JP7310688B2 - Method and apparatus for estimating thickness of deposits on inner wall surface of exhaust gas passage - Google Patents

Description

The present invention relates to a method and apparatus for estimating the thickness of deposits on the inner wall surface of an exhaust gas passage, and in particular, it is possible to estimate the thickness of deposits with high accuracy, extend the continuous operation time, and increase the operating rate of the inner wall surface of the exhaust gas passage. The present invention relates to a deposit thickness estimation method and apparatus.

When operating a waste incineration facility, melting facility, various heating furnaces, melting furnaces, etc., such as those illustrated in FIG. In FIG. 1, 10 is an incinerator or a melting furnace, 12 is a secondary incineration combustion furnace (also referred to as a secondary combustion furnace), 13 is a burner, 14 is a cooling tower, 15 is water spray equipment, 16 is dust removal equipment, 18 is an induced draft fan, and 20 is a chimney. Note that the temperature reduction for efficiently removing HCl and SOx can be performed by the water spray equipment 15 provided in the temperature reduction tower 14 as shown in FIG. A boiler 22 for heating is provided, or as shown in FIG. This is done by lowering the temperature by spraying.

By the way, the adhesion to the wall surface is caused by the components that volatilize at high temperatures during incineration or melting and the components that accompany combustion exhaust gas, etc. (1) When the temperature drops, they condense and adhere to the wall surface, and (2) The flow velocity is slow. (3) adhesion and deposition on condensate;

This deposit is caused by the blockage of the flue caused by the deposit itself, which increases the pressure inside the furnace and the flue. It is a big problem because it causes various operational and equipment troubles, and in some cases, emergency shutdown of the furnace. In addition, the same situation occurs when the deposits fall down and accumulate in the lower hopper, clogging nearby pipes. For this reason, aiming at automatic separation as much as possible, (i) adding a chemical that makes it difficult to adhere to the fuel (Patent Document 1), or (ii) injecting a dispersion liquid onto the side wall (Patent Document 2). , (iii) Bin blowing with air, air hammer, hammering, and other automatic peeling equipment are being used, but a drastic solution has not been reached.

JP-A-2002-285179 Japanese Patent Application Laid-Open No. 2006-29701

As an actual adhesion countermeasure, the operation of the equipment is eventually stopped and cleaning (scraping) is performed, but regarding the adhesion status of the adhesion, especially when the target is waste, the same thing happens every time. Instead, they may differ greatly in both quantity and quality. In such a case, the number of personnel, tools, required days, etc., to be prepared for the cleaning work may be excessive or insufficient, resulting in over-preparation or poor preparation, resulting in waste. Therefore, when the operation of the equipment is stopped and the peeling work is performed, by estimating the adhesion situation in advance, it is possible to more appropriately predict the cleaning date, number of personnel, equipment, work days, etc., so it is highly effective in eliminating waste. .

In addition, in the event of an emergency shutdown of the furnace, depending on the cause, gas, slag, and ash may be ejected from the furnace or flue. At this stage, it is necessary to release the gas from an emergency discharge valve or the like, contain it with an inert gas, or cool it while replacing it. In particular, in the case of an emergency stop in a melting furnace, slag, raw materials, etc. will solidify, and in order to restart the melting furnace, it is necessary to chip off the slag with a rock drill and remove deposits from the furnace wall. Therefore, an emergency shutdown is the most desirable situation for the operation of the furnace, and it is desirable to predict the occurrence of a problem as early as possible.

FIG. 4 shows the case of a cooling tower 14 (inner diameter 3 m to 6 m, height 7 m to 10 m) to which deposits (hereinafter referred to as deposits 8) tend to adhere due to a rapid temperature drop. In order to estimate the adhesion state of the deposit 8 inside the melting facility (for example, the deposit thickness t, which is about 800 mm at the maximum), (1) measure the internal state from the outside and estimate the internal state. , and (2) measuring internal conditions from within.

Regarding the former (1) measuring the internal situation from the outside and estimating the internal situation, it is conceivable to use ultrasonic waves, radioactivity, and the like. The accuracy of ultrasonic waves is not very high, and the measurement results are greatly affected by the presence of cavities and the degree of coarseness and fineness of the adhesion state, so there is a problem with estimation accuracy. On the other hand, since radioactivity has extremely high sensitivity, it may be possible to measure with high accuracy even if there is a lot of dust in the flue. Not yet. A method of opening a viewing window and measuring with a camera or light has also been studied, but there are problems such as dust adhering to the viewing window and restrictions on the installation position of the viewing window.

