JPH0642298A - Ventilation controlling device of tunnel - Google Patents

Abstract

PURPOSE:To further reduce the deterioration in contaminant concentration or minimize the generation frequency of excess ventilation by judging the control output of a control arithmetic operation means for a ventilator and correct the judgment based on the fluctuation patterns of measured values of a contaminant concentration meter within a specified time in the past. CONSTITUTION:A control output for a ventilator is determined by a feed back control arithmetic operation means 41 based on measured values of a contaminant concentration meter 3 and a predetermined target value. Then, fluctuation patterns of the measured values of the contaminant concentration meter within a specified time in the past are recognized with a pattern recognition means 42 by using a neural network. The control output of the feed back control arithmetic operation device 41 is judged so as to see if it is proper or not based on the fluctuation patterns of the contaminant concentration recognized by the pattern recognition means 42. Furthermore, if the control output is not proper, compensation processing will be carried out by a compensation means 43.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a ventilation control device for a tunnel which controls the number of operating ventilators according to the concentration of pollution in the tunnel, and in particular, implements appropriate ventilation control while observing the changing tendency of pollution. A ventilation control device for a tunnel.

[0002]

2. Description of the Related Art Generally, road tunnels are polluted by exhaust gas emitted from automobiles. Especially, in the case of a long tunnel or a tunnel with a lot of traffic, the pollutant concentration in the tunnel is measured, and The number of operating ventilators is controlled so that the pollution concentration is always below the allowable value. By the way, pollutants include various substances,
Among them, two variables, that is, smoke concentration and carbon monoxide concentration (hereinafter referred to as CO concentration) were often controlled.

However, after that, the smoke density is often controlled. This soot concentration is the smoke transmittance (hereinafter, V
This is because it can be represented by the light transmittance. That is, the VI measurement value is 100
The closer it is to%, the better the visibility is, and from another perspective, it means that the soot concentration is low. This means that if the VI measurement value is controlled so as to exceed a certain allowable value, the CO concentration can be maintained at almost the allowable value or less.

Therefore, in the recent tunnel ventilation control system,
Feedback control is mainly performed on the VI measurement value, and interrupt control is performed when the CO concentration exceeds a predetermined level.

As the most appropriate conventional tunnel ventilation control system, for example, Japanese Patent Application Laid-Open No. 56-64099 has been proposed. This control method is based on VI as shown in FIG.
When "upper limit 1-lower limit 1" is set as the target region of the measured value, several levels of control levels, for example, upper limit 2, lower limit 2 and lower limit 3 are set above and below the target region, and the VI measured value is at the upper limit side management level. Reduce the number of operating ventilators,
On the contrary, when the lower limit side control level is exceeded, the number of operating ventilators is increased and the VI measurement value is controlled so as to enter the target range.

More specifically, the VI measurement value has a lower limit of 1
When the lower limit 1 is continuously exceeded for t ₁ after exceeding, the number of operating ventilators is increased by a preset number. After this correction of the number of operating units, the change state is observed without operating the ventilator during the effect waiting time of t ₂ minutes, and when the VI measurement value is still below the lower limit 1, Increase the number of operating vehicles again. On the other hand, when the vehicle returns to the target area after the lapse of t ₂ minutes, the operating number of the ventilators is maintained and the ventilators are not operated.

[0007]

Therefore, in the tunnel ventilation control system as described above, the increase or decrease in the number of operating ventilators is determined depending on whether or not the VI measurement value exceeds a predetermined management level. Therefore, the following various problems occur.

A. When the VI measurement value is higher than the target area and when the VI measurement value has a sharp downward tendency, the VI measurement value itself is higher than the target area, so control is performed to reduce the number of operating ventilators. As a result, there is a problem that the number of operating ventilators is excessively reduced, and the VI measurement value drops significantly and sharply below the target region.

B. Next, when the VI measurement value tends to fall within the target region, the VI measurement value itself is within the target region, and therefore the number of operating ventilators is not operated. Therefore,
In this case, the VI measurement value often falls below the target area, and when the VI measurement value falls below the target area, the number of operating vehicles is controlled for the first time.
There is a problem that causes deterioration of the I measurement value.

