Indicator value calculation device, indicator value calculation method, and indicator value calculation program

Abstract

We will calculate index values ​​for heat transfer tube corrosion from readily available data. [Solution] The index value calculation device (2) comprises a composition identification unit (201) that identifies the composition of ash adhering to heat transfer tubes that recover waste heat in a boiler installed in the target facility, and an index value calculation unit (202) that uses a learned model (211) to calculate an index value related to the corrosion of the heat transfer tubes from the composition identified by the composition identification unit (201).

Description

Technical Field

[0001] The present invention relates to an index value calculation device that calculates an index value related to corrosion of a heat transfer tube that recovers waste heat in a boiler, and the like.

Background Art

[0002] In a boiler provided in a power generation plant that generates power using waste heat generated by burning fuel, waste heat is recovered by heat transfer tubes such as water tubes and superheater tubes. Such heat transfer tubes are exposed to a high-temperature state for a long period of time and are also affected by corrosion components contained in the fuel, so corrosion tends to progress. Therefore, continuous management of the heat transfer tubes and repair or replacement at appropriate frequencies are necessary so that the corrosion of the heat transfer tubes does not progress to an unacceptable level.

[0003] As a prior art related to the management of heat transfer tubes, for example, the following Patent Document 1 can be cited. Patent Document 1 below discloses a technique for predicting the life of a boiler heat transfer tube from the measured value of the hydrogen chloride concentration in the combustion gas of the boiler.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, the above-mentioned conventional technologies have a problem that it is not easy to obtain data used for prediction. Specifically, in the technique of Patent Document 1, the hydrogen chloride concentration in the combustion gas of the boiler is continuously measured and its integrated value is calculated, and the wall thickness of the boiler heat transfer tube is estimated based on a previously obtained correlation coefficient and the calculated integrated value. Therefore, it is necessary to continuously measure the hydrogen chloride concentration in the combustion gas of the boiler for a certain period of time. Thus, it cannot be said that it is easy to obtain data used for prediction in the technique of Patent Document 1.

[0006] One aspect of the present invention aims to realize an index value calculation device that can calculate index values ​​related to the corrosion of heat transfer tubes from readily available data. [Means for solving the problem]

[0007] To solve the above problems, an index value calculation device according to another aspect of the present invention comprises: a composition identification unit that identifies the composition of ash adhering to heat transfer tubes that recover waste heat in a boiler installed in a target facility; and an index value calculation unit that calculates the index value in the target facility from the composition identified by the composition identification unit, using a trained model constructed by machine learning with the composition of ash adhering to heat transfer tubes of boilers in the target facility or other facilities as an explanatory variable and an index value related to the corrosion of the heat transfer tubes as the objective variable.

[0008] Furthermore, in order to solve the above problems, an index value calculation method according to another aspect of the present invention is an index value calculation method executed by one or more information processing devices, comprising: a composition identification step of identifying the composition of ash adhering to heat transfer tubes that recover waste heat in a boiler provided at a target facility; and an index value calculation step of calculating the index value at the target facility from the composition identified in the composition identification step, using a trained model constructed by machine learning with the composition of ash adhering to heat transfer tubes of boilers at the target facility or other facilities as an explanatory variable and an index value related to the corrosion of the heat transfer tubes as an objective variable. [Effects of the Invention]

[0009] According to one aspect of the present invention, it becomes possible to calculate index values ​​related to the corrosion of heat transfer tubes from readily available data. [Brief explanation of the drawing]

[0010] [Figure 1] This is a block diagram showing an example configuration of an index value calculation device according to Embodiment 1 of the present invention. [Figure 2]This diagram shows the arrangement and configuration of heat transfer tubes in a waste incineration facility. [Figure 3] This diagram shows the steam flow passing through the secondary and tertiary superheaters, and the exhaust gas flow passing around the secondary and tertiary superheaters. [Figure 4] This flowchart shows an example of the process performed by the above-mentioned index value calculation device. [Figure 5] This is a block diagram showing an example configuration of an index value calculation device according to Embodiment 2 of the present invention. [Figure 6] This flowchart shows an example of the process performed by the above-mentioned index value calculation device. [Modes for carrying out the invention]

[0011] [Embodiment 1] (Device configuration) Figure 1 is a block diagram showing an example configuration of the index value calculation device 1 according to this embodiment. The index value calculation device 1 is a device that calculates an index value related to the corrosion of heat transfer tubes used to recover waste heat generated when fuel is burned and transmit it to a boiler.