On the other hand, with respect to the latter (2) measuring the internal conditions from the inside, if the outside air is ventilated to the furnace and flue, etc. when the furnace is stopped, or after cooling by opening the manhole 14a, measurement can be performed. Although it is possible, the measuring equipment is exposed to high temperatures and harmful gases such as dust, SOx, and HCl during operation, so advanced or large-scale equipment devices are required. , did not reach a practical level.

The dimensions of target flues and reaction vessels vary from small to large, but most are in the range of several meters to 5 meters in inner diameter and 5 to 10 meters in height. For such shapes, deposits grow towards the center of the cross-section inside the container and can be up to 800 mm or more in length. When the position and size of the adhering matter are measured using light, sound waves, or the like, the adhering matter itself becomes an obstacle depending on the position of the sensor and affects the measurement. For this reason, attempts have been made to insert sensors longer or install many sensors on the wall. However, when the temperature inside the flue or reaction vessel is high, many complicated and large-sized cooling devices are required to cool the sensor and ensure the cleanliness of the measurement surface. I didn't. In addition, the flue gas of incinerators and melting furnaces generally contains several g to 10 g/m ³ of dust, which interferes with measurement.

Based on these facts, the state of adhesion can be grasped by empirically determining the number of days of operation when adhesion begins to affect the operation, stopping the operation after this number of days, opening the manhole 14a, etc., and cooling. Cleaning was done after grasping the shape. In other words, it was a response based on experience, and there was no method that could measure the adhesion amount during operation and in the presence of dust.

When the sensor is inserted into the flue from the outside, as mentioned above, it is necessary to maintain a constant temperature through air or cooling water for cooling. It blows into the reaction vessel. Therefore, if the amount of cooling medium increases, problems such as an increase in the amount of exhaust gas and a local decrease in gas temperature may occur.

The present invention has been made in view of this situation, and it is possible to increase the operating rate by estimating the thickness of the deposit with high accuracy, extending the continuous operation time, and improving the efficiency of the cleaning work. An object of the present invention is to provide a method and apparatus for estimating the thickness of deposits on the inner wall surface of an exhaust gas passage.

The inventors have identified the following characteristics of the measuring instrument: 1. 2. With respect to high temperature resistance, it is possible to perform three-dimensional measurement in as short a time as possible with as few measurement points as possible, and to increase the measurement accuracy. 3. It can be measured even in a sandstorm environment due to the presence of dust. 4. The measured properties of the gas do not change so much as to affect the measurement and operation; On the premise that the auxiliary measurement equipment will not become large-scaled, we examined proposals that can be drastically improved in terms of (1) the measurement work and the measurement equipment, and (2) the measurement environment.

(1) Regarding measurement work and measuring equipment, (a) high-accuracy measurement that enables three-dimensional measurement in as few measuring points as possible in the shortest possible time, and (b) the measuring equipment must have the necessary high-temperature resistance. and (c) that the measuring equipment and auxiliary measurement equipment should not be large-scaled.

(2) With regard to the measurement environment, the main items to be examined are (I) measures that can be taken even in sandstorm environments due to the presence of dust, and (II) measures that do not affect the measurement due to gas properties. be.

First, with regard to the measuring equipment, for example, as a method that can measure the shape of the deposit 8 adhering to part or most of a large space of 3 m to 5 m in diameter and 5 m to 10 m in height in a short time with high three-dimensional accuracy, As shown in FIG. 5, a three-dimensional optical sensor (also called a 3D sensor) 32 using visible, ultraviolet, infrared or laser light was considered suitable.

The measurement position is longer than the internal extension length (deposit thickness) t of the deposit 8, and the measurement point is one on the circumference of the cross section as shown in FIG. 6(A). Therefore, as shown in FIG. 6B, there are two locations on the circumference of the same cross section (approximately facing positions) to three locations (positions dividing the circumference into three). This reduces blind spots and enables overall measurement with a minimum number of devices. FIG. 6(C) shows a longitudinal section corresponding to FIG. 6(B).

