JP7623129B2 - Soil purification device and soil purification method - Google Patents

Description

The present invention relates to a soil purification device and a soil purification method for purifying soil containing contaminants.

Conventionally, soil contaminated with volatile organic compounds (VOCs), heavy metals, perfluoroalkyl and polyfluoroalkyl compounds (PFAS), polychlorinated biphenyls (PCBs), dioxins, oils, etc. (hereinafter referred to as target substances) has been purified, for example, by in-situ thermal desorption. In-situ thermal desorption is a purification method in which the target substances are volatilized and recovered by heating them in-situ, without excavating the contaminated soil.

The target substances listed above cause soil contamination by adsorbing to soil particles, and dissolve into groundwater, causing groundwater contamination. Various methods have been proposed for heating soil in conventional purification equipment using in-situ thermal desorption, such as a method in which a heating well (or heating hole) is installed that leads to the area of soil to be purified, and a heater is inserted into the hollow part of the heating well to heat the soil to a predetermined temperature.

In this method, the temperature of the soil is raised to about 350°C, which causes the target substances to be thermally decomposed underground or adsorbed to the soil particles to be desorbed and the fluidity of the target substances to be increased, causing them to volatilize or dissolve into the surrounding groundwater. The volatilized target substances are drawn to the surface from the recovery well and either thermally decomposed in an exhaust gas treatment device or the like installed on the surface, or are recovered by being adsorbed onto activated carbon or the like. The target substances that have dissolved into the liquid phase are pumped to the surface together with groundwater from the recovery well and detoxified in a wastewater treatment device or the like.

For example, Patent Document 1 discloses a purification system that is installed in a preheat hole (heating hole) excavated in the contaminated soil around a recovery well and has a heater that preheats the groundwater that seeps into the preheat hole.

Patent Document 2 shows a system (ISTD: In-Situ Thermal Desorption) that directly heats soil in situ using a heating well to desorb target substances from the soil or decompose them underground, as well as the shape of the heater used for heating.

Patent document 3 discloses a system (IPTD: In-Pile Thermal Desorption) in which excavated contaminated soil is sealed in a space surrounded by structures such as walls and heated with a heater. Patent document 4 shows the shape of a heater used to heat the soil in situ.

JP 2006-272273 A Patent No. 4509558 Patent No. 5486684 U.S. Pat. No. 1,020,1042

However, the ground to be purified is composed of multiple different strata (loam, sand, gravel, silt, clay, etc., or alternating layers of these), and the thermal conductivity of the ground varies greatly depending on the properties of the ground, such as porosity, density, specific heat, and the degree of saturation due to the presence of groundwater. Conventional heaters are designed to heat the entire depth range of the heating well uniformly, resulting in inefficient heating. In order to accommodate differences in thermal conductivity in the depth direction based on the properties of the soil, it is possible to use multi-stage heaters installed in the depth direction, or to change the density of the coils that make up the heater in the length direction. However, since multiple heating wells are distributed two-dimensionally in the area of the ground to be purified, there is a problem that if the heating of all heating wells in the depth direction is to be managed individually, the control becomes three-dimensional and extremely complicated.

To address these issues, it is possible to create a model that takes into account the ground characteristics in advance and set the heating pattern using an analytical method based on numerical simulation (heat conduction analysis). However, the results of numerical simulation (physical equations) do not necessarily match the current complex conditions of the ground, and in this case it is necessary to adjust the analytical parameters and repeat the numerical simulation. In this way, the operation of changing the current pattern to the heater while analyzing the constantly changing temperature data during in-situ purification becomes more complicated the more heating wells there are, and there is a problem that the analysis takes a huge amount of time.

The present invention aims to provide a soil purification device and a soil purification method that can efficiently heat the ground by controlling multiple heating wells according to the properties of the ground.

To achieve the above object, the present invention provides a soil purification device that heats a target area of the ground to volatilize and purify the target substances, comprising: a plurality of heater devices arranged along the depth direction of the target area to be treated; a plurality of detectors arranged around the heater devices; a plurality of recovery wells arranged along the depth direction of the ground to recover the target substances volatilized from the ground heated by the heater devices; and a control unit that controls the plurality of heater devices based on the detection results of the detectors, wherein the control unit controls the power to be supplied to the heater devices based on different control patterns, performs machine learning of the temperature change pattern of the ground corresponding to the control pattern detected by the detectors, and controls the plurality of heater devices based on the optimal control pattern selected based on the learning results.