[0012] The index value calculation device 1 can also calculate index values ​​related to the corrosion of heat transfer tubes in a waste incineration facility equipped with a boiler for power generation, for example. In this case, the fuel would be waste. The index value calculation device 1 can also calculate index values ​​related to the corrosion of heat transfer tubes in a power generation facility that burns biomass fuel to generate electricity, for example, in this case the fuel would be biomass fuel.

[0013] As shown in the figure, the index value calculation device 1 includes a control unit 10 that controls all parts of the index value calculation device 1, and a storage unit 11 that stores various data used by the index value calculation device 1. The index value calculation device 1 also includes a communication unit 12 for the index value calculation device 1 to communicate with other devices, an input unit 13 that receives input of various data to the index value calculation device 1, and an output unit 14 for the index value calculation device 1 to output various data.

[0014] The control unit 10 also includes a content identification unit 101, a temperature calculation unit 102, and an index value calculation unit 103. The storage unit 11 stores operating data 111, a temperature calculation formula 112, and a learned model 113.

[0015] The content identification unit 101 identifies the sulfur and chlorine content of the boiler fuel in the target facility for which index values ​​related to heat transfer tube corrosion are to be calculated. If the boiler fuel is garbage, i.e., waste, the content identification unit 101 identifies the sulfur and chlorine content of the waste to be incinerated. For example, the sulfur content may be the weight percentage concentration of sulfur contained in the waste to be incinerated (the value obtained by dividing the weight of the sulfur component contained in the waste by the dry weight obtained by subtracting the weight of water from the weight of the waste and converting it to a percentage). The type of index value used to indicate the content is arbitrary. For example, a ratio to a predetermined standard value may be used as the index value to indicate the content. The same applies to the chlorine content. The sulfur and chlorine content may be input via the input unit 13, or obtained from other devices via communication via the communication unit 12.

[0016] The temperature calculation unit 102 calculates the exhaust gas temperature, which is the temperature of the exhaust gas emitted during fuel combustion, and the tube wall temperature of the heat transfer tubes by performing thermal calculations related to the heat transfer tubes. The methods for calculating the exhaust gas temperature and tube wall temperature will be explained later in the sections "Method for Calculating Exhaust Gas Temperature" and "Method for Calculating Tube Wall Temperature".

[0017] The index value calculation unit 103 calculates an index value related to the corrosion of the heat transfer tube. More specifically, the index value calculation unit 103 uses a learned model 113 to calculate the index value from the sulfur content and chlorine content identified by the content identification unit 101, and the exhaust gas temperature and pipe wall temperature calculated by the temperature calculation unit 102. Note that the use of exhaust gas temperature and pipe wall temperature is not mandatory.

[0018] The index value calculated by the index value calculation unit 103 can be any value relating to the corrosion of the heat transfer tube. For example, the index value calculation unit 103 may calculate an index value that indicates the rate of wall thinning of the heat transfer tube, i.e., how much the thickness of the heat transfer tube will decrease over a predetermined period, as an index value to predict the degree of corrosion progression of the heat transfer tube. Alternatively, for example, the index value calculation unit 103 may calculate an index value indicating the remaining lifespan of the heat transfer tube based on the calculated wall thinning rate, the current wall thickness of the heat transfer tube, and the minimum wall thickness of the heat transfer tube. Similarly, the index value calculation unit 103 can also calculate predicted values ​​such as the amount of wall thinning over a predetermined period or the wall thickness of the heat transfer tube after a predetermined period as index values. Note that a trained model 113 may be constructed with the remaining lifespan of the heat transfer tube as the target variable, in which case the index value calculation unit 103 can calculate the remaining lifespan of the heat transfer tube without going through the calculations described above.

[0019] Operating data 111 is the operating data for the target facility used to calculate the index value, and is used to calculate the exhaust gas temperature and pipe wall temperature. Temperature calculation formula 112 is a mathematical formula for calculating the exhaust gas temperature and pipe wall temperature. Details of temperature calculation formula 112 and operating data 111 will be explained later in the sections "Method for Calculating Exhaust Gas Temperature" and "Method for Calculating Pipe Wall Temperature".

[0020] The trained model 113 is a model for calculating index values ​​related to the corrosion of heat transfer tubes. More specifically, the trained model 113 is a trained model constructed by machine learning with the sulfur and chlorine content of the fuel as explanatory variables and the index value related to the corrosion of heat transfer tubes that recover the waste heat generated when the fuel is burned as the dependent variable. The machine learning may be performed using data acquired at the target facility for which the index value is to be calculated, or it may be performed using data acquired at other facilities. Details of the trained model will be explained in the section "About the Trained Model" below.