Regarding the effect of dust concentration, the following studies were conducted. Although it depends on the type of incinerator or melting furnace and the type of object, the dust concentration at the furnace outlet is generally less than ten g/m ³ . Since the exhaust gas temperature is about 700° C. to 300° C., the substantial concentration is about 1 to 5 g/m ³ . As a result of examining whether or not it is possible to measure the number of particles and the particle size distribution using a laser scattering particle size meter using dust actually collected from a melting furnace, the total number of particles of 1 μm or more was 0.35 g/ There was a measurement limit of about ^m3 . Although the degree of influence of dust concentration on particle size measurement and shape measurement by light is not necessarily the same, when the dust concentration is so high that the particle concentration cannot be measured, it is considered that light will not be transmitted. In other words, when measuring the deposit thickness t, if there is a lot of dust, the light cannot pass through and the measurement cannot be performed. Due to the influence of dust or the like, transmission of the laser used in the 3D sensor 32 is obstructed, and the ratio of the laser reflected back to the target point whose position is to be detected decreases. Here, the position detection rate [%] on the vertical axis of FIG. 7 indicates the rate of correct position detection for the laser sent to the measurement target point under this influence.

In the case of waste incinerators and melting furnaces, the amount of dust varies greatly depending on the type of waste and the type of furnace. As mentioned above, when comparing the dust concentration at the outlet of the furnace and the measurement limit dust concentration of the number of particles, the difference is several times to 10 times or more. it is not easy to do. Therefore, it was necessary to find a way to solve this problem, that is, to take measures that would allow measurement even in a sandstorm-like environment due to the presence of dust without stopping or cooling the furnace.

Incinerators and melting furnaces are generally operated with as little fluctuation as possible in order to maintain stability and carry out continuous operation for a long period of time. On the other hand, in order to measure the shape of deposits in the flue, it is necessary to greatly reduce the dust concentration as described above. Therefore, the inventors stopped the operation of the furnace for a short period of time, and carried out a stop test, thinking of measuring during that time. Since harmful gases are generated, as shown in Fig. 1, it is necessary to properly manage the processes after the secondary combustion using combustion aids such as gas, oil and coke, and combustion support materials such as air and oxygen. One, and especially in the case of melting furnaces, was the occurrence of areas where the melt partially solidified, which in some cases could be a problem during re-operation. In other words, it was found that a complete stoppage of operations, even for a short period of time, could lead to problems. From these results, it is possible to create a state in which the generation of dust is greatly suppressed for 30 minutes or more by maintaining the furnace temperature by burning the combustion aid and the combustion support material when the waste load is greatly reduced or zero. It was found that the furnace could be restarted without causing problems in the furnace.

Here, the reason why the state in which the generation of dust is greatly suppressed is set for 30 minutes or longer is to secure the minimum time required for the measurement.

That is, for example, in the case of the cooling tower 14, a suitable position is searched for in the manhole 14a or the handhole installed from the top to the bottom of the cooling tower 14, and the measurement is performed. 14a and the cover of the flange of the hand hole are removed, 2) connect cooling air and cooling water to the measuring instrument as necessary, insert and install, 3) perform measurement, 4) measuring instrument after measurement 5) Close the lid of the flange as before to complete.

The measurement time is about 5 to 10 minutes, but the work time for this is 10 minutes for installation and 10 minutes or more for cleanup, so the shortest time required for measurement is about 30 minutes.

However, depending on the type and shape of the furnace, the type of waste, etc., it may take 10 minutes or more to create a state in which the generation of dust is greatly suppressed. In this case, the minimum time required for stabilization is about 20 minutes.

Disadvantages of a longer measurement time include lower availability and longer time required to return to steady state operation.

Further, the grounds for setting the load factor to 20% or less for the significantly low-load operation is that the load factor of 20% is the maximum load for lowering the dust concentration required for measurement.

That is, in the case of ordinary waste incinerators and melting furnaces, the concentration of dust at the outlet of the furnace is about 1 to 5 g/m ³ at about 500° C. (as described above). If the load factor is lowered to 20%, the amount of combustible gas and combustion-supporting gas such as combustion air will also decrease at the same rate, so it is thought that there will be no change in the dust concentration. % and 0.3 to 1.5 g/m ³ at a load rate of 20% ^. The reason for this is thought to be that the flow velocity in the furnace and the flue became slow, which increased the coagulation and sedimentation effects, but the details are unknown. However, the upper limit of the load factor is 20% because there is a possibility that the dust concentration range in which 3D optical measurement can be performed is possible at a load factor of about 20%.

