Property prediction device, property prediction program, and analysis system

Abstract

This invention provides a property prediction device, a property prediction program, and an analysis system that can improve the accuracy of predicting the properties of raw material oil. [Solution] The property prediction device (14) includes a data acquisition unit 42 that acquires information about the components contained in the raw oil (12) and measurement data (58, 62) which is a collection of measured values ​​corresponding to the information about the components; a feature generation unit 46 that generates a feature set which is a collection of feature quantities related to the measured values ​​from the acquired measurement data (58, 62); and a property prediction unit 50 that predicts the properties of the raw oil (12). The property prediction unit 50 includes a raw oil prediction model (PM) that, upon input of the generated feature set, outputs the yield of fractions that can be obtained from the raw oil (12) according to a predetermined temperature zone.

Description

Technical Field

[0001] The present disclosure relates to a property prediction device, a property prediction program, and an analysis system.

Background Art

[0002] Conventionally, various analysis techniques for analyzing properties of raw material oil or its fractions have been known. For example, a method has been proposed in which measurement data obtained through an analyzer is subjected to analysis processing to predict the properties of raw material oil or its fractions.

[0003] Patent Document 1 discloses a method for predicting the properties of crude oil or its fractions using gas chromatography and a mass spectrometer.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] By the way, fractions obtained when raw material oil is fractionally distilled (fractionated) by several temperature sections may be used as products or base materials. In this case, for example, the properties of raw material oil regarding the amount of fractions obtained for each specific temperature section, such as the yield (amount distilled) in the fraction with a temperature section of 0 - 80°C, are important from the viewpoint of predicting the production amount of products and base materials. Thus, improvement in the prediction accuracy regarding the properties of raw material oil has been continuously desired.

[0006] The present disclosure has been made in view of the above circumstances, and an object thereof is to provide a property prediction device, a property prediction program, and an analysis system capable of improving the prediction accuracy of the properties of raw material oil.

Means for Solving the Problems

[0007] To solve the above problems, a property prediction device in one aspect of the present disclosure includes a data acquisition unit that acquires measurement data which is a collection of information about components contained in raw oil and measured values ​​corresponding to the information about components, and a feature generation unit that generates a feature set which is a collection of features related to the measured values ​​from the measurement data acquired by the data acquisition unit. The system includes a properties prediction unit that predicts the properties of the raw material oil, the properties prediction unit includes a raw material oil prediction model that, upon input of the set of features generated by the feature generation unit, outputs the yield of fractions obtainable from the raw material oil according to predetermined temperature zones, there are multiple temperature zones, and the number of raw material oil prediction models that output the yield is equal to the number of temperature zones.

[0008] A property prediction program in another aspect of the present disclosure causes one or more computers to perform an acquisition process to acquire measurement data which is a collection of measurement values ​​corresponding to information about components contained in the raw oil and information about the components; a generation process which generates a feature set which is a collection of feature values ​​relating to the measurement values ​​from the measurement data acquired through the acquisition process; and a property prediction process which predicts the properties of the raw oil. The property prediction process is performed using a raw oil prediction model that, upon input of the feature set generated through the generation process, outputs the yield of fractions that can be obtained from the raw oil according to predetermined temperature zones. There are multiple temperature zones, and the number of raw oil prediction models that output the yield is equal to the number of temperature zones.

[0009] An analysis system in another aspect of the present disclosure comprises: an analysis device that outputs measurement data which is a collection of measured values ​​corresponding to information about components contained in a raw material oil; and a property prediction device that performs analytical processing on the measurement data output from the analysis device to predict the properties of the raw material oil, wherein the property prediction device performs a generation process to generate a feature set which is a collection of feature quantities relating to the measured values ​​from the measurement data; and a property prediction process to predict the properties of the raw material oil, wherein the property prediction process is performed using a raw material oil prediction model that, upon input of the feature set generated through the generation process, outputs the yield of fractions obtainable from the raw material oil according to predetermined temperature zones, wherein there are multiple temperature zones, and the number of raw material oil prediction models that output the yield is equal to the number of temperature zones. [Effects of the Invention]