According to the present invention, the control unit can machine-learn the control patterns of multiple heater devices arranged in the area to be treated, thereby reducing energy costs and significantly shortening the construction period.

The heater device of the present invention has a heater section having a plurality of heater elements connected in series along the depth direction, and a switching section that energizes one or more of the heater elements selected from the plurality of heater elements, and the detection section has a detection section having a plurality of temperature sensors arranged along the depth direction, and the control section energizes the switching section with different energization patterns based on the control pattern to energize the heater section and generate different temperature changes in the depth direction, and may perform machine learning of the temperature change pattern in the depth direction of the ground detected by the detection section based on the energization pattern, and control the switching section with the optimal energization pattern selected based on the learning result.

According to the present invention, when a heater device has multiple heater elements in the depth direction, the current conduction pattern of the switching unit that energizes each heater element can be learned by machine learning, thereby making it possible to select the optimal control pattern for the complex control of multiple heater devices in three dimensions.

The control unit of the present invention may machine-learn the temperature change patterns in multiple depth regions along the depth direction relative to the current flow pattern based on the detection results of the detection unit.

According to the present invention, even if the ground has different thermal conductivity in the depth direction, the temperature change pattern can be learned by machine learning, and the optimal control pattern for the heater device can be selected.

The control unit of the present invention may control the heater device with the minimum power supplied to heat the ground to a target temperature based on the learning results.

According to the present invention, the heater device can be controlled with minimal power depending on the ground properties, which can significantly reduce energy costs.

The target temperature of the present invention may be set based on the properties of the ground and the type of target material contained in the ground.

According to the present invention, the heater device can be operated efficiently by setting the target temperature according to the environment where the ground properties and the types of target substances contained in the ground differ.

The present invention is a soil purification method for purifying a soil by heating a target area of the ground to volatilize the target substances, characterized in that the method controls power to be supplied to a plurality of heater devices arranged along the depth direction of the target area of the ground based on different control patterns, performs machine learning of a temperature change pattern of the ground corresponding to the control pattern based on detection results detected by a plurality of detectors arranged around the heater devices, controls the plurality of heater devices based on the optimal control pattern selected based on the learning results, and recovers the target substances volatilized from the ground heated by the heater devices from a plurality of recovery wells arranged along the depth direction of the ground.

According to the present invention, by learning the control patterns of multiple heater devices arranged in the area to be treated by machine learning, it is possible to reduce energy costs and significantly shorten construction time.

According to the present invention, it is possible to efficiently heat the ground by controlling multiple heating wells according to the properties of the ground.

1 is a cross-sectional view showing a schematic configuration of a soil purification apparatus according to an embodiment of the present invention. 1 is a plan view showing a schematic configuration of a soil purification apparatus. FIG. 2 is a diagram illustrating a schematic configuration of a heater device. FIG. 4 is a diagram showing a circuit configuration of a switching unit. 5 is a diagram showing the relationship between energization of a heater element in a switching portion and a switch that is turned on. FIG. FIG. 2 is a block diagram showing the configuration of a soil purification device. FIG. 2 is a diagram showing a state in which the soil purification apparatus is operated. 3 is a flowchart showing a flow of processes executed in the soil purification apparatus.

Below, an embodiment of a soil purification device and a soil purification method according to the present invention will be described with reference to the drawings. The soil purification device is a device that heats the area of the ground to be treated and purifies the target substances by volatilizing them.

As shown in Figures 1 and 2, the ground G of the treatment area A, where soil purification is required, is partitioned underground by a water shielding wall W. The water shielding wall W is, for example, driven into the ground in the depth direction. The upper end of the water shielding wall W is exposed at the ground surface E, and the lower end reaches an impermeable layer F through which groundwater does not easily penetrate. A groundwater level L is formed in the ground G of the treatment area A. The ground G of the treatment area A has strata formed in the depth direction, for example, from multiple soil layers with different properties in relation to thermal conductivity.