C. Further, when the VI measurement value is below the target region and has a rapid upward tendency, the VI measurement value is originally below the target region, and therefore the number of operating ventilators is increased based on the VI measurement value. As a result, the VI measurement value easily exceeds the target range due to an increase in the number of operating ventilators. In this case, the ventilator is operated more than necessary, and wasteful power is consumed.

Therefore, if the number of operating ventilators is controlled by observing only the VI measurement value as described above, it is easy to cause an extreme decrease (deterioration) of the VI measurement value or excessive ventilation. In particular, the deterioration of the VI measurement value is very dangerous for driving a car. On the other hand, in the case of over-ventilation, although the visibility in the tunnel is good, since the ventilator is operated more than necessary, useless power is consumed and power cost increases.

Therefore, in order to improve the above-mentioned various problems, it is important to quickly grasp the changing tendency of the VI measurement value, but it is difficult to grasp the changing tendency for the following reasons. .

One of them is, in a general tunnel ventilation control system, a dead time until a sensor detects a change in the pollution concentration in the tunnel, and a delay time until the effect appears after the actual control. Therefore, the correct trend of change cannot be grasped only by changing the VI measurement value in a short time. In other words, it is necessary to see the change tendency of the VI measurement value of at least about 5 to 10 minutes in the past, but since no operation is performed during that time, the same problem as the above-mentioned a, b, and c occurs,
The pollution status in the tunnel is greatly deteriorated.

The other one is that not only are there various types of vehicles passing through the tunnel, from those with a large amount of exhaust gas to those with a small amount of exhaust gas, but the traffic volume also changes frequently. As a result, the distribution of the pollution concentration is biased, and the VI measurement value also largely changes, so it is very difficult to grasp the change tendency of the VI value.

The present invention has been made in view of the above situation, and corrects the pollution concentration in the past by performing a correction process to always maintain the pollution concentration stably near the target region, and An object is to provide a ventilation control device for a tunnel that realizes energy-saving operation.

[0016]

In order to solve the above problems, the invention according to claims 1 and 2 controls a ventilator in a tunnel based on a measured value of a pollution concentration meter installed in the tunnel. In the ventilation control device of the tunnel
In addition to feedback control calculation means for obtaining a control output for the ventilator using a measured value of a pollution concentration meter such as a meter (fume transmittance meter) or anemometer and a preset pollution concentration target value, a neural network And a contamination concentration fluctuation pattern recognition means for recognizing a fluctuation pattern of the measured value of the pollution concentration meter within a predetermined time in the past, and the feedback control calculation based on the fluctuation pattern of the pollution concentration recognized by the pattern recognition means. A ventilation control device for a tunnel provided with a control output correcting means for judging whether or not the control output of the means is an appropriate control output, and if it is an inappropriate control output, performing a correction process.

[0017]

Therefore, in the inventions corresponding to claims 1 and 2, by taking the above-mentioned means, the feedback control calculation means uses the measured value of the pollution concentration meter and the preset pollution concentration target value. The control output for the ventilator is calculated by using this control output.However, this control output alone can quickly respond to the state of the pollutant concentration at the measurement timing, which changes moment by moment, but the ventilator is controlled while considering the change tendency for a certain period of time. Can not.

Therefore, after the patterns having a plurality of types of change tendency are stored in advance in the contamination concentration fluctuation pattern recognition means composed of a neural network, the VI measurement value of a predetermined time in the past is actually inputted to the neural network, and the VI is concerned. It is determined which pattern the variation pattern of the measured value belongs to, and the pattern data of the determination result is sent to the control output correction means.

