The flow rate adjusting device according to the present embodiment is roughly divided into a detection system, an operation system and a control system. As shown in FIG. 1, the detection system includes an exhaust pump 74 that causes an air flow in the throttle bore 71 of the throttle body 70.
And a negative pressure sensor 76 attached to the flow path 75. The operation system includes a lever 78 that is provided coaxially with the valve 72 of the throttle body 70, a TAS 79 that pushes the lever 78, and a motor 80 that drives the TAS 79. The motor 80 is a step motor. As the control system, the pulse generation circuit 60
Is equipped with.

The pulse generation circuit 60 includes a negative pressure sensor 76.
Output value (voltage signal) of a predetermined time interval (for example, 0.2
A / D converter 10 for sampling every second) and converting from analog to digital, and A / D converter 10
Of the output signal of the comparator 11 which outputs a difference (hereinafter, simply referred to as “difference”) from the target value of the output signal of the A / D converter 10, and the output signal of the A / D converter 10 are compared with the value at the time of the previous sampling, and the time change (hereinafter, “ Differentiator 12) that outputs "deviation")
And a fuzzy inference unit 1 that determines the drive output to the motor 80 using a fuzzy control method based on the difference and deviation.
A D / A converter 13 for converting the output signal of the fuzzy inference unit 1 from digital to analog, and a V / F converter 14 for generating a pulse having a rate proportional to the voltage which is the output signal of the D / A converter 13. Is provided.

Here, the comparator 11 outputs the output signal of the A / D converter 10 to the target negative pressure value of -450 m.
The difference is output in comparison with a digital value corresponding to mHg, and the negative pressure measured by the negative pressure sensor 76 is larger in absolute value than the target value (-450 mmHg) (-600 m).
mHg), a positive value is output.
Further, the differentiator 12 outputs a positive value when the negative pressure measured by the negative pressure sensor 76 is decreasing, that is, when the pressure is increasing (-600 mmHg → -550 mmHg, etc.). There is. The maximum pulse rate that the V / F converter 14 can generate is 300 pulses per second.

The fuzzy inference unit 1 will be described here. In the fuzzy inference unit 1, some conditions and membership functions are defined for each of the difference and the deviation in order to determine the driving output. Further, it has a rule matrix for determining the drive output to the motor 80 according to the difference condition and the deviation condition.

First, the difference condition and its membership function will be described. The condition of difference is the difference expressed in a form including an ambiguous expression depending on its value. Here, "positively very large", "positively fairly large",
"Positively large", "Positively slightly large", "Almost 0",
Nine types of conditions are defined: "negatively large,""negativelylarge,""negativelylarge," and "negatively large." The relationship between each of these conditions and the difference value is as shown in Table 1, and there is an intermediate range between two adjacent conditions in which it cannot be uniquely determined. This intermediate range corresponds to the ambiguity that human judgment has.

[0031]

[Table 1]

A specific value of the difference, that is, what condition is assigned to the output value of the comparator 11, is determined by a probability distribution function called a membership function. The difference membership function is given as the graph of FIG. In the graph of FIG. 2, the vertical axis represents the probability (%) and the horizontal axis represents the difference (mmHg). Then, for the condition “just very large”, the curve a is
The condition "positively large" is curve b, the condition "positively large" is curve c, the condition "positively large" is curve d, and the condition "nearly 0" is curve e. The curve f is for "a little negatively large", and the curve g is for the condition "negatively large".
The probability distribution is shown by the curve h for the condition “significantly large in negative” and the curve i for the condition “very large in negative”. For example, if the difference value is 150 mmHg indicated by a broken line in the graph, a probability of 60% is assigned to the condition "positively large" and a probability of 40% is assigned to the condition "positively large". This corresponds to the fact that 60% of people would judge "just big" and 40% would judge "just quite big" if they judged.