[0021] As described above, the index value calculation device 1 comprises a content identification unit 101 that identifies the sulfur content and chlorine content of the boiler fuel, and an index value calculation unit 103 that calculates an index value from the sulfur content and chlorine content identified by the content identification unit 101, using a trained model 113 constructed by machine learning with the sulfur content and chlorine content of the fuel as explanatory variables and an index value related to the corrosion of heat transfer tubes that recover the waste heat generated when the fuel is burned as the objective variable.

[0022] Since the sulfur and chlorine content of fuel is known to have a significant impact on the corrosion of heat transfer tubes, the above configuration allows for the calculation of index values ​​related to heat transfer tube corrosion. Furthermore, the sulfur and chlorine content of fuel can be determined without operating the target facility for which the index values ​​are to be calculated, and the data is relatively easy to obtain. Therefore, the above configuration has the effect of allowing the calculation of index values ​​related to heat transfer tube corrosion from relatively easily obtainable data. For example, with the above configuration, it becomes possible to calculate index values ​​related to heat transfer tube corrosion during the design phase of the facility for which the index values ​​are to be calculated, and to reflect those index values ​​in the design.

[0023] (Regarding pre-trained models) As described above, the trained model 113 is a trained model constructed by machine learning with the sulfur content and chlorine content of the fuel as explanatory variables and an index value related to the corrosion of the heat transfer tube that recovers the waste heat generated when the fuel is burned as the dependent variable.

[0024] The explanatory variables of the trained model 113 only need to include at least the sulfur content and chlorine content of the fuel, and it is also possible to include various other information related to the corrosion of heat transfer tubes as explanatory variables. In this embodiment, in addition to the sulfur content and chlorine content of the fuel, an example is described in which the exhaust gas temperature (more precisely, the exhaust gas temperature around the heat transfer tubes), which is the temperature of the exhaust gas emitted when the fuel is burned, and the tube wall temperature of the heat transfer tubes are also used as explanatory variables. Since it is known that exhaust gas temperature and tube wall temperature also have a significant impact on the corrosion of heat transfer tubes, the above configuration makes it possible to calculate highly accurate index values ​​regarding the corrosion of heat transfer tubes.

[0025] Furthermore, as described above, the index value calculation device 1 may include a temperature calculation unit 102 that calculates the exhaust gas temperature and pipe wall temperature by thermal calculations relating to the heat transfer tubes. In this case, the index value calculation unit 103 uses a trained model 113 in which the exhaust gas temperature and pipe wall temperature are included as explanatory variables to calculate the index value from the exhaust gas temperature and pipe wall temperature calculated by the temperature calculation unit 102 and the sulfur content and chlorine content identified by the content identification unit 101.

[0026] Normally, exhaust gas temperature and pipe wall temperature need to be measured by operating the target facility for which the index values ​​are to be calculated. However, with the above configuration, the exhaust gas temperature and the pipe wall temperature of the heat transfer tubes are calculated by thermal calculations related to the heat transfer tubes. Therefore, with the above configuration, index values ​​related to the corrosion of heat transfer tubes can be calculated without operating the facility for which the index values ​​are to be calculated.

[0027] Of course, the exhaust gas temperature and pipe wall temperature do not necessarily have to be calculated by the index value calculation device 1. The index value may be calculated using the exhaust gas temperature and pipe wall temperature input by the user of the index value calculation device 1 via the input unit 13, or obtained from other devices via communication via the communication unit 12. In this case, the temperature calculation unit 102 may be omitted.

[0028] The trained model 113 can be constructed by machine learning using training data that shows the relationship between explanatory variables and the target variable. The training data may be a correspondence between the sulfur and chlorine content of the fuel used at the target facility for which the index value is to be calculated, the exhaust gas temperature and pipe wall temperature calculated from (or collected at) the operating data collected at the facility, and the index value at the facility (e.g., wall thinning rate). In this way, by using data collected at the facility for which the index value is to be calculated as training data, the accuracy of calculating the index value at the facility can be improved.

[0029] Furthermore, the values ​​of the explanatory variables, such as sulfur content, and the target variable, the index value, of the trained model 113 generally show common trends across typical power generation facilities. Therefore, even if the trained model 113 is machine-trained using data collected at a specific facility as training data, it is not limited to use at that facility, but becomes a general-purpose model that can be used to calculate index values ​​at other facilities as well.

[0030] In other words, the training data for the trained model 113 may be data acquired at the target facility for which the index value is to be calculated, or it may be data acquired at another facility. However, in either case, if the operating conditions of the facility affect the corrosion of the heat transfer tubes, it is preferable to use training data acquired under operating conditions equivalent to those of the target facility, at least in terms of such conditions. For example, if it is decided that the target facility will be operated at a steam temperature of 400 degrees, it is preferable to use a trained model 113 constructed using training data acquired under conditions of a steam temperature of 400 degrees or close to it.