On the other hand, if only the waste load factor is set to 20% and the calorific value due to the amount of combustible gas and combustion-supporting gas is increased, and the total calorific value is maintained at the same level before and after changing the load factor, this does not lower the furnace temperature. It is effective, but if such an operation method is selected, the amount of exhaust gas does not change greatly, so the dust concentration decreases almost in proportion to the load factor, and is estimated to be 0.2 to 1 g/m ³ .

Based on the above two facts, 3D shape measurement can be performed smoothly by setting the upper limit of the waste load factor to 20% and operating at a load factor of 20% or less.

When the waste supply is stopped and the combustion aid and power required to maintain the furnace temperature are supplied (approximately 20% or less of the rated load of the furnace) for 30 minutes, the effects are as follows.

Measurements shall be taken once a week for 30 minutes and shall be returned for 1 hour. The furnace waste load factor is 0 for 30 minutes and averaged to 50% for 1 hour after recovery. In the current situation, the continuous operation time is 14 days, and the time from reactor shutdown to restart is 5 days. Availability = (14 x 24)/(14 + 5) x 24 = 0.736
Availability of invention results = (21 x 24 - 1.5 x 0.5 x 3) / (21 + 5) x 24 = 0.804
becomes. Although this difference may seem small at first glance, 1) cleaning time and cleaning costs can be reduced, and 2) efficiency can be increased by more than 6% without changing equipment and personnel.

It is desirable to conduct measurements frequently, but considering that reliability can be improved by combining long-term measurements and operation data, data accumulation and operation conditions, etc., such as once every few days or once a week By incorporating the huge amount of data in machine learning, the learning effect can be enhanced, and as a result, the frequency of measurement can be greatly reduced.

Another thing to be careful about is the measurement temperature, and it is necessary to cool the measurement probe of the 3D sensor 32 to below the heat-resistant temperature. Although this temperature varies depending on the manufacturer and model of the measuring instrument, it is said to be approximately 70°C.

In the case of an indirect heat exchange system using cold water using a water-cooling jacket 34 as shown in FIG. 8A, the equipment is relatively small but complicated. In the figure, 36 is a cable space. When gas (air) is used for cooling, as shown in FIG. 8(B), the measurement probe of the 3D sensor 32 is placed in the air cooling jacket 38, and the air cooling jacket 38 is filled with gas (air, non-flammable nitrogen N _{2 ,} etc.). gas), and the gas after cooling can be dissipated into the furnace or flue. The water-cooling jacket 34 and the air-cooling jacket 38 are provided with transparent peepholes made of heat-resistant glass or the like, through which light for measurement is transmitted and received and photographing is performed with a camera. FIG. 8C shows an example of using both water cooling and air cooling.

The charging and installation of the jackets 34 and 38 containing the measurement probes of the 3D sensor 32 is performed by opening the charging port of the pedestal provided in advance at an appropriate place (for example, manhole 14a or handhole) in the furnace or flue. Install. The measurement data becomes user-friendly by putting it into a 3D model.

In addition, by using the 3D laser measurement method, the 3D sensor (transmitting/receiving) 32 needs to be inserted above the height of the sticking object 8 projecting from the wall surface. As shown in C), using two 3D sensors 32A and 32B, by measuring surfaces facing each other from substantially facing positions, blind spots of each 3D sensor 32A and 32B and the deposit 8 itself It is possible to greatly reduce the part that interferes and cannot be measured. Since the growth of the deposit 8 itself changes slowly on a daily or weekly basis, in order to know this change, it is sufficient to measure once a day using one measuring instrument and protective component in order. is not required to be deeply charged into the flue or reaction vessel, cooling the 3D sensors 32A and 32B and maintaining a clean surface of the measuring device are facilitated, and operation is simplified.

The amount of cooling gas used in the air-cooling jacket 38 is a dusty environment at 800°C, and the heat-resistant effect was confirmed by using the air-cooling jacket 38 with a peephole and a ceramic inner surface with heat insulating material, in which the 3D sensor 32 was inserted. As a result, the amount of room-temperature cooling gas required to maintain 70°C for at least 15 minutes was at most 2% or less of the amount of gas passing through the furnace and flue.