Here, the control output correction means receives the control output from the feedback control calculation means to control the ventilator, but when the pattern data is received from the pollution concentration fluctuation pattern recognition means, the control is performed from this pattern data. Whether the output is appropriate or not is judged, and if it is an unsuitable control output, an appropriate correction process is performed while considering the change tendency of the contamination concentration, so that the contamination concentration is stably maintained near the target area. It also contributes to energy saving operation.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. FIG. 1 is a functional block diagram for explaining a main part of the device of the present invention, FIG. 2 is a diagram schematically showing a configuration of a tunnel structure and ventilation equipment to which the device of the present invention is applied, and FIG. 3 is a pollution concentration fluctuation pattern recognition of FIG. It is a figure which shows the neural network which comprises a means.

First, the structure of the tunnel structure and the ventilation equipment will be described with reference to FIG. 1 is a tunnel in which various vehicle travels, the jet fan 2 ₁ to 2 ₁₀ is installed in the ceiling of the tunnel 1 in Fig. Further, for measuring the physical quantity related to the pollution concentration, for example, when the VI meter 3 for measuring the fume transmittance, that is, the VI measurement value is in either or both of the wellheads A and B, or in the direction of the ventilation wind indicated by the arrow, the leeward direction. The VI measurement value which is installed near the wellhead B on the side and measured by the VI meter 3 is sent to the ventilation control device 4.

The ventilation control device 4 calculates a control output (the number of operating jet fans) such that the VI measurement value measured by the VI meter 3 maintains the vicinity of the VI target value, and this control output is used as the jet fan. It is sent to the control device 5. The jet fan control unit 5 has the function of selectively operating and stopping the jet fan 2 ₁ to 2 ₁₀ based on a control output transmitted from the ventilation control device 4 (number of operating).

Specifically, the ventilation control device 4 compares the VI measurement value measured by the VI meter 3 with a predetermined VI target value as shown in FIG. 1, and the VI measurement value is the VI target value. In addition to the feedback control calculation means 41 for calculating the control output for the jet fan so as to approach the value,
A contamination concentration fluctuation pattern recognition means 42, a control output correction means 43, a storage means 44 and a tunnel ventilation process 45 are provided.

This contamination concentration fluctuation pattern recognition means 42
Recognizes the type of variation pattern of the VI measurement value measured by the VI meter 3 in the past several minutes, for example, using a neural network. As shown in FIG. 3, this neural network has a hierarchical structure composed of at least three layers of an input layer, an intermediate layer and an output layer, and a neuron group in the input layer has five patterns 1 shown in FIG. ~
By inputting each one of 5 and varying the weighting factors of the neurons in the intermediate layer and the output layer, a predetermined level is output from an output terminal having an output layer. For example, pattern 1 showing the tendency of change in the neuron group in the input layer
When the VI measurement value belonging to is input, the weighting coefficient is set so that the output terminal z _{1 in the} output layer outputs the level “1” or a value near the level “1”. Similarly,
In the case of the pattern 2 showing the change tendency of the VI measurement value, the weighting coefficient is set so as to output the level "1" or a value near the level "1" from the output terminal z _{2 having the} output layer.
Hereinafter, the pattern is similarly stored.

It should be noted that each of the patterns 1 to 5 represents the changing tendency of the VI value with respect to the time t in the past several minutes, and the VI measurement value belonging to the pattern 1 is almost the same as the pattern 1 as shown by the dotted line in the figure. This means that the variation pattern has a similar change tendency, and in this case, it means that the output terminal z ₁ having the output layer outputs a level “1” or a value close to the level “1”.

The control output correction means 43 receives the control output of the feedback control calculation means 41, uses a fuzzy control rule stored in advance in the storage means 44, and uses a calculation parameter, for example, an upper limit V representing a target area.
In addition to I _U , lower limit VI _L , and VI target value, it also has a function of executing fuzzy calculation using VI measurement value and the like to correct the control output.

[0027] The tunnel ventilation process 45, which includes a jet fan 2 ₁ to 2 ₁₀ and the jet fan control unit 5 to be controlled as shown in FIG. 2, the jet fan based on the output from the control output correction unit 43 To start and stop. Next, the operation of the apparatus configured as described above will be described.