A soil purification device 1 is installed in the ground G of the treatment area A. The soil purification device 1 includes, for example, a water barrier W, multiple heating wells 10 that heat the ground G, multiple detection units 20 that detect the temperature of the ground around the heating wells 10, multiple recovery wells 30 that recover pollutants that volatilize from the heated ground, an exhaust gas treatment device 40 that collects and decomposes the target substances recovered by the multiple recovery wells 30, a power supply device 50 that supplies power to the heating wells 10, a control device 60 that controls the heating wells 10 and the power supply device 50, and a diffusion prevention cover 70 that covers the upper part of the ground G of the treatment area A.

The diffusion prevention cover 70 is provided to prevent the target substance contained in the ground G of the treatment target area A from evaporating and diffusing directly toward the ground. The diffusion prevention cover 70 is provided to cover the ground surface E of the treatment target area A, which is surrounded by the water barrier wall W.

A plurality of heating wells 10 are provided in the treatment target area A (see FIG. 2). The heating wells 10 are arranged two-dimensionally distributed in a planar view in the treatment target area A. The heating wells 10 generate heat in the treatment target area A, and heat a predetermined area T of the ground G surrounding the wells 10, which is to be heated. The predetermined area T is an area set in the ground G to be heated by one heating well 10. The predetermined area T is set, for example, in a roughly hexagonal shape in a planar view in the treatment target area A.

The multiple heating wells 10 are arranged, for example, on lattice points that divide multiple predetermined areas T in the treatment target area A. A recovery well 30 is arranged at the center of the predetermined area T. For example, three heating wells 10 are arranged around one recovery well 30. Three detection units 20 are arranged between each of the three heating wells 10.

The heating well 10 is disposed in a depth direction below the ground surface E. The heating well 10 includes a drilled hole 10H formed vertically downward from the ground surface E, and a heater device 11 inserted into the drilled hole 10H. The drilled hole 10H is formed in a depth direction from the ground surface E to near the boundary of the impermeable layer F.

The heater device 11 includes, for example, a cylindrical casing 12, a heater unit 13 inserted into the casing 12, and a switching unit 15 (see FIG. 4) (described later) that controls the heater unit 13. The casing 12 is formed, for example, from a steel pipe with suitable thermal conductivity and closed upper and lower ends. The casing 12 is inserted into the drilling hole 10H. The heater unit 13 extending in the depth direction is inserted into the casing 12.

The heater section 13 has multiple (e.g., five) heater elements R1-R5 connected in series along the depth direction. The multiple heater elements R1-R5 are supplied with power from the power supply unit 50 by switching the circuit with a switching section as described below. Of the multiple heater elements R1-R5, electricity is passed through one or more heater elements selected by the switching section to create various heating states in the depth direction of the ground G. The detailed configuration and control of the heater section 13 will be described later.

A number of detection units 20 are arranged around the heating well 10. The detection units 20 include a number of temperature sensors 21 arranged along the depth direction of the ground G. The temperature sensors 21 detect the temperature at each depth in the ground G. The data detected by the detection units 20 is transmitted to the control device 60. The configuration of the control device 60 will be described later.

The recovery well 30 recovers target substances that have volatilized into the gas phase (vaporized target substances) from the ground G heated by the heater devices 11 of the multiple heating wells 10 arranged around it, and target substances that have dissolved into the groundwater. The recovery well 30 is arranged in the depth direction below the ground surface E. The recovery well 30 includes a drilled hole 30H formed vertically below the ground surface E, and a casing 31 inserted into the drilled hole 30H.

The recovery well 30 is formed in the depth direction from the ground surface E to near the boundary of the impermeable layer F. The casing 31 is formed, for example, in a cylindrical shape. The underground portion of the casing 31 is formed, for example, from punching metal. A filter (not shown) is provided around the casing 31, which blocks soil particles while allowing volatilized pollutants to pass through. The upper and lower ends of the casing 31 are closed. A pipe 32 is connected to the upper end of the casing 31. An exhaust gas treatment device 40 is connected to the downstream side of the pipe 32.