The present invention has been made based on the above study results, and is a method for estimating the thickness of deposits adhering to the inner wall surface of an exhaust gas passage, which uses a three-dimensional optical sensor inserted into the exhaust gas passage. a three-dimensional measurement step of directly three-dimensionally measuring the shape of the deposit on the inner wall surface of the exhaust gas passage from the inside ; actual measurement data of the thickness of the deposit obtained from the three-dimensional measurement of the deposit in the three-dimensional measurement step; The problem is solved by providing a learning step of accumulating operation data to learn the deposit thickness and a prediction step of predicting the deposit thickness using the learning result obtained from the learning step. be.

The present invention also provides a device for estimating the thickness of deposits adhering to the inner wall surface of an exhaust gas passage, which is inserted into the exhaust gas passage and directly measures the shape of the deposit on the inner wall surface of the exhaust gas passage three-dimensionally from the inside. learning means for accumulating actual measurement data of the deposit thickness obtained from three-dimensional measurement of the deposit shape by the three-dimensional optical sensor and operation data to learn the deposit thickness; Prediction means for predicting the thickness of deposits using the learning result obtained from the learning means. It is something to do.

Here, the three-dimensional optical sensors used in the three-dimensional measurement process can be arranged at a plurality of locations on the cross section of the exhaust gas passage.

In addition, the measurement of the shape of deposits by the three-dimensional optical sensor used in the three-dimensional measurement process is carried out when the furnace is stopped for a short period of time or during low-load operation when the dust concentration is 0.35 g/m ³ or less. can be done.

Also, the three-dimensional optical sensor used in the three-dimensional measurement process can be used by inserting it into an air-cooling and/or water-cooling jacket.

The present invention measures the state of adhesion of deposits 8 inside an incinerator/melting furnace, exhaust gas treatment equipment, and flue with a 3D camera or the like, and obtains three-dimensional deposit thickness data by data processing and operation of the equipment. By combining the data, highly accurate measurements (thickness, spread, hardness, etc.) are performed, and the deposit 8 is effectively removed using a suitable instrument determined from the results. At that time, these data are machine-learned by artificial intelligence (AI) to control the type and combination of raw materials, processing amount, processing conditions, etc. to lengthen the operation time of one continuous unit, and clean the deposits 8 efficiently. At the same time, grasp the comprehensively effective operation method to enhance safety. Through these measures, we aim to save energy, save labor, improve safety, and reduce costs. The principle and function are as follows.

<1> Cooling the high temperature part of the 3D sensor 32 with air cooling, water cooling, etc. as illustrated in FIG. As shown in FIG. 7, the operation is performed by adjusting the raw fuel supply, the set temperature, etc. so that the concentration is 0.35 g/m ³ or less.

<2> By matching the three-dimensional shape grasp of the adhesion state by the 3D sensor 32 and the operation data 40 and 42 as shown in FIG. Along with the judgment, a jig or a safety tool suitable for the thickness t of the adhering matter and the removing work of the adhering matter 8 is selected to perform the removing work.

<3> Combining the contents of the operation data 40 and 42 with the three-dimensional measurement data 33 of the deposit 8 .
The operation data 40 and 42 include the properties (components and amounts) of the object to be treated, the types and amounts of additives used during treatment, the incineration/melting temperature, the temperature at the furnace outlet, the composition and amount of exhaust gas discharged from the furnace, the dust composition and dust. amount, secondary combustion conditions, gas cooling method, additives in the exhaust gas treatment process of harmful gas removal equipment, etc., and the adhesion position, adhesion amount, adhesion hardness, and adhesion composition (particularly related to salt composition) of the deposit 8 The main purpose is to infer associations. If necessary, perform intermittent measurements with increased frequency, clarify the adhesion status and changes over time, optimize the selection of the order in which objects such as waste are placed in the furnace, and adjust the temperature and load of the furnace. It is possible to optimize the operating conditions of the furnace and more accurately determine when to stop the furnace, leading to a longer continuous operation period of the furnace.