A physical quantity related to the pollution concentration in the tunnel is measured from the VI meter 3 installed in the tunnel and sent to the feedback control calculation means 41. Here, the feedback control calculation means 41 monitors the VI measurement value of the VI meter 3 at every predetermined sampling cycle (for example, 1 minute cycle), and calculates the control output using the following calculation formula, for example. △ VI = VI target value-VI measurement value (1) △ N _JF = K _JF · △ VI (2) N _JFOUT = N _JF + △ N _JF・ ・ ・ (3)

Here, ΔVI: is a VI value deviation (%), K
_JF : Control gain (unit /%), N _JF : Current number of jet fan operating units (unit), △ N _JF : Correction amount of jet fan operating unit (unit), N _JFOUT : Control output to jet fan (unit) Is. The control gain K _JF has a ΔVI of 10
It is set to a predetermined value (0.1) such that it increases by, for example, one unit each time it increases by%.

When the number of operating jet fans is corrected by the feedback control calculating means 41 in this way, thereafter, the jet fans are not operated for a predetermined required effect waiting time.

At this time, in the contamination concentration fluctuation pattern recognition means 42, for example, the VI measurement value (VI ₁ to VI of the VI total 3 for the past 5 minutes obtained by sampling every 10 seconds).
VI ₃₀ ) is converted into a normalized value having a minimum value of 0 and a maximum value of 1 to facilitate pattern identification and supplied to the input layer of the neural network.
It is identified which of the five types of patterns the VI measurement value variation pattern _per minute belongs to, for example, five patterns shown in FIG. 4, and when it belongs to pattern 2, for example, the output end z _{1 of} the output layer,
Of z ₂ , ..., A level close to 1 is output from the output terminal z ₂ . That is, when the variation pattern of the VI measurement value is close to the patterns 1 to 5 shown in FIG. 4, the outputs closest to the level 1 are obtained from the output ends z _{1 to} z _{5 of the} corresponding output layers.

Therefore, when the output from the j-th neuron in the intermediate layer is y _j and the output from the k-th neuron in the output layer is z _k , this z _k is a value in the range of 0 to 1, and is most equal to 1. A pattern corresponding to a close value is selected as a VI measurement value variation pattern. However, when the value output from each of z _{1 to} z ₅ does not have a value of, for example, 0.7 or more, another pattern (for example, pattern 6) is used. Also, the weighting factors in the neural network are learned in advance by backpropagation. For example, when there are five types of patterns, first, the pattern 1 is input, the weighting factors of the intermediate layer and the output layer are changed, and stored so that the level 1 is output from z _1. Then, the pattern 2, the pattern 3 ,. The weighting factors of the intermediate layer and the output layer are changed in accordance with the above, and stored so that level 1 is output from z ₂ , z ₃ ,. Next, a method of calculating the output by the neural network will be described. First, the output y _{j of the} middle layer is

[0033]

[Equation 1]

It is obtained from Here, y _j : output of the intermediate layer, w _ji : weighting coefficient between the i-th neurons of the input layer, V
I _i : output of the i-th neuron in the input layer (input and output are the same in the input layer), f: function representing the input / output characteristics of the neuron. Next, the output z _k of the output layer is calculated as in the case of the intermediate layer as described below.

[0035]

[Equation 2]

Here, z _{k is the} output of the output layer, and w _kj is the weighting coefficient between the j-th neuron of the intermediate layer and the k-th neuron of the output layer. Further, as the function f, for example, a sigmoid function as shown below is used. f (u) = 1 / (1 + e- ^u ) (8)

Further, a backpropagation method for learning the weighting coefficient in advance will be described. Here, the error function in the output layer is defined as in equation (9).

[0038]

[Equation 3]

Here, z _{k is the} output of the k-th neuron in the output layer, and T _{k is the} teaching data (desired value) for the k-th neuron in the output layer. Back propagation is
This is a method of correcting the weighting coefficient so that this error function approaches the minimum value, and the correction amounts _{Δw kj} and Δw _ji of the weighting coefficient are calculated by the following equations.