According to Table 1 and FIG. 2, as the difference value deviates from 0 in the difference membership function, the intervals between the conditions become narrower. For example, 75 mmHg between the condition “almost 0” (e) and the condition “just slightly larger” (d).
While there is an interval of (the interval of values at which the probability is 100%, the same applies hereinafter), the interval between the condition "positively large" (d) and the condition "positively large" (c) is 60 mmHg,
Condition "positively large" (c) and condition "positively large"
The distance from (b) is 45 mmHg, and the distance between the condition "positively quite large" (b) and the condition "positively very large" (a) is 30 mmHg. The reason for providing such an arrangement is to make the output value of the negative pressure sensor 76 match the target value as quickly as possible. That is, in the area where the difference value is far from 0, the condition "positively very large" is satisfied.
While the adjustment speed is increased by allocating a large number of probabilities to a condition that means that the difference value is large, such as (a) and the condition “just quite large” (b), the difference value is 0.
In the region close to, the range in which a large number of probabilities are assigned to a condition that means that the difference value is small, such as the condition “almost 0” (e) or the condition “just a little larger” (d), is widened,
It prevents excessive adjustment.

In Table 1 and FIG. 2, the difference value is (mm
This is indicated by Hg), but this is because priority is given to intuitive ease of understanding, and it is the corresponding digital value that is actually processed by the fuzzy inference unit 1. In addition, as described in the section of the prior art, the throttle body 70 cannot be adjusted in the reverse direction when the return spring is excessive, so that the above-mentioned nine types of conditions are
"Negatively large", "Negatively large", "Negatively large", and "Negative very large" are not actually used, but "positively very large" and "positively large" are used. large",
"Positively large", "Positively slightly large", "almost 0" 5
Adjustment is made according to one condition.

Next, regarding the deviation, the conditions and the membership function thereof are determined in the same manner as in the case of the difference.
The condition of deviation is also “very large positively”, “significantly large positively”, “positively large”, “positively a little large”, “almost 0”, “negatively a little large”, “negatively large”. , "Negatively large", "Negative very large",
The relationship between each condition and the deviation value is shown in Table 2.

[0036]

[Table 2]

The deviation membership function is shown in the graph of FIG. In this graph, the condition "Absolutely large" is the A curve, the condition "Absolutely large" is the A curve, the condition "Positively large" is the C curve, and the condition "Positive" is “A little big” is the curve of d, condition “almost 0” is the curve of o, condition “slightly large” is the curve of k, condition “largely negative” is the curve of k The probability distribution is shown by the curve of "Negatively large" and by the curve of "Condition is very large". For example, the value of the deviation is 50 indicated by a broken line in the graph.
If mmHg, probability 70% for condition "positively large", probability 30% for condition "positively large".
Are allocated respectively.

According to Table 2 and FIG. 3, as the deviation value deviates from 0 in the deviation membership function, the intervals between the conditions become wider. For example, the condition "almost 0"
10 between (e) and the condition "just a little larger" (d)
While the distance is only mmHg, the distance between the condition "positively large" (d) and the condition "positively large" (c) is 20.
mmHg, and the interval between the condition "positively large" (c) and the condition "positively large" (b) is 30 mmHg, and the condition "positively large" (b) and the condition "positively very large". "(A) and the space | interval are 40 mmHg. The reason for providing such an arrangement is to make the output value of the negative pressure sensor 76 match the target value as quickly as possible. That is, in a region where the deviation value is far from 0 to a certain extent, there are many conditions such as the condition “positively very large” (a) and the condition “positively large” (b), which means that the deviation value is large. While the adjustment speed is increased by assigning the probability of, the deviation value such as the condition "nearly 0" (e) or the condition "just a little larger" (d) is small in order to prevent over-adjustment. The range in which many probabilities are assigned to the meaning condition is limited to the region where the deviation value is very close to zero.

In both Table 2 and FIG. 3, the deviation value is (m
The description in mHg) is for easy understanding as in the case of the difference, and what is actually processed by the fuzzy inference unit 1 is the corresponding digital value. Further, as in the case of the difference condition, the adjustment is made in one way due to the characteristics of the throttle body 70, and therefore, among the above-mentioned five types of conditions, “a little larger in negative”, “larger in negative”, and “negative” are set. The four, which are "significantly large" and "very negatively large," are not actually used, but "positively very large", "positively significantly large",
"Positively large", "Positively slightly large", "almost 0" 5
Adjustment is made according to one condition.