[0010] According to this disclosure, the accuracy of predicting the properties of raw material oil can be improved. [Brief explanation of the drawing]

[0011] [Figure 1] This is an overall configuration diagram of the analysis system in one embodiment of the present disclosure. [Figure 2] Figure 1 shows an example of the hardware configuration of the property prediction device. [Figure 3] This is a block diagram showing an example of the functional configuration of the property prediction device in Figure 1. [Figure 4] This figure shows an example of a combination of measurement conditions and predicted values ​​for raw material oil. [Figure 5] This figure shows an example of the model structure of a predictive model used to predict the properties of raw material oil. [Figure 6] Figures 1 to 3 are flowcharts illustrating an example of the learning process performed by the property prediction device. [Figure 7] This figure shows an example of measurement data obtained by high-temperature gas chromatography. [Figure 8] This diagram schematically illustrates one example of a method for formatting a training dataset. [Figure 9] It is a diagram showing an example of a method for generating a prediction model in FIG. 5. [Figure 10] It is a functional block diagram regarding the prediction operation by the control unit in FIG. 3. [Figure 11] It is a flowchart showing an example of a prediction operation by the property prediction device in FIGS. 1 to 3. [Figure 12] It is a diagram showing an example of an analysis result screen displayed on the display unit in FIG. 3.

Mode for Carrying Out the Invention

[0012] First, some aspects of the present disclosure will be described.

[0013] The property prediction device according to the first aspect of the present disclosure includes a data acquisition unit that acquires measurement data, which is an aggregate of information regarding components contained in the feedstock and measurement values corresponding to the information regarding the components, a feature quantity generation unit that generates a feature quantity set, which is an aggregate of feature quantities regarding the measurement values, from the measurement data acquired by the data acquisition unit, and a property prediction unit that predicts the property of the feedstock. When the property prediction unit inputs the feature quantity set generated by the feature quantity generation unit, it includes a feedstock prediction model that outputs the yield of a fraction that can be obtained from the feedstock according to a predetermined temperature section. There are a plurality of the temperature sections, and the number of the feedstock prediction models that output the yield is equal to the number of the temperature sections.

[0014] In the property prediction device according to the second aspect of the present disclosure, the property prediction unit further predicts at least one of the density of the feedstock, the sulfur content of the feedstock, and the nitrogen content of the feedstock, in addition to the yield, as the property of the feedstock, and the feedstock prediction model may be provided for each type of the property of the feedstock predicted by the property prediction unit.

[0015] In the property prediction device according to the third aspect of the present disclosure, the property prediction unit may further include a fraction prediction model that further predicts the property of the fraction and outputs a fraction property value indicating the property of the fraction corresponding to the temperature section when the feature amount set generated by the feature amount generation unit is input.

[0016] In the property prediction device according to the fourth aspect of the present disclosure, the property of the fraction includes at least one of density, sulfur content, and nitrogen content, and the fraction prediction model may be provided with a number corresponding to a combination of the number of the plurality of temperature sections and the number of properties of the fraction predicted by the property prediction unit.

[0017] In the property prediction device according to the fifth aspect of the present disclosure, the measurement data is data measured by a gas chromatograph device, the measured value is a signal intensity, and the information regarding the component may be a retention time.

[0018] In the property prediction device according to the sixth aspect of the present disclosure, the temperature section is determined based on the type of fraction desired by the user for the feedstock, the feedstock is crude oil, and the type may include at least one of LPG, naphtha, gasoline, kerosene, jet fuel, light oil, lubricating oil, heavy oil, residual oil, and asphalt.

[0019] In the property prediction device according to the seventh aspect of the present disclosure, the feature amount generation unit may generate the feature amount set by compressing the number of dimensions of the measurement data.

[0020] In the property prediction device according to the eighth aspect of the present disclosure, the prediction model may be a mathematical model in which learning processing is performed using different learning data sets for each property of the feedstock within the temperature section for a learning device having common input-output characteristics.

[0021] In the property prediction device according to the ninth aspect of the present disclosure, the feature amount generation unit may generate a feature amount set common to two or more properties of the feedstock.