The exhaust gas treatment device 40 sucks air from inside the casing 31 through the piping 32. Pollutants that have volatilized from the heated ground G flow into the casing 31 through a filter. The pollutants are collected in the exhaust gas treatment device 40 through the piping 32. The exhaust gas treatment device 40 performs decomposition processing of the target substances collected from the collection well 30. The exhaust gas treatment device 40 is placed, for example, on the ground outside the treatment target area A.

The exhaust gas treatment device 40 recovers the target substances that have volatilized into the gas phase from the heated ground G via the piping 32 and have been attracted into the inside of the casing 12 of the recovery well 30. The exhaust gas treatment device 40 is equipped with an adsorbent such as activated carbon. The adsorbent captures the volatile target substances that have volatilized into the gas phase. The captured target substances are decomposed thermally or chemically treated by the exhaust gas treatment device 40 to be rendered harmless.

Next, the detailed configuration of the heater device 11 will be described.

As shown in FIG. 3, the heater section 13 in the heater device 11 is formed of an electric heating coil that generates heat when electricity is passed through it along the depth direction. Power is supplied to the heater section 13 from a power supply device 50 via a pair of terminals 13T. The heater section 13 has multiple terminals connected to it corresponding to the depth region R, and multiple heater elements R1-R5 connected in series by the terminals are formed along the depth direction. Each heater element R1-R5 heats the corresponding depth region R having a predetermined depth in the depth direction. For example, multiple terminals T1-T4 are connected between the heater elements R1-R5. The multiple terminals T1-T4, 13T are connected to a switching section 15.

As shown in FIG. 4, the switching unit 15 is connected to the power supply device 50. The power supply device 50 is, for example, a DC power supply. The power supply device 50 may be an AC power supply. The switching unit 15 is controlled by the control device 60 and supplies power from the power supply device 50 to each of the heater elements R1-R5 of the heater device 11. The switching unit 15 is, for example, configured with a switching circuit connected to multiple terminals T1-T4, 13T. The switching unit 15 has multiple switches S1-S10 connected to multiple terminals T1-T4, 13T.

As shown in FIG. 5, the switching unit 15 switches two switches selected from the multiple switches S1-S10 that are in the OFF state to the ON state, heating one or more consecutive heater elements and heating an arbitrary depth region R in the depth direction of the ground G. The above combination of heater elements R1-R5 and the circuit configuration of the switching unit 15 having the switches S1-S10 are only an example, and the switching unit 15 may be configured not only to heat one or more consecutive heater elements, but also to heat the ground G by any combination of one or more of the multiple heater elements R1-R5.

The switching unit 15 may be a device that controls a mechanical switch to an on or off state, or may be a device that controls an on or off state on a circuit configured using semiconductor elements. Selecting a heating method that efficiently heats the ground G by switching the multiple switches S1-S10 in the switching unit 15 between on and off states is an extremely complicated task. By efficiently controlling the heater device 11 to heat the ground, the total amount of energy consumed during the purification period can be reduced, shortening the purification period and reducing energy costs. The control of the proposed soil purification device 1 is described below.

As shown in FIG. 6, the multiple heater devices 11 in the soil purification apparatus 1 are controlled by a control device 60. The control device 60 includes, for example, a control unit 61 that controls the heater devices 11, and a storage unit 62 that stores data related to the control. The storage unit 62 is realized by a storage medium such as, for example, a RAM, a ROM, a HDD, or a flash memory.

The control unit 61 is realized, for example, by a processor such as a CPU executing a program (software) stored in the storage unit 62. Some or all of the functions of these components may be realized by hardware (including circuitry) such as an LSI, ASIC, FPGA, or GPU, or may be realized by a combination of software and hardware. The program may be stored in a storage device such as a HDD or flash memory in advance, or may be stored on a removable storage medium such as a DVD or CD-ROM, and installed by inserting the storage medium into a drive device.

The control unit 61 controls the multiple heater devices 11 based on the detection results of the detection unit 20. The control unit 61 selects an optimal control method for the multiple heater devices 11 based on machine learning using, for example, a neural network. The machine learning is first performed using a single heating well 10.