Next, the rule matrix will be described. The rule matrix is a list in which control rules for driving the motor 80 are specified for each combination of the difference condition and the deviation condition, and as shown in Table 3, for each combination,
One of the five control rules of "high-speed forward rotation", "medium-speed forward rotation", "low-speed forward rotation", "slow-speed forward rotation", and "rotation stop" is designated. However, as mentioned above, neither the difference condition nor the deviation condition is used, since the conditions of "a little large in negative", "large in negative", "significantly large in negative", and "very large in negative" are not used. It is within the range surrounded by the thick line in Table 3.

[0041]

[Table 3]

This rule matrix is one in which a control rule is applied to each condition combination based on the human sense in order to make an adjustment close to manual adjustment. For example, when the pointer reading of the negative pressure sensor 76 is still far from the target value (difference condition “very large positively”), but the pointer hardly moves (deviation condition “nearly 0”), the operator TAS79. Will turn faster. This corresponds to "high speed forward rotation" in the a-o column of Table 3. Then, even if the reading of the pointer is still far from the target value (difference condition "positively very large"), if the movement toward the target value becomes faster (deviation condition "positively considerably large" or The operator will slow down the speed at which the TAS 79 is turned (e.g. "just very large"). This corresponds to "low-speed forward rotation" in the a-a column of Table 3 and "slow-speed forward rotation" in the a-a column. Then, when the reading value approaches the target value (such as the difference condition “positively very large” → “positively large”), the TAS 79 will be stopped at last. This corresponds to “stop rotation” in the c-a column of Table 3. Control rules are similarly assigned to the other columns based on human senses.

[0043]

[Table 4]

Table 4 shows the rotation speed of the TAS 79 for each control rule, that is, the rotation speed of the motor 80 and the drive pulse signal therefor. When the control rule is high speed forward rotation,
The TAS 79 and the motor 80 rotate in the forward direction at their maximum rotation speed. The forward direction is a direction in which the opening degree of the valve 72 in the throttle body 70 is increased. At this time, the rate of pulses applied to the motor 80 by the V / F converter 14 is 600 pulses per second. Medium speed, low speed, and slow speed are three-fourths (450 pulses per second), one-half (300 pulses per second), and one-fourth as high speeds, respectively.
(150 pulses per second). However, this pulse rate is a case where a probability of 100% is assigned to each corresponding control rule, and the probability is actually distributed over a plurality of difference conditions and deviation conditions as described above, and thus the probability is Are distributed over a plurality of control rules, the actual drive pulse rate is determined by taking the probability distribution into consideration. The details will be described later.

Table 4 also lists the control rules for rotating the motor 80 in the reverse direction, but because of the one-way adjustment of the throttle body 70 described above, it is actually surrounded by a bold frame. Only forward or stop control rules are used, not reverse rotation. For this reason, in the rule matrix of Table 3, the reverse rotation control rule is not adopted. For example, the i-a column in Table 3 (combination of the difference condition “very large in negative” and the deviation condition “very large in positive”) and the like indicates that if reverse adjustment is allowed, the reverse rotation control rule is set. Is placed, but the rule of "stop rotation" is actually placed.

Next, a flow rate adjusting method executed by this flow rate adjusting device will be described step by step.

First, as shown in FIG. 1, the throttle body 70 is attached to the flow rate adjusting device. At this time the valve 72
Should be fully closed. Then, when the exhaust pump 74 is driven to cause the air flow F, the negative pressure sensor 76 is set to -700.
A negative pressure of about -670 mmHg is shown. Then, in the pulse generation circuit 60, the output voltage of the negative pressure sensor 76 is sampled every 0.2 seconds, and the A / D converter 10 converts it into a digital value. This digital value is input to the comparator 11 and the differentiator 12.