[0031] The learning algorithm for the pre-trained model 113 is not particularly limited, as long as it can derive the target variable from the explanatory variables as described above. For example, learning algorithms such as neural networks or support vector machines may be applied. Alternatively, a feedforward neural network such as ELM (Extreme Learning Machine) may be applied. The inventors' experiments have confirmed that by applying ELM as the pre-trained model 113, it is possible to calculate index values ​​with sufficient accuracy for practical use using relatively little training data.

[0032] The index value calculation device 1 may also generate the trained model 113 and the training data used for its machine learning, as described above. In this case, a training data generation unit for generating training data and a training unit for generating the trained model 113 can be added to the control unit 10 in Figure 1.

[0033] (Regarding the arrangement and configuration of heat transfer tubes) The index value calculation device 1 can also calculate index values ​​related to the corrosion of heat transfer tubes in a waste incineration facility equipped with a boiler for power generation. Below, the arrangement and configuration of heat transfer tubes in the waste incineration facility 5 will be described based on Figure 2. Figure 2 is a diagram showing the arrangement and configuration of heat transfer tubes in the waste incineration facility 5.

[0034] In the waste incineration facility 5 shown in Figure 2, waste is burned in a combustion chamber 51. The high-temperature exhaust gas generated by the combustion of waste flows into a flow channel 52 located above the combustion chamber 51. The walls of the combustion chamber 51 and the flow channel 52 are made of water pipe walls 53. As partially enlarged in Figure 2, the water pipe walls 53 are constructed by placing water pipes 531 inside an insulating wall 532. Water circulates inside the water pipes 531.

[0035] Furthermore, the flow path 52 is equipped with an exhaust gas economizer 54 for preheating the boiler feedwater, a steam drum 55 for evaporating the circulating water inside the water tube 531 into steam, a primary superheater 56, a secondary superheater 57, and a tertiary superheater 58. The primary superheater 56 is composed of a continuous superheater tube 562 from the inlet 561 to the outlet 563, as shown in an enlarged view in Figure 2. Although not shown in the illustration, the secondary superheater 57 and tertiary superheater 58 are also composed of a continuous, meandering superheater tube similar to the primary superheater 56.

[0036] The water pipe 531 is connected to the exhaust gas economizer 54, and the circulating water in the water pipe 531, heated by the exhaust gas, is supplied to the exhaust gas economizer 54. This circulating water is further heated in the exhaust gas economizer 54 and supplied to the steam drum 55, which evaporates this circulating water and converts it into steam.

[0037] The steam generated by the steam drum 55 is supplied to the primary superheater 56 where it is superheated. More specifically, the steam generated by the steam drum 55 enters the superheater tube 562 from the inlet 561 of the primary superheater 56, is superheated in the superheater tube 562, and is discharged from the outlet 563. Subsequently, the superheated steam is superheated again in the secondary superheater 57, and then further superheated in the tertiary superheater 58 before being supplied to a steam turbine (not shown). The steam turbine rotates using this superheated steam, and electricity is generated using this rotational power.

[0038] Thus, in the waste incineration facility 5, the waste heat generated by burning waste in the combustion chamber 51 is recovered by the water tube 531 and the superheater tubes in the primary superheater 56 to the tertiary superheater 58. In other words, the water tube 531 and the superheater tubes in the primary superheater 56 to the tertiary superheater 58 are heat transfer tubes that recover waste heat. The index value calculation device 1 can calculate index values ​​related to the corrosion of such water tube 531 and superheater tubes.

[0039] (Method for calculating exhaust gas temperature) The method for calculating the exhaust gas temperature by the temperature calculation unit 102 will be described based on FIG. 3. FIG. 3 is a diagram showing the flow of steam passing through the inside of the secondary superheater 57A and the tertiary superheater 58A, and the flow of exhaust gas passing around the secondary superheater 57A and the tertiary superheater 58A. Although not shown in the figure, a primary superheater is arranged on the downstream side in the exhaust gas flow direction with respect to the secondary superheater 57A.

[0040] The tertiary superheater 58A shown in the figure is provided adjacent to the secondary superheater 57A. As shown by the dashed arrow in FIG. 3, the steam that enters the secondary superheater 57A passes through the inside of the secondary superheater 57A, then enters the tertiary superheater 58A, passes through the inside of the tertiary superheater 58A, and is discharged from the tertiary superheater 58A. On the other hand, the exhaust gas flows from the tertiary superheater 58 towards the secondary superheater 57A.