Machine learning uses, for example, supervised learning, which performs control based on training data and changes parameters used in calculations based on detection results. In supervised learning, for example, a numerical simulation is performed in advance for all heater devices 11 to generate initial training data. The numerical simulation, for example, causes multiple heater elements R1-R5 in the heater device 11 to generate heat based on different control patterns, and analyzes the state of temperature rise in the ground G. In the simulation, a thermal conductivity is set for each depth region R in the ground G. The thermal conductivity is set according to the properties of the ground G, such as the type of stratum, the presence or absence of groundwater, and the flow rate of the groundwater.

In the simulation, for example, a temperature rise in depth region R of ground G is predicted based on a combination of the on states of one or more switches selected from all switches S1-S10. In the simulation, the temperature at each depth of ground G is analyzed until it reaches a predetermined temperature (e.g., a rise of 5°C per hour) in a predetermined time. Based on the analysis results in the simulation, a control pattern for switches S1-S10 that consumes the least amount of energy to raise ground G to the predetermined temperature is selected as initial training data.

The control unit 61 starts heating in the heating well 10 with a control pattern that energizes the initial heater elements R1-R5 based on the selected initial teacher data. The control unit 61 repeats machine learning to change the control pattern of the heater elements R1-R5 until the energy consumption until the detection value for each depth of the ground G detected by the temperature sensor 21 installed immediately adjacent to the heater device 11 reaches the target temperature is minimized. The target temperature is set based on the properties of the ground G and the type of target material contained in the ground G.

The target substances contained in the ground G each have a different boiling point. Therefore, it is desirable to heat the ground G to a temperature higher than the highest boiling point of each target substance in order to volatilize all of the target substances. Therefore, the target temperature is set, for example, to match the highest boiling point of the target substances contained in the ground G. The target temperature may be set individually according to the depth region R in the depth direction of the ground G depending on the distribution state of the contaminant in the stratum.

The control unit 61 stores control data related to the control pattern and data on the detection results corresponding to the control data in the memory unit 62. For example, the control unit 61 changes the method of supplying electricity to the heater device 11 from the initial control pattern and controls it based on a control pattern with different power. If the energy consumption required for the temperature at each depth of the ground G to reach a predetermined temperature in a predetermined time has decreased based on the detection results of the detection unit 20, the control unit 61 selects the changed control pattern as the control pattern.

When the energy consumption required for the temperature at each depth of the ground G to reach a predetermined temperature in a predetermined time increases based on the detection results of the detection unit 20, the control unit 61 changes the current control pattern and performs control based on a different control pattern until the power decreases. For example, based on the selected control pattern, the control unit 61 applies a different current pattern to the switching unit 15 to apply current to the heater unit 13, thereby causing different temperature changes in the depth direction.

The control unit 61, for example, energizes the switching unit 15 with different current patterns based on the selected control pattern to energize the heater unit 13 and generate different temperature changes in the depth direction. By repeating the above process, the control unit 61 uses the current pattern for the switching unit 15 as teacher data and machine-learns the temperature change pattern of the ground G corresponding to the control pattern in the ground G detected by the detection unit 20. Based on the detection results of the detection unit 20, the control unit 61 learns each temperature change pattern in multiple depth regions along the depth direction of the ground G for the current pattern. Based on the learning results, the control unit 61 selects a control pattern that minimizes the energy consumption until the temperature at each depth of the ground G reaches a predetermined temperature in a predetermined time.

Based on the learning results, the control unit 61 controls the heater device 11 with the minimum power supplied to heat the ground G to the target temperature using the selected control pattern. In repeating the above process, if necessary, heating may be stopped and the ground left to stand for a certain period of time to return the ground temperature to its initial state. If the degree of temperature change in the ground G has changed based on the detection results of the detection unit 20, the control unit 61 may repeat machine learning anew and change the control pattern.

As shown in FIG. 7, data from a temperature sensor 21 installed as close as possible to the heating well 10 (e.g., within 1 m horizontal distance) is used to acquire data for machine learning, which is easier to follow changes. Based on the obtained learning result (correct label), the control unit 61 repeats the same learning (iteration) using the next detection unit 20 horizontally adjacent to the detection unit 20 to be measured (e.g., 1 m horizontal distance).