In the comparator 11, this digital value is a value (digital) corresponding to the target negative pressure value of -450 mmHg.
Output the difference compared to. At this point, the negative pressure sensor 76
Since the absolute value of the indicated value is larger than the target value, the comparator 11
The difference value output by is positive and is 220 to 220 mmHg conversion.
The value is about 250. On the other hand, the differentiator 12 compares the digital value input from the A / D converter 10 with the value at the previous sampling and outputs the deviation. At this time, the valve 72 is not fully closed and the instruction value of the negative pressure sensor 76 does not change, so the value of the deviation output from the differentiator 12 is 0 (mmHg conversion).

The difference value output from the comparator 11 and the differentiator 12
The value of the deviation output by is input to the fuzzy inference unit 1. Therefore, the fuzzy inference unit 1 uses the respective membership functions based on these two inputs, and the motor 80
Output to. This decision is made as follows. First, the probability distribution to each difference condition is determined according to the difference membership function (FIG. 2, Table 1) based on the difference value. At this point in time, the difference value is about 220 to 250 (mmHg conversion), so it is located in the plateau portion of the curve a as shown in FIG. 2, and therefore the difference condition “positively very large” is 100.
% Probability is assigned, and no probability is assigned to other conditions. On the other hand, based on the deviation value, the probability distribution to each deviation condition is determined according to the deviation membership function (FIG. 3, Table 2). At this point, the deviation value is 0 (mmHg conversion), so it is located at the top of the curve in FIG.
Therefore, a probability of 100% is assigned to the deviation condition “almost 0”, and no probability is assigned to other conditions.

Thus, when the probability distributions of the difference condition and the deviation condition are determined, the probability is assigned to the control rules by referring to the rule matrix (Table 3). At this point, the only condition that a finite probability that is not 0% is allocated is that the difference condition is "a very large", and the deviation condition is that it is "nearly 0". Both are assigned a probability of 100%. Therefore, the control rule to which the probability is assigned is only the "high-speed forward rotation" in the a-e column in Table 3. Then, for this "fast forward rotation", the lower of the probability assigned to the difference condition and the probability assigned to the deviation condition is selected, and is assigned as the assignment probability to the control rule. Whichever is lower is equivalent to AND in a logical operation. At this time point, all the probabilities are 100%, and thus the control rule "high speed forward rotation" is assigned a probability of 100%.

At this point, the probabilities are not assigned to other control rules, so that "fast forward rotation" is the only control rule having a finite probability. Therefore, a digital signal corresponding to "high speed forward rotation" is output from the fuzzy inference unit 1, converted into an analog voltage signal by the D / A converter 13, and further converted into a pulse having a rate proportional to the voltage by the V / F converter. It is output to the motor 80. The pulse rate is 600 pulses per second as shown in Table 4, and this pulse signal causes the motor 80 and the T
The AS 79 rotates at high speed in the forward direction. The forward end of the TAS 79 is advanced by rotation in the forward direction, so that the lever 78 is pressed and rotated, and the valve 72 of the throttle body 70 is gradually opened. Due to this increase in valve opening, the negative pressure sensor 7
Since the negative pressure measured by 6 becomes weak (the pressure rises and approaches the atmospheric pressure, the absolute value of the reading becomes small), the flow rate adjustment is continued based on the new output value of the negative pressure sensor 76.

For example, if the measured value of the negative pressure sensor 76 is -600 mmHg at a certain sampling time and -650 mmHg at the previous sampling time, the control at that time is performed as follows.

At this time, the difference value output by the comparator 11 is 150 (mmHg conversion), and the deviation value output by the differentiator 12 is 50 (mmHg conversion). Therefore, in the fuzzy inference unit 1, according to the difference membership function (FIG. 2, Table 1), the difference condition of the curve b is 40% in the difference condition “positively large”, and the difference condition of the curve c is 60% in the positive condition.
%, Allocate the probability (see the broken line in FIG. 2). Then, according to the membership function of deviation (Fig. 3, Table 2), 70% is allocated to the deviation condition "positively large" of the curve a, and 30% is allocated to the deviation condition "positively large" of the curve c. (See dashed line in Figure 3).