[0041] In FIG. 3, the exhaust gas temperature at the inlet of the secondary superheater 57A is represented as T g,2,in (°C), and the exhaust gas temperature at the outlet of the secondary superheater 57A is represented as T g,2,out (°C). Similarly, the exhaust gas temperature at the inlet of the tertiary superheater 58A is represented as T g,3,in (°C), and the exhaust gas temperature at the outlet of the tertiary superheater 58A is represented as T g,3,out (°C).

[0042] Among the above temperatures, the exhaust gas temperature T g,3,in at the inlet of the tertiary superheater 58A can be measured by providing a thermometer near the inlet of the tertiary superheater 58A. Also, the exhaust gas temperature T g,3,out at the outlet of the tertiary superheater 58A can be expressed as in the following formula (1) using the specific heat c (kJ / m 3 N°C) of the exhaust gas and the enthalpy h g,3,out (kJ / m 3 N) of the exhaust gas at the outlet of the tertiary superheater 58A. Note that the specific heat c is a constant.

Equation

number

number

number

number

[0043] Furthermore, the temperature calculation unit 102 uses the measured values ​​of steam pressure and steam temperature in the secondary superheater 57A to calculate the steam enthalpy h v,2,out h v,2,in Calculate these values ​​and the h calculated as described above g,3,out By substituting the measured values ​​of exhaust gas flow rate F and steam flow rate V into the above formula (5), h g,2,out The temperature calculation unit 102 can calculate the h g,2,out By substituting this into the above formula (2), the exhaust gas temperature T g,2,out It is possible to calculate this.

[0044] The formula used in the above calculation can be stored in the memory unit 11 as the temperature calculation formula 112. The measured values ​​used in the above calculation can be input as operation data 111.

[0045] (Method for calculating pipe wall temperature) Next, we will explain how to calculate the tube wall temperature. The tube wall temperature of the superheater tube is Tm (°C), the exhaust gas temperature around the superheater tube is T g If we denote the temperature as (°C), then the pipe wall temperature T m is, exhaust gas temperature T g The heat transfer rate Q (kcal / h), the outer diameter r2 of the superheater tube, and the gas heat transfer coefficient α are all considered. g (kcal / m 2 It can be expressed using the temperature (h°C) and the length L (m) of the superheater tube as shown in the following formula (6).

number

number

number

number

[0046] For example, the tube wall temperature T near the outlet of the tertiary superheater 58A. m,3,out When calculating the exhaust gas temperature T near the outlet of the tertiary superheater 58A, the temperature calculation unit 102 calculates the exhaust gas temperature T g,3,out The heat transfer coefficient K and the fluid (i.e., steam) temperature T near the outlet of the tertiary superheater 58A. f,3,out Substitute this into the above formula (8). Formula (8) can be stored in the memory unit 11 as the temperature calculation formula 112, and T f,3,out This should be entered as driving data 111.

[0047] (Process flow) The processing flow (indicator value calculation method) performed by the indicator value calculation device 1 will be explained based on Figure 4. Figure 4 is a flowchart showing an example of the processing performed by the indicator value calculation device 1. In the following, an example of calculating the indicator value related to the corrosion of the superheater tubes of the tertiary superheater 58A shown in Figure 3 will be explained.

[0048] In S11, the temperature calculation unit 102 receives the input of operation data 111. The operation data 111 may be input via the input unit 13, or it may be input from another device via communication through the communication unit 12. Specifically, the operation data 111 includes various measured values ​​used to calculate the exhaust gas temperature and pipe wall temperature, as described in the "Method for Calculating Exhaust Gas Temperature" above.

[0049] In S12, the temperature calculation unit 102 uses the operating data 111 input in S11 and the temperature calculation formula 112 to calculate the exhaust gas temperature near the tertiary superheater 58A and the tube wall temperature of the superheater tubes of the tertiary superheater 58A. For example, as explained in the "Method for Calculating Exhaust Gas Temperature" above, the temperature calculation unit 102 uses formulas (1) to (5) and the operating data 111 to calculate the exhaust gas temperature T at the outlet of the tertiary superheater 58A. g,3,out The temperature calculation unit 102 then calculates the exhaust gas temperature T as described in the "Method for Calculating Pipe Wall Temperature" above. g,3,out And using the operating data 111 and formula (8), the tube wall temperature T of the superheater tube of the tertiary superheater 58A is calculated. mYou may calculate this.

[0050] In S13, the content identification unit 101 identifies the sulfur and chlorine content of the boiler fuel at the target facility. For example, if the boiler fuel is incinerated waste, S13 identifies the sulfur and chlorine content of the waste. The sulfur and chlorine content may be input via the input unit 13, or obtained from another device via communication through the communication unit 12. Furthermore, the sulfur and chlorine content identified in S13 may be determined by actually analyzing the fuel (e.g., waste), or they may be estimated values ​​calculated based on the general composition of the fuel.