Typically, the multiple heating wells 10 are installed at intervals of 2 to 5 m. The control unit 61 performs heating learning (epochs) using the temperature sensor 21 three to four times. Based on the learning results, the control unit 61 selects an optimal control pattern to be applied to the switching unit 15 that minimizes the power input to the multiple heater devices 11, and controls the switching unit 15 based on the selected optimal control pattern to heat the multiple heater devices 11. At this time, the control unit 61 controls the switching unit 15 with a current pattern to which the optimal combination of on states of switches S1-S10 selected based on the learning results is applied.

After these several epochs, the "trained model" may be applied individually to other nearby heating wells. Alternatively, a machine learning model may be reconstructed using multiple heating wells 10.

Machine learning using a single heating well 10 is preferably performed at the location where in-situ purification is actually performed. However, it is time-consuming to obtain initial data using various control patterns in-situ. In the stage of setting the initial teacher data for the combination pattern of heater elements R1-R5, in addition to using the analysis results of numerical analysis by simulation using the above-mentioned computer model, data may also be obtained using an experimental tank that simulates indoor and outdoor ground conditions. The above-mentioned machine learning is an example of supervised learning. The control unit 61 may also perform machine learning using unsupervised learning that does not use initial teacher data.

Next, we will explain the soil purification method using the soil purification device 1.

Figure 8 shows a flow chart of the process executed in the soil purification device 1. Based on a model in which the properties of the ground G have been digitized, the control unit 61 performs a simulation to control the switching unit 15 to operate the heater device 11 and heat the ground G. Based on the analysis results of the simulation, the control unit 61 generates initial data for a control pattern for heating the ground G (step S100).

Based on the initial data of the control pattern, the control unit 61 controls the switching unit 15 to operate the heater device 11 and heat the ground G (step S102). The control unit 61 performs machine learning to change the control pattern based on the detection result of the detection unit 20 so as to reduce power consumption (step S104).

The control unit 61 energizes the switching unit 15 with different current patterns based on the control pattern, thereby energizing the heater unit 13 and causing different temperature changes in the depth direction. The control unit 61 learns the temperature change pattern in the depth direction of the ground G detected by the detection unit 20 based on the different current patterns.

Based on machine learning, the control unit 61 selects a control pattern that minimizes power consumption (step S106). Based on the selected optimal current pattern, the control unit 61 controls the switching unit 15 to operate the multiple heater devices 11 (step S108).

As described above, the soil purification device 1 can efficiently heat the ground by controlling the heating of the multiple heater devices 11 (heating wells) according to the properties of the ground G. The soil purification device 1 has multiple heater elements R1-R5 in the depth direction, and can change the heating method according to the properties of the depth region of the ground G, allowing the ground G to be efficiently heated. The soil purification device 1 can use machine learning to select a control pattern that heats the depth direction of the ground G with minimal power consumption from among complex control patterns that control the multiple heater elements R1-R5, reducing the energy cost for purifying the ground G and significantly shortening the construction period.

Although one embodiment of the present invention has been described above, the present invention is not limited to the above embodiment and can be modified as appropriate without departing from the spirit of the present invention. For example, the soil purification device 1 may be applied to purifying not only the treatment target area A of the ground on a horizontal surface, but also the ground on an inclined surface or a vertical surface. Furthermore, the control device 60 may be configured with a server device connected via a network. The control device 60 may perform machine learning of multiple soil purification devices.

1 soil purification device, 10 heating well, 10H drilling hole, 11 heater device, 12 casing, 13 heater unit, 13T terminal, 15 switching unit, 20 detection unit, 21 temperature sensor, 30 recovery well, 30H drilling hole, 31 casing, 32 piping, 40 exhaust gas treatment device, 50 power supply unit, 60 control device, 61 control unit, 62 memory unit, 70 diffusion prevention cover, A treatment target area, E ground surface, F impermeable layer, G ground, L water level, R depth area, R1-R5 heater element, S1-S10 switch, T specified area, T1-T4 terminal, W water barrier

Claims (6)