<4> Understand the effect of adhesion countermeasures during equipment operation.
Hammering equipment, air hammers, bin blows, etc. may be installed to prevent adhesion, but at present, it is difficult to determine the effect because they are installed in a black box that cannot be seen from the outside. If intermittent measurements with increased frequency are performed to clarify the adhesion status, the conditions of adhesion countermeasures (method, frequency, force), and changes over time in the effect, it will be possible to select an effective countermeasure method, and finally This will lead to a longer continuous operating period of the furnace.

As described above, in the present invention, the accuracy of predicting the deposit thickness t from the operation data is achieved by fusing and learning the deposit thickness measurement data based on the measured values of the 3D sensor 32 measured from the inside and the operation data of the furnace. increase Then, the deposit thickness t is determined by the predicted value, and when the thickness reaches a certain value or more (at the point when the operation of the furnace begins to be adversely affected), the operation is stopped and cooled, and the removal work is performed manually or by a robot. conduct.

Usually, waste incinerators with a processing scale of about 100t to 200t/day are often used. There are operations that start up and shut down in units of one day, operations that operate on weekdays and stops on weekends, and operations that operate continuously until cleaning work is required due to clogging or the like. It is common to heat and keep the temperature in consideration of start-up at the time of shutdown even for those that operate on a daily or weekly basis. Melting furnaces are the same size or slightly smaller than incinerators. Once started, it is common to operate continuously until cleaning work is required due to clogging or the like, or refractory parts need to be replaced due to wear or deterioration. In particular, since small-diameter pipes such as furnace outlets are prone to clogging, equipment such as installing automatic cleaning equipment and providing flanges on both ends of straight pipes for easy cleaning are being devised.

For both incinerators and melting furnaces, the efficiency of operation depends on the throughput per hour, the frequency of equipment maintenance, the life of the equipment, the amount of utilities used, and the number of personnel required to operate. When comparing efficiency, it is necessary to fix the operation because whether it is continuous operation, night stop, or weekend stop depends on the ease of securing workers and wages. Therefore, the description will focus on the continuous operation of a melting furnace, which is realistic and the effects are likely to be clear. When a melting furnace is operated continuously, it depends on the method, but in the case of a melting furnace using fuel such as coke, it is usually possible to operate continuously for about several months.

On the other hand, the period during which problems such as adhesion and clogging occur in the exhaust gas treatment system depends on the object to be treated, but it is as short as several weeks and as long as about one month.

For example, in the case of continuous operation for 2 weeks, a total of 5 days are required: 1.5 days for cooling, 2 days for cleaning, and 1.5 days for start-up after shutting down the furnace. is 5 days and the operating rate is 64% (9/14).

On the other hand, if the adhesion thickness estimation technology according to the present invention is used, it becomes possible to estimate the bottleneck portion, and there is a possibility that the operation can be extended by about one week. The operating rate in this case is 76% (16/21), which can be increased by about 20%. In addition to the operating rate, it is possible to estimate the location and thickness of deposits, so that it is possible to set up appropriate scaffolding and prepare cleaning tools, which can be expected to significantly reduce cleaning time. In this regard, it is possible to reduce the number of personnel, especially in the case of cleaning work when the deposits are radioactive, because the work can be completed in a shorter time than the work time stipulated by law. In addition, by comparing the object and treatment conditions with the radioactivity measurement values, the possibility of finding conditions that can extend the operating time increases, so the possibility of further increasing the operating rate increases.

A block diagram showing an example of a process targeted by the present invention and a range of processes targeted by the present invention. Block diagram showing a modification of the process of cooling the exhaust gas in FIG. Block diagram showing another modification of the process of cooling the exhaust gas in FIG. Longitudinal cross-sectional view showing deposits adhering to the cooling tower of FIG. Longitudinal cross-sectional view of a desuperheating tower showing an arrangement example of 3D sensors used in the embodiment of the present invention (A) (B) horizontal cross-sectional view and (C) vertical cross-sectional view for explaining the blind spot of the 3D sensor used in the embodiment of the present invention FIG. 2 is a diagram showing dust concentration and 3D shape measurement limit in an embodiment of the present invention. Sectional view showing the jacket of the 3D sensor used in the embodiment of the present invention FIG. 2 is a block diagram showing the configuration of a deposit thickness prediction process according to an embodiment of the present invention; FIG. 2 is a flow chart showing the processing procedure in the embodiment of the present invention; FIG. A diagram showing a learning method according to an embodiment of the present invention.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In addition, the present invention is not limited by the contents described in the following embodiments. In addition, the configuration requirements in the embodiments described below include those that can be easily assumed by those skilled in the art, those that are substantially the same, and those within the so-called equivalent range. Furthermore, the constituent elements disclosed in the embodiments described below may be combined as appropriate, or may be selected and used as appropriate.