[0040]

[Equation 4] In the above formula, ε is a parameter that determines the magnitude of one modification, and when the partial differential terms of formulas (10) and (11) are expanded and arranged, the modification amount of the weighting factor is as follows. Can be represented. Δw _kj = −ε · δ _k · y _j (12) Δw _ji = −ε · δ _j · VI _i (13)

[0041]

[Equation 5] Further, in order to suppress the vibration at the time of learning the weighting factor and accelerate the convergence of the learning, it is effective to use the following expressions instead of the expressions (12) and (13). _{Δw kj} (t) = −ε · δ _k · y _j + α · _{Δw kj} (t-1) (16) Δw _ji (t) = −ε · δ _j · VI _i + α · Δw _ji (t-1) (17) Here, α (0 <α <1) is a parameter for stabilizing learning, and t represents the number of times of learning.

Then, if the fluctuation pattern is recognized by the pollution concentration fluctuation pattern recognition means 42 as described above,
Data representing the type of the variation pattern is sent to the control output correction means 43. Here, the control output correction means 43
The control output of the feedback control calculation means 41 is corrected based on the variation pattern type data. This correction process is performed using a fuzzy control rule or the like stored in the storage means 44 in advance. Specifically, (1) If the variation pattern of the VI measurement value is a pattern other than the pattern 3 or the pattern belonging to FIG. 4, the correction process is not performed.

(2) When the control output for increasing the number of operating vehicles is output from the control calculation means 41, the variation pattern of the VI measurement value is the pattern 1 or the pattern 2, and
When the relationship of VI measurement value> lower limit VI _L is satisfied, the operating number of jet fans is not increased because the control operation unit 41 has already increased the operating number. That is, the current number of operating vehicles is maintained. This is a rule when the VI measurement value tends to increase, and is to prevent the jet fan from rotating too much.

(3) On the other hand, when the control output for reducing the number of operating vehicles is output, the variation pattern of the VI measurement value is the pattern 4 or the pattern 5 and the upper limit VI _U > V.
When there is a relation of I measured value> VI target value, the operating number of jet fans is not decreased because the control operation means 41 has already decreased the operating number of jet fans. That is, the current number of operating vehicles is maintained. This is a rule when the VI measurement value tends to decrease, and prevents the VI measurement value from extremely decreasing.

(4) In the case of the control output in which the correction amount of the number of operating vehicles is zero, the variation pattern of the VI measurement value at that time is the pattern 4 or the pattern 5, and the VI target value + ΔVI> VI When there is a relation of measured value> VI target value-ΔVI, the operating number of jet fans is increased by the operating number set by the control calculation means 41. Then, after the correction processing is performed by the control output correction means 43 as described above, the number of operating jet fans is controlled via the jet fan control device 5.

Therefore, according to the configuration of the above-described embodiment, the number of operating units is determined while the feedback control calculation unit 41 executes the calculation control such that the VI measurement value follows the VI target value. Different from the control to increase / decrease the number of operating vehicles depending on whether the measured VI value exceeds or exceeds the upper / lower limit of the target area, the direction of rapidly converging to the VI target value according to the change of the pollution state of the tunnel. Since the number of operating vehicles can be determined and the operation waiting time required after the number of operating vehicles is corrected is set so that the operation is not performed, the stability of control can be ensured.

Further, a plurality of patterns showing VI changes within a fixed time are stored in advance in the neural network, and it is recognized to which pattern the fluctuation pattern of the VI measured value in the past fixed time belongs. So the past V
The change tendency can be surely grasped from the I measurement value data, and the number of output terminals of the output layer constituting the neural network is matched with the number of types of patterns, and the output level of those output terminals is most predetermined. Since a pattern close to the level is taken out and which pattern it belongs to is recognized, it is possible to sufficiently cope with changes in various VI measurement values, and it is possible to know a future change prediction when the number of operating vehicles is changed.