A soil purification device that purifies a target area of the ground by heating the target area to volatilize the target substances,
A plurality of heater devices arranged along a depth direction of the processing target area;
A plurality of detectors arranged around the heater device;
A plurality of recovery wells are arranged in the ground along a depth direction and recover the target substance volatilized from the ground heated by the heater device;
a control unit that controls the plurality of heater devices based on a detection result of the detection unit,
The control unit controls the power supplied to the heater device based on different control patterns,
A temperature change pattern of the ground corresponding to the control pattern detected by the detection unit is machine-learned, and the plurality of heater devices are controlled by the optimal control pattern selected based on the learning result .
The heater device includes a heater section having a plurality of heater elements connected in series along a depth direction, and a switching section that energizes one or more of the heater elements selected from the plurality of heater elements,
The detection unit includes a detection unit having a plurality of temperature sensors arranged along a depth direction,
The control unit energizes the switching unit with different current patterns based on the control pattern to energize the heater unit and cause different temperature changes in the depth direction, performs machine learning on the temperature change pattern in the depth direction of the ground detected by the detection unit based on the current pattern, and controls the switching unit with the optimal current pattern selected based on the learning results . The control unit performs machine learning of each of the temperature change patterns in a plurality of depth regions along a depth direction with respect to the current conduction pattern based on the detection result of the detection unit.
The soil purification apparatus according to claim 1 . A soil purification device that purifies a target area of the ground by heating the target area to volatilize the target substances,
A plurality of heater devices arranged along a depth direction of the processing target area;
A plurality of detectors arranged around the heater device;
A plurality of recovery wells are arranged in the ground along a depth direction and recover the target substance volatilized from the ground heated by the heater device;
a control unit that controls the plurality of heater devices based on a detection result of the detection unit,
The control unit controls the power supplied to the heater device based on different control patterns,
A temperature change pattern of the ground corresponding to the control pattern detected by the detection unit is machine-learned, and the plurality of heater devices are controlled by the optimal control pattern selected based on the learning result .
The soil purification apparatus, wherein the control unit controls the heater device with a minimum amount of power supplied to heat the ground to a target temperature based on the learning result. The target temperature is set based on the properties of the ground and the type of the target substance contained in the ground.
The soil purification apparatus according to claim 3 . A soil purification method for purifying a soil treatment target area by heating the treatment target area of the ground to volatilize target substances, comprising the steps of:
A control unit controls a plurality of heater devices arranged along a depth direction of the processing target area so that electric power is supplied to the heater devices based on different control patterns;
Based on the detection results detected by a plurality of detection units arranged around the heater device,
A temperature change pattern of the ground according to the control pattern is learned by machine learning;
controlling the plurality of heater devices in the control unit using the optimal control pattern selected based on a learning result;
The target substance volatilized from the ground heated by the heater device is collected from a plurality of collection wells arranged in the depth direction of the ground ;
The heater device includes a heater section having a plurality of heater elements connected in series along a depth direction, and a switching section that energizes one or more of the heater elements selected from the plurality of heater elements,
The detection unit includes a detection unit having a plurality of temperature sensors arranged along a depth direction,
The control unit applies different current patterns to the switching unit based on the control pattern to apply current to the heater unit and cause different temperature changes in the depth direction, performs machine learning on the temperature change pattern in the depth direction of the ground detected by the detection unit based on the current pattern, and controls the switching unit with the optimal current pattern selected based on the learning results . A soil purification method for purifying a soil treatment target area by heating the treatment target area of the ground to volatilize target substances, comprising the steps of:
A control unit controls a plurality of heater devices arranged along a depth direction of the processing target area so that electric power is supplied to the heater devices based on different control patterns;
Based on the detection results detected by a plurality of detection units arranged around the heater device,
A temperature change pattern of the ground according to the control pattern is learned by machine learning;
controlling the plurality of heater devices in the control unit using the optimal control pattern selected based on a learning result;
The target substance volatilized from the ground heated by the heater device is collected from a plurality of collection wells arranged in the depth direction of the ground ;
The soil remediation method, wherein the control unit controls the heater device with a minimum amount of power supplied to heat the ground to a target temperature based on the learning result.

Images