[0051] In S14, the index value calculation unit 103 uses the learned model 113 to calculate an index value for corrosion of the superheater tubes of the tertiary superheater 58A from the exhaust gas temperature and tube wall temperature calculated in S12, and the sulfur content and chlorine content identified in S13. Specifically, the index value calculation unit 103 inputs the exhaust gas temperature, tube wall temperature, sulfur content, and chlorine content into the learned model 113, and the learned model 113 outputs the index value. This completes the process shown in Figure 4. The index value calculation unit 103 may store the calculated index value in the storage unit 11, output it to the output unit 14, or transmit it to another device via the communication unit 12.

[0052] As described above, the index value calculation method according to this embodiment includes a content identification step (S13) in which the sulfur content and chlorine content of the boiler fuel are identified, and an index value calculation step (S14) in which an index value is calculated from the sulfur content and chlorine content identified in S13 using a trained model 113 constructed by machine learning with the sulfur content and chlorine content of the fuel as explanatory variables and an index value relating to the corrosion of heat transfer tubes that recover the waste heat generated when the fuel is burned as the objective variable. Thus, according to the index value calculation method according to this embodiment, an index value relating to the corrosion of heat transfer tubes can be calculated from readily available data such as the sulfur content and chlorine content of the fuel.

[0053] [Embodiment 2] Other embodiments of the present invention are described below. For the sake of clarity, components having the same function as those described in the above embodiments will be denoted by the same reference numerals, and their descriptions will not be repeated.

[0054] (Device configuration) The configuration of the index value calculation device 2 according to this embodiment will be described with reference to Figure 5. Figure 5 is a block diagram showing an example of the configuration of the index value calculation device 2. Similar to the index value calculation device 1 of Embodiment 1, the index value calculation device 2 is a device that calculates an index value related to the corrosion of heat transfer tubes used to recover waste heat generated when fuel is burned and transmit it to a boiler.

[0055] The index value calculation device 2 differs from the index value calculation device 1 of Embodiment 1 in that the control unit 10 and storage unit 11 are replaced by a control unit 20 and a storage unit 21, respectively. The control unit 20 includes a composition identification unit 201 and an index value calculation unit 202. The storage unit 21 stores a trained model 211.

[0056] The composition identification unit 201 identifies the composition of the ash adhering to the heat transfer tubes that recover waste heat in the boiler. For example, the composition identification unit 201 may identify the weight percentage concentration of each component contained in the ash, which is determined by component analysis of the ash adhering to the heat transfer tubes, as the composition of the ash. Details of the composition identified by the composition identification unit 201 will be explained in the section "About the trained model" below.

[0057] The ash composition may be input via the input unit 13, or obtained from another device via communication through the communication unit 12. The ash composition may be determined by analyzing the ash actually adhering to the heat transfer tubes as described above, or it may be a design assumption value (a value that is assumed to be the composition based on the design of the facility).

[0058] The index value calculation unit 202 calculates index values ​​related to the corrosion of the heat transfer tubes. More specifically, the index value calculation unit 202 uses a trained model 211 to calculate index values ​​from the composition of the ash identified by the composition identification unit 201. The index values ​​to be calculated can be related to the corrosion of the heat transfer tubes, as in Embodiment 1. For example, the index value calculation unit 202 may calculate index values ​​indicating the thinning rate or the remaining lifespan of the heat transfer tubes.

[0059] The trained model 211, like the trained model 113 in Embodiment 1, is a model for calculating index values ​​related to heat transfer tube corrosion. The explanatory variable of the trained model 211 is the composition of ash adhering to the heat transfer tubes of a boiler, which is different from the trained model 113. The trained model 211 may be machine-trained using data acquired at the target facility for which the index values ​​are to be calculated, or it may be machine-trained using data acquired at other facilities, and in this respect it is the same as the trained model 113.

[0060] As described above, the index value calculation device 2 comprises a composition identification unit 201 that identifies the composition of ash adhering to heat transfer tubes that recover waste heat in a boiler installed at the target facility, and an index value calculation unit 202 that uses a trained model 211 constructed by machine learning with the composition of ash adhering to heat transfer tubes of boilers at the target facility or other facilities as explanatory variables and an index value related to the corrosion of said heat transfer tubes as the objective variable to calculate the index value at the target facility from the composition identified by the composition identification unit 201.

[0061] Since the composition of ash adhering to heat transfer tubes is known to correlate with the corrosion of those tubes, the above configuration allows for the calculation of an index value related to heat transfer tube corrosion. Furthermore, the above configuration has the effect of allowing the calculation of an index value related to heat transfer tube corrosion from relatively easily obtainable data, namely the composition of ash.