Then, the process proceeds to the rule matrix (Table 3). At this point, a finite probability is allocated to the two conditions for both the difference condition and the deviation condition.
Probabilities are assigned to the control rules in the four columns a, b, c, a, and c. This is shown in Table 5. Of these, focusing on the b-a column, this column is a combination of the difference condition "positively large" (b, probability 40%) and the deviation condition "positively large" (a, probability 70%). The lower one of 40% and 70% (corresponding to AND operation), 40%, is assigned as a probability to the control rule "slow speed forward rotation". Similarly, in the other columns of Table 5, the lower one of the probabilities assigned to both conditions is assigned as the probability of the control rule. In Table 5, the probability is displayed in the column marked with * in each column. Also, as is clear from Table 5, the sum of the probabilities assigned to each column is not always 10
It does not always become 0% or an integral multiple thereof.

[0055]

[Table 5]

In Table 5, b-a, c-a,
The control rules in the three columns of c-c are "slow speed forward rotation" and are common. Therefore, the highest 60% of the allocation probabilities in these three columns is the allocation probability for the control rule “slow forward rotation”. The highest value is selected corresponds to the OR in the logical operation. On the other hand, in Table 5, “low speed forward rotation” in the column b-c is not common to the other columns, and therefore the allocation probability of 30% is assigned to the control rule “low speed forward rotation” as it is. Probability.

When the distribution of the probabilities to the respective control rules is determined in this way, the drive pulse rate to the motor 80 is determined by the weighted average. Since the pulse rate for each control rule is shown in Table 4, the pulse rate P in this case is

P = X / Y (pulses per second) X = 300 x 30 + 150 x 60 Y = 30 + 60

And P = 200 (pulses per second). Therefore, a digital signal corresponding to this pulse rate is output from the fuzzy inference unit 1 and the D / A converter 1
It is converted into an analog voltage signal in 3, and further converted into a pulse by the V / F converter and output to the motor 80.
The pulse signal causes the motor 80 and the TAS 79 to rotate in the forward direction, and the valve opening of the valve 72 further increases.

In Table 5, the control rules in the four columns are common to each other. Therefore, for the common rules, the highest probability is selected and the probability of the control rule is selected, and then the weighted average is calculated. However, if the control rules in each column are all different, the weighted average is calculated using the allocation probability of each column as it is as the probability of the control rule in that column.

The valve 72 is opened by such control by the fuzzy inference unit 1, and the reading value of the negative pressure sensor 76 approaches the target value of -450 mmHg. This is shown in the graph of FIG. In this graph, the horizontal axis shows the elapsed time from the start of driving the motor 80, and the upper side shows the motor 80.
The driving pulse rate is shown on the lower side, and the reading value of the negative pressure sensor 76 is shown on the lower side. At the start of driving the motor 80, the measurement value of the negative pressure sensor 76 is about -670 mm.
Hg is shown and is constant until about 1 second later.
This is because the valve 72 does not move even when the motor 80 is driven because there is a slight gap between the tip of the TAS 79 and the lever 78 at the start of driving the motor 80.
At this time, the difference condition is "very large".
The deviation conditions are "nearly zero" and 100
% Probability is assigned and the motor 80 is normally rotated at a high speed of 600 pulses per second.