First, the configuration and arrangement of sensors in the embodiment of the present invention will be described.

As shown in FIG. 5, the sensor for carrying out the present invention is inserted into the exhaust gas passage (reducing temperature tower 14) using, for example, a manhole 14a, and the shape of deposits on the inner wall surface of the exhaust gas passage is measured three-dimensionally. It has a 3D sensor 32 to measure. In this embodiment, the two 3D sensors 32A and 32B are the water-cooled jacket 34 shown in FIG. 8A, the air-cooled jacket 38 shown in FIG. 8B, or the water-cooled/air-cooled They are inserted into the jacket and arranged opposite to each other as shown in FIGS. 6(B) and 6(C).

Then, the output of the sensor is, as shown in FIG. , coke, etc.) charging amount, components, air/oxygen blowing amount, furnace, flue temperature, flue gas amount (air volume, wind speed), cooling water amount, ash, slag, fly ash generation amount, control items and control values Range, component analysis values, other operation (plan) data 40, and processing amount (charge amount), air/oxygen blowing amount, furnace, flue temperature, exhaust gas amount (air volume, wind speed), exhaust gas composition ( O ₂ , SO ₂ , SO, HCl, etc.), cooling water amount, electric power/fuel consumption, and other operational (actual measurement) data 42 are introduced, and a data accumulation learning unit 110 for accumulating and learning these, and data accumulation A computer 100 having a deposit thickness prediction model 120 created based on the output of a learning unit 110 and a deposit thickness prediction unit 130; (Decision) is introduced into a deposit thickness prediction processing circuit with an output 140 .

Here, FIG. 10 shows the procedure for estimating the deposit thickness in this embodiment.

First, in step 110, the shape of the adhering matter 8 is three-dimensionally measured.

Next, in step 120, deposit thickness measurement data (three-dimensional measurement data) 33 by 3D measurement and operation data 40, 42 are accumulated to learn the deposit thickness t.

Next, in step 130, the learning result is used to predict the deposit thickness t.

Furthermore, FIG. 11 shows how the deposit thickness is learned in step 120 .

First, at the learning stage in FIG. , to create a training data set 50 . At this time, abnormal data are excluded, and measurement and image data are merged.

Next, the learning data set 50 is input to the data accumulation learning unit 110 including the learning program 112, the pre-learning parameters 114, and the hyperparameters 116 that determine the configuration of the learning device, and the trained model 122, the trained parameters 124, A deposit thickness prediction model 120 having an inference program 126 is learned.

The learning result is input to the deposit thickness prediction unit 130, and based on the actual input data 128, the learned model 122 determines and outputs the deposit thickness t.

Based on the results, for example, (a) selection of raw materials and operating conditions is performed to minimize the deposit thickness t with sufficient types and amounts of raw materials. Alternatively, (b) when there are restrictions on the selection of raw materials, the deposit thickness is minimized by selecting raw materials and operating conditions that can minimize the deposit thickness t. Alternatively, (c) when stopping and cleaning are carried out as planned, regardless of raw material conditions, an operation selection is made to quickly increase the deposit thickness t, thereby maximizing the deposit thickness. Furthermore, desirable raw material conditions can also be known.

As described above, in the data accumulation learning unit 110 of the computer 100, the 3D sensor 32 measures the adhesion state of the deposits 8 in the furnace of the incineration/melting furnace, the exhaust gas treatment equipment, and the flue, and the data is obtained by data processing. By combining deposit thickness measurement data (also referred to as three-dimensional measurement data) 33 by 3D measurement, facility operation (plan) data 40, and operation (actual measurement) data 42, more accurate estimation (thickness, spread , hardness, radioactivity concentration, etc.), and using the deposit thickness prediction model 120 created based on this result, the deposit Along with determining when to remove the deposit 8, an instrument suitable for the removal work is selected and used to effectively remove the deposit 8 manually or by a robot.