[0062] (Regarding pre-trained models) The trained model 211 is a trained model constructed by machine learning, with the composition of ash adhering to the heat transfer tube as the explanatory variable and an index value related to the corrosion of the heat transfer tube as the dependent variable. The explanatory variables of the trained model 211 may represent the composition of all components constituting the ash, or they may represent the composition of some of the components. The composition identification unit 201 identifies the content of the components that are the explanatory variables of the trained model 211.

[0063] For example, the composition identification unit 201 may at least specify the sulfur content and chlorine content of the ash adhering to the heat transfer tube as the composition of the ash. Since the sulfur content and chlorine content of the ash adhering to the heat transfer tube are known to have a strong correlation with the corrosion of the heat transfer tube, the above configuration makes it possible to calculate highly accurate index values ​​regarding the corrosion of the heat transfer tube.

[0064] Furthermore, for example, the composition identification unit 201 may at least specify the alkali metal content in the ash adhering to the heat transfer tube as part of the composition of the ash. Since the alkali metal content in the ash adhering to the heat transfer tube is known to correlate with the corrosion of the heat transfer tube, the above configuration makes it possible to calculate highly accurate index values ​​regarding the corrosion of the heat transfer tube. Examples of alkali metals contained in the ash adhering to the heat transfer tube include sodium and potassium.

[0065] Furthermore, for example, the composition identification unit 201 may at least specify the heavy metal content in the ash adhering to the heat transfer tube as part of the composition of the ash. Since the heavy metal content in the ash adhering to the heat transfer tube is known to correlate with the corrosion of the heat transfer tube, the above configuration makes it possible to calculate highly accurate index values ​​regarding the corrosion of the heat transfer tube. Examples of heavy metals contained in the ash adhering to the heat transfer tube include copper, zinc, and lead.

[0066] Alternatively, a trained model 211 may be used, in which the composition of all major components contained in the ash adhering to the heat transfer tubes is used as an explanatory variable. In this case, the composition identification unit 201 may identify the content of oxygen, magnesium, aluminum, silicon, phosphorus, and calcium, in addition to the sulfur, chlorine, sodium, potassium, copper, zinc, and lead mentioned above.

[0067] Similar to the trained model 113 in Embodiment 1, the trained model 211 can be constructed by machine learning using training data that shows the relationship between explanatory variables and the target variable. The training data may be a correspondence between the composition of ash collected from the facility to which the index value is to be calculated or from other facilities, and the index value at said facility (e.g., thinning rate). Also, similar to the trained model 113, the algorithm of the trained model 211 is not particularly limited, as long as it can derive the target variable from the explanatory variables as described above.

[0068] Furthermore, the index value calculation device 2 may also generate the trained model 211 and the training data used for its machine learning, as described above. In this case, a training data generation unit for generating training data and a training unit for generating the trained model 211 can be added to the control unit 20 in Figure 5.

[0069] (Process flow) The processing flow (indicator value calculation method) performed by the indicator value calculation device 2 will be explained based on Figure 6. Figure 6 is a flowchart showing an example of the processing performed by the indicator value calculation device 2. In the following, an example of calculating the indicator value for corrosion of the superheater tube 562 in the waste incineration facility 5 shown in Figure 2 will be explained.

[0070] In S21, the composition identification unit 201 identifies the composition of the ash adhering to the superheater tube 562. The ash composition may be input via the input unit 13, or it may be input from another device via communication through the communication unit 12. As described above, the ash composition identified in S21 may be determined by analyzing the ash actually adhering to the superheater tube 562, or it may be a design assumption value.

[0071] In S22, the index value calculation unit 202 uses the trained model 211 to calculate an index value related to the corrosion of the superheater tube 562 from the ash composition identified in S21. Specifically, the index value calculation unit 202 inputs the ash composition, i.e., the content of each component contained in the ash, into the trained model 211, and the trained model 211 outputs an index value. This completes the process shown in Figure 6. The index value calculation unit 202 may store the calculated index value in the storage unit 21, output it to the output unit 14, or transmit it to another device via the communication unit 12.

[0072] As described above, the index value calculation method according to this embodiment includes a composition identification step (S21) in which the composition of ash adhering to heat transfer tubes that recover waste heat in a boiler installed at the target facility is identified, and an index value calculation step (S22) in which an index value at the target facility is calculated from the composition identified in S21 using a trained model 211 constructed by machine learning with the composition of ash adhering to heat transfer tubes of boilers at the target facility or other facilities as an explanatory variable and an index value related to the corrosion of said heat transfer tubes as the objective variable. This makes it possible to calculate an index value related to the corrosion of heat transfer tubes from readily available data such as the composition of ash.