When the tip of the TAS 79 moves forward by the forward rotation of the motor 80 and comes into contact with the lever 78, thereafter, the motor 80 continues.
The movement of the TAS 79 due to the drive of is transmitted to the valve 72 via the lever 78, and the valve 72 is gradually opened. Therefore, the measured value of the negative pressure sensor 76 gradually increases as the opening degree of the valve 72 increases. When this rise starts, the probability distribution for the difference condition is from "very large positively" to "positively large", "positively large", or "positively slightly large", or even "nearly zero". Fluctuates,
The deviation condition is also from "almost 0" to "just a little larger".
Or the probability distribution fluctuates to be "positively large", "positively fairly large", or "positively very large". As a result, the allocation of probabilities to each control rule also fluctuates,
The drive pulse rate of the motor 80 also changes. In FIG. 4, the pulse rate is once 0 pulse per second approximately 0.6 seconds after the start of driving, but this is because the deviation condition is assigned a very positive value of 100% and the difference condition. It is considered that since the probability of 100% in total was assigned within the range of “positively large” to “almost 0”, the probability distribution of the control rule temporarily became “rotation stop” 100%. In this way, the target value I (-450m
When the speed approaching mHg) becomes faster, the speed of the TAS 79 is slowed down and further stopped to prevent the adjustment from going too far.

When the TAS 79 is temporarily stopped, the probability distribution of the deviation condition shifts from "positively very large" to "positively slightly large" or "nearly 0", so that the probability distribution of the control rule is "rotation stopped". From "low speed forward" and "medium speed forward"
And so on, and the TAS 79 rotates again. Then, as the reading value of the negative pressure sensor 76 approaches the target value I, the probability distribution of “almost 0” for the difference condition increases and the pulse rate decreases. When the read value matches the target value I, the probability of the difference condition "almost 0" becomes 100%, the control rule "rotation stop" is assigned a probability of 100%, and the TAS 79 stops. This completes the flow rate adjustment. In FIG. 4, the time required to complete the adjustment is about 1.2 seconds, which is comparable to the time required for manual adjustment by a skilled worker.

As described in detail above, according to the device and the method for adjusting the main flow rate of the throttle body according to the present embodiment, regarding the "difference" and "deviation" of the output value of the negative pressure sensor 76. Probability distribution to each condition is determined by each membership function, and the driving speed of the motor 80 is determined by referring to the rule matrix determined based on human senses by the probability distribution to determine the opening degree of the valve 72. Since the flow rate is adjusted by increasing the flow rate, the flow rate can be adjusted by matching the output value of the negative pressure sensor 76 with the target value with high accuracy and in the same time as when manually adjusting the flow rate. Nor. As a result, the main flow rate of the throttle body can be automatically and efficiently adjusted, and manual adjustment can be substituted to save labor.

Further, the rule matrix defining the control rule for each combination of the conditions of "difference" and "deviation" is:
Since the rules from the high speed rotation to the rotation stop are arranged based on the human sense, the same membership function and rule matrix are used for adjustment regardless of the model of the throttle body 70 (especially the diameter of the throttle bore 71). be able to. Therefore, the throttle body 7
Depending on the model of 0, there is no need to perform adjustment work on the device side,
It is possible to adjust various types of throttle bodies 70 with high efficiency.

In particular, in the present embodiment, the arrangement of each condition on the membership function of the difference is such that the larger the value of the difference is, the narrower the interval becomes.
(E) and the condition "a little larger than" (d) means that the difference value is small, the range in which a large number of probabilities are assigned is large, and the excessive adjustment is effectively prevented. . As a result, the maximum drive pulse rate of the motor 80 can be set to a high rate of 600 pulses per second, and the adjustment time can be shortened. In addition, the arrangement of each condition on the membership function of deviation is set to be wider as the deviation value becomes larger. Therefore, the condition "exactly very large" is used except when the deviation value becomes very small. A large number of probabilities are assigned to conditions such as the condition (a) and the condition "positively large" (b), which means that the deviation value is large, and the adjustment is performed at the highest possible pulse rate. This also contributes to shortening the adjustment time.

Although the present invention has been described with reference to the embodiments, the embodiments are not intended to limit the present invention, and various modifications and improvements can be made without departing from the spirit of the present invention. Of course there is. For example, in the above-described embodiment, the flow rate is adjusted by using only the forward rotation in consideration of the unidirectionality of the return spring of the throttle body 70. Alternatively, the rotation in the opposite direction may be used for the adjustment.