Further, in addition to the pattern specified by the neural network, the feedback control calculation means 41 determines the future change prediction according to the fuzzy control rule using the VI measurement value, the VI target value, the upper and lower limits, etc. Since the determined number of operating units is corrected, the number of ventilators is not excessively reduced or excessively reduced.
As a result, deterioration of pollutant concentration and frequent repetition of excessive ventilation are eliminated, and useless power consumption can be suppressed, which greatly contributes to energy saving operation.

Although the feedback control means 41 of the above embodiment determines the number of jet fans to be operated based on the equations (1) to (3), the number of jet fans to be operated depends on whether the target area is exceeded or not. Or the well-known P
The number of operating vehicles may be determined using I or PID arithmetic control.

Further, in the feedback control means 41,
Although only the VI measurement value was used as the physical quantity related to the pollution concentration, the control of the wind speed in the roadway is also very important for the ventilation control because the decrease of the wind speed in the roadway leads to the deterioration of the pollution concentration.

FIG. 5 is a diagram showing a configuration in which such wind speed is also taken into consideration. That is, the anemometer 46 is installed in the tunnel, and the wind speed value is supplied to the feedback control calculation means 41 'in addition to the VI measurement value.
The number of operating vehicles may be determined by performing fuzzy calculation based on a fuzzy control rule using variables.

Furthermore, in the above-mentioned embodiment, the jet fan is taken as an example of the ventilator, but it can be similarly applied to a blower or a dust collector which can continuously control the air volume. However,
The control output in the case of the jet fan was the operating number, but the control output in the case of the blower and the dust collector is the operating air volume.

Although five types of patterns are used as the basic patterns, for example, two types of patterns of five types are used.
You may make it memorize | store the pattern compositely combining two or more. In addition, the present invention can be modified in various ways without departing from the scope of the invention.

[0054]

As described above, according to the present invention, the fluctuation pattern of the contamination concentration is recognized, and the control output of the feedback control calculation means is corrected based on the recognized fluctuation pattern. It is possible to provide a tunnel ventilation control device that can perform a correction process quickly by catching the change tendency of the concentration, reduce the pollution frequency and the frequency of excessive ventilation, and realize energy-saving operation.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing an embodiment of a ventilation control device for a tunnel according to the present invention.

FIG. 2 is a diagram showing a layout structure of a tunnel structure and ventilation equipment to which the ventilation control device for a tunnel according to the present invention is applied.

FIG. 3 is a configuration diagram of a neural network that constitutes a contamination concentration fluctuation pattern recognition means of the device of the present invention.

FIG. 4 is a diagram showing a basic pattern stored in a neural network.

FIG. 5 is a functional block diagram showing another embodiment of the device of the present invention.

FIG. 6 is a time chart illustrating a conventional tunnel ventilation control system.

[Explanation of symbols]

1 ... tunnel 2 ₁ to 2 ₁₀ ... jet fans (ventilators), 3 ... VI thermometer, 4 ... ventilation controller, 41, 41 '...
Feedback control calculation means, 42 ... Contamination concentration fluctuation pattern recognition means, 43 ... Control output correction means, 44 ... Storage means.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Internal reference number FI Technical display location G05D 21/00 A 7314-3H G06F 15/18 8945-5L G06G 7/60

Claims (2)

[Claims] 1. A tunnel ventilation control device for controlling a ventilator in a tunnel based on a measurement value of a pollution concentration meter installed in the tunnel, comprising: a measurement value of the pollution concentration meter and a preset pollution concentration; A feedback control calculation means for obtaining a control output for the ventilator using a target value; and a pollution concentration fluctuation pattern recognition means for recognizing a fluctuation pattern of the measured value of the pollution concentration meter within a predetermined time in the past using a neural network. And whether or not the control output of the feedback control calculation means is an appropriate control output based on the variation pattern of the contamination concentration recognized by the pattern recognition means, and if it is an inappropriate control output, a correction process is performed. A ventilation control device for a tunnel, comprising: 2. The pollution concentration meter is a VI meter (fume transmittance meter).
The ventilation control device for a tunnel according to claim 1, wherein one or both of the above and an anemometer are used.