[0073] [Variation] The entities executing each process described in the above embodiments are arbitrary and are not limited to the examples above. For example, the index value calculation method shown in Figure 4 may be configured so that multiple information processing devices execute it instead of the index value calculation device 1. For example, the processes S11 and S12 in Figure 4 may be executed by one information processing device, and the exhaust gas temperature and pipe wall temperature calculated by that information processing device may be output to other information processing devices. In this case, the other information processing devices may perform the processes S13 and S14. The same applies to the index value calculation method shown in Figure 6, which may be configured so that multiple information processing devices execute it instead of the index value calculation device 2.

[0074] [Examples of implementation using software] The functions of the index value calculation devices 1 and 2 (hereinafter referred to as "devices") can be realized by a program that causes a computer to function as the device, and by a program that causes a computer to function as each control block of the device (particularly each part included in the control units 10 and 20) (index value calculation program).

[0075] In this case, the device includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., memory) as hardware for executing the program. By executing the program using this control device and storage device, the functions described in each of the embodiments are realized.

[0076] The above program may be recorded on one or more computer-readable recording media, not temporary ones. These recording media may or may not be provided by the above device. In the latter case, the program may be supplied to the above device via any wired or wireless transmission medium.

[0077] Furthermore, some or all of the functions of each of the above control blocks can also be realized by logic circuits. For example, an integrated circuit in which logic circuits functioning as each of the above control blocks are formed is also included in the scope of the present invention. In addition, it is also possible to realize the functions of each of the above control blocks by, for example, a quantum computer.

[0078] The present invention is not limited to the embodiments described above, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention.

[0079] (Additional notes) To solve the above problems, an index value calculation device according to one aspect of the present invention comprises: a content identification unit that identifies the sulfur content and chlorine content of a fuel; and an index value calculation unit that calculates the index value from the sulfur content and chlorine content identified by the content identification unit, using a trained model constructed by machine learning with the sulfur content and chlorine content of the fuel as explanatory variables and an index value relating to the corrosion of a heat transfer tube that recovers the waste heat generated when the fuel is burned as the objective variable.

[0080] To solve the above problems, an index value calculation method according to one aspect of the present invention is an index value calculation method executed by one or more information processing devices, comprising: a content identification step of identifying the sulfur content and chlorine content of boiler fuel; and an index value calculation step of calculating the index value from the sulfur content and chlorine content identified in the content identification step, using a trained model constructed by machine learning with the sulfur content and chlorine content of the fuel as explanatory variables and an index value relating to the corrosion of heat transfer tubes that recover waste heat generated when the fuel is burned as the objective variable. [Explanation of Symbols]

[0081] 1. Indicator Value Calculation Device 101 Content Specification Department 102 Temperature calculation section 103 Indicator Value Calculation Unit 113 Pre-trained models 2. Indicator Value Calculation Device 201 Composition Identification Department 202 Indicator Value Calculation Unit 211 Pre-trained models

Claims

1. A composition identification unit that identifies the composition of ash adhering to the heat transfer tubes that recover waste heat in the boiler installed in the target facility, An index value calculation device comprising: an index value calculation unit that calculates the index value at the target facility from the composition identified by the composition identification unit, using a trained model constructed by machine learning with the composition of ash adhering to the heat transfer tubes of a boiler at the target facility or other facilities as an explanatory variable and an index value related to the corrosion of the heat transfer tubes as the dependent variable.

2. The index value calculation device according to claim 1, wherein the composition identification unit identifies at least the sulfur content and chlorine content in the ash adhering to the heat transfer tube as the composition of the ash.

3. The index value calculation device according to claim 1 or 2, wherein the composition identification unit identifies at least the alkali metal content in the ash adhering to the heat transfer tube as the composition of the ash.

4. The index value calculation device according to any one of claims 1 to 3, wherein the composition identification unit identifies at least the heavy metal content in the ash adhering to the heat transfer tube as the composition of the ash.

5. A method for calculating an index value, which is performed by one or more information processing devices, A composition identification step to identify the composition of ash adhering to the heat transfer tubes that recover waste heat in the boiler installed at the target facility, An index value calculation method comprising: an index value calculation step, which calculates the index value at the target facility from the composition identified in the composition identification step, using a trained model constructed by machine learning with the composition of ash adhering to the heat transfer tubes of a boiler at the target facility or other facilities as an explanatory variable and an index value related to the corrosion of the heat transfer tubes as the dependent variable.

6. An index value calculation program for causing a computer to function as an index value calculation device according to claim 1, wherein the composition identification unit and the index value calculation unit are the components of the computer.

Images