[0068]

As is clear from the above description, according to the present invention, adjustment can be performed in a time equivalent to that of manual adjustment by a skilled person, there is no fear of excessive adjustment, and regardless of the model to be adjusted. A flow rate adjusting device that can be adjusted by the same control is provided. This makes it possible to replace manual flow rate adjustment with automatic adjustment efficiently.
The effect that it plays in the industry is great.

[Brief description of drawings]

FIG. 1 is a diagram showing a main flow rate adjusting device for a throttle body according to an embodiment.

FIG. 2 is a graph showing a membership function of difference.

FIG. 3 is a graph showing a deviation membership function.

FIG. 4 is a graph showing a change over time of the drive pulse and the negative pressure when the flow rate is adjusted by the main flow rate adjusting device according to the embodiment.

FIG. 5 is a view showing a conventional main flow rate adjusting device for a throttle body.

FIG. 6 is a graph showing a change over time of a drive pulse and a negative pressure when a flow rate is adjusted by a conventional main flow rate adjusting device.

[Explanation of symbols]

1 Fuzzy inference section 11 comparator 12 Differentiator (deviation calculation means) 60 pulse generation circuit (control means) 80 motor 76 Negative pressure sensor

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁷ identification code FI G05D 7/06 G05D 7/06 Z (56) Reference JP-A-3-279028 (JP, A) JP-A-8-49572 ( JP, A) JP 7-110702 (JP, A) JP 5-332704 (JP, A) JP 5-173605 (JP, A) JP 7-40763 (JP, A) JP Japanese Patent Laid-Open No. 5-288647 (JP, A) Japanese Patent Laid-Open No. 5-59991 (JP, A) Japanese Patent Laid-Open No. 5-5448 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) F02D 41 / 04 310 F02D 9/02 F02D 41/20 310 F02D 41/34 F02D 45/00 364 G05D 7/06

Claims (4)

(57) [Claims] 1. A flow rate adjusting device comprising a motor for driving a valve, a pressure sensor for measuring pressure at a position downstream of the valve, and a control means for controlling the motor, wherein the control means includes the pressure a difference obtained by subtracting the target value of Karaso indicated value of the sensor,
And difference membership function that defines the relationship between the probability distribution assigned to the plurality of difference condition expressed by the degree of their, minus the current indicator value from the previous indicated value of the pressure sensor
Deviation and a deviation membership function that defines the relationship between the probability assigned to the plurality of deviation condition expressed distributed by the degree of that, the motor for each combination the difference condition and said difference condition is not both negative Specify the rotation speed and
Includes the difference condition when the force sensor reading is equal to the target value
Specify "stop rotation" for the combination, and
As for the condition, the difference condition corresponding to a large difference is faster.
Deviation condition that specifies rotation speed and is equivalent to a positive deviation
An output rule matrix specifying a rotation speed as slow as possible is stored, and at least one of the plurality of difference conditions and the plurality of deviation conditions is arranged at unequal intervals on the difference membership function or the deviation membership function, The control means determines a probability distribution to each of the difference conditions according to an instruction value of the pressure sensor, determines a probability distribution to each of the deviation conditions according to a value of the time change, and the difference condition and the deviation condition. For each combination, the lower of the probability assigned to the difference condition and the probability assigned to the deviation condition is selected as the assignment probability of that combination, and the weight of the rotation speed determined in the output rule matrix is determined by this assignment probability. A flow rate adjusting device, characterized in that a weighted average is calculated and used as an output to the motor. 2. The flow rate adjusting device according to claim 1, wherein the plurality of difference conditions are arranged at a narrower distance from a target value on the difference membership function. 3. A flow rate control device according to claim 1 or claim 2, flow of the plurality of deviation condition, characterized in that it is located farther wide-interval from zero on the deviation membership functions Adjustment device. 4. The flow rate adjusting device according to claim 1, wherein the same rotation speed is designated in a combination of the difference condition and the deviation condition to which the allocation probability is given. Flow rate adjusting device, the weighted average is calculated with the highest allocation probability among the allocation probabilities as the allocation probability to the rotation speed.