As described above, if the deposit thickness measurement data 33 obtained by 3D measurement and the furnace operation data 40 and 42 are integrated and learned to estimate the deposit thickness t, the deposit thickness t can be estimated from the operation data 40 and 42. Prediction accuracy is improved. As a result, the deposit thickness t can be predicted with high accuracy without 3D measurement, so the frequency of low-load operation of the furnace required for 3D measurement is reduced, and the operating efficiency of the furnace is improved.

It can also be used as an index for searching for an operation method when the deposit thickness t is thin.

In the present embodiment, the present invention is applied to the temperature reduction tower of the incinerator and the melting furnace, but the application of the present invention is not limited to this, and the target range and , incinerators and melting furnaces.

t Deposit thickness 8 Deposit 10 Incinerator or melting furnace 12 Secondary combustion furnace 14 Temperature reduction tower 22 Boiler 32, 32A, 32B Three-dimensional optical sensor (3D sensor)
33: Deposit thickness measurement data by three-dimensional measurement (three-dimensional measurement data)
34 Water-cooled jacket 38 Air-cooled jacket 40 Operation (planned) data 42 Operation (actual measurement) data 50 Learning data set 100 Computer 110 Data accumulation learning unit 112 Learning program 114 Pre-learning parameter 116 Hyper Parameter 120 Deposit thickness prediction model 122 Learned model 124 Learned parameter 126 Inference program 128 Input data 130 Deposit thickness prediction unit 140 Deposit thickness (decision) output unit

Claims (8)

A method for estimating the thickness of deposits adhering to the inner wall surface of an exhaust gas passage, comprising:
A three-dimensional measurement step of directly three-dimensionally measuring the shape of deposits on the inner wall surface of the exhaust gas passage from the inside using a three-dimensional optical sensor inserted into the exhaust gas passage;
a learning step of learning the deposit thickness by accumulating actual measurement data of the deposit thickness obtained from the three-dimensional measurement of the deposit in the three-dimensional measurement step and operation data;
A prediction step of predicting the deposit thickness using the learning result obtained from the learning step;
A method for estimating the thickness of deposits on an inner wall surface of an exhaust gas passage, comprising: 2. The method for estimating the thickness of deposits on the inner wall surface of the exhaust gas passage according to claim 1, wherein the three-dimensional optical sensors used in the three-dimensional measurement step are arranged at a plurality of locations on the cross section of the exhaust gas passage. The measurement of the shape of deposits by the three-dimensional optical sensor used in the three-dimensional measurement step is performed when the furnace is stopped for a short time or during low-load operation when the dust concentration is 0.35 g/m ³ or less. The method for estimating the thickness of deposits on the inner wall surface of the exhaust gas passage according to claim 1 or 2. 4. Estimation of deposit thickness on inner wall surface of exhaust gas passage according to any one of claims 1 to 3, characterized in that said three-dimensional optical sensor used in said three-dimensional measurement step is used by being inserted into an air-cooling and/or water-cooling jacket. Method. A device for estimating the thickness of deposits adhering to the inner wall surface of an exhaust gas passage,
a three-dimensional optical sensor that is inserted into the exhaust gas passage and directly measures the shape of deposits on the inner wall surface of the exhaust gas passage three-dimensionally from the inside ;
learning means for learning the thickness of the deposit by accumulating actual measurement data of the deposit thickness obtained from the three-dimensional measurement of the shape of the deposit by the three-dimensional optical sensor and operation data;
Prediction means for predicting deposit thickness using the learning result obtained from the learning means;
An apparatus for estimating the thickness of deposits on the inner wall surface of an exhaust gas passage, comprising: 6. The apparatus for estimating the thickness of deposits on the inner wall surface of the exhaust gas passage according to claim 5, wherein the three-dimensional optical sensors are arranged at a plurality of locations on the cross section of the exhaust gas passage. 6. The three-dimensional optical sensor measures the shape of deposits when the dust concentration becomes 0.35 g/m ³ or less due to short-time shutdown or low-load operation of the furnace. 7. The device for estimating the thickness of deposits on the inner wall surface of the exhaust gas passage according to 6 above. 8. An apparatus for estimating the thickness of deposits on an inner wall surface of an exhaust gas passage according to any one of claims 5 to 7, wherein said three-dimensional optical sensor is inserted in an air-cooled and/or water-cooled jacket.

Images