[0022] The property prediction device in the tenth aspect of this disclosure may further include a data processing unit that generates integrated data by associating the collection of measurement data with the collection of property values ​​relating to the raw oil, and divides the integrated data for each property relating to the raw oil within the temperature section to generate the learning dataset.

[0023] The property prediction device in the eleventh aspect of this disclosure may further include a learning processing unit that performs learning processing on a learner having common input / output characteristics using the learning dataset generated by the data processing unit for each property of the raw material oil in the temperature compartment, thereby generating the raw material oil prediction model or the fraction prediction model for each property of the raw material oil in the temperature compartment.

[0024] The property prediction device in the twelfth aspect of this disclosure may further include an output processing unit that instructs an output means to output a property analysis table including the raw oil property values ​​predicted by the property prediction unit.

[0025] A property prediction program in a 13th aspect of this disclosure causes one or more computers to perform an acquisition process to acquire measurement data which is a collection of measurement values ​​corresponding to information about components contained in the raw oil and information about the components; a generation process which generates a feature set which is a collection of feature values ​​related to the measurement values ​​from the measurement data acquired through the acquisition process; and a property prediction process which predicts the properties of the raw oil. The property prediction process is performed using a raw oil prediction model that, upon input of the feature set generated through the generation process, outputs the yield of fractions that can be obtained from the raw oil according to predetermined temperature zones. There are multiple temperature zones, and the number of raw oil prediction models that output the yield is equal to the number of temperature zones.

[0026] The analysis system in the 14th aspect of this disclosure comprises an analysis device that outputs measurement data which is a collection of measured values ​​corresponding to information about components contained in a raw material oil, and a property prediction device that performs analytical processing on the measurement data output from the analysis device and predicts the properties of the raw material oil, wherein the property prediction device performs a generation process to generate a feature set which is a collection of feature quantities related to the measured values ​​from the measurement data, and a property prediction process to predict the properties of the raw material oil, wherein the property prediction process is performed using a raw material oil prediction model that, upon input of the feature set generated through the generation process, outputs the yield of fractions that can be obtained from the raw material oil according to predetermined temperature zones, wherein there are multiple temperature zones, and the number of raw material oil prediction models that output the yield is equal to the number of temperature zones.

[0027] Embodiments of this disclosure will be described below with reference to the accompanying drawings. To facilitate understanding of the description, the same reference numerals are used for identical components and steps in each drawing whenever possible, and redundant descriptions are omitted. Where terms such as “first,” “second,” etc. are used in this specification or claims, unless otherwise specified, they do not indicate any order or importance, but are used to distinguish one configuration from another. The word “or” is interpreted in its broadest sense, i.e., “at least one,” unless otherwise specified. The word “part” may be replaced with other words such as, for example, unit, module, device, or element.

[0028] [Configuration of Analysis System 10] <Overall Structure> Figure 1 is an overall diagram of the analysis system 10 in one embodiment of the present disclosure. The analysis system 10 is provided for, for example, to analyze the raw oil 12 or fractions of the raw oil 12 and to manage the analysis results. The analysis system 10 is provided for, for example, to predict the properties of the raw oil 12 or the properties of fractions of the raw oil 12 and to manage the prediction results. The analysis system 10 predicts the distillation properties of the raw oil 12, for example. In this embodiment, the raw oil 12 is petroleum as extracted from the oil well (i.e., crude oil). The raw oil 12 may be, for example, petroleum products produced from crude oil. Examples of petroleum products include gasoline, jet fuel, kerosene, diesel fuel, and fractions obtained by fractional distillation from crude oil. A fraction is each part obtained when a mixed liquid is fractionally distilled through several temperature zones. A temperature zone is the lower limit or upper limit of the temperature required to obtain a specific fraction by fractional distillation. That is, a temperature zone is the boundary for obtaining a specific fraction from the raw oil 12. Therefore, if the raw material oil 12 contains multiple compounds with different boiling points, by appropriately defining multiple temperature zones, the user can obtain fractions from the raw material oil 12 according to the temperature zone. Here, the term "temperature zone" in this disclosure is not limited to the temperature range for obtaining fractions from the raw material oil 12. For example, the term "temperature zone" in this disclosure includes the case where no distillation operation has been performed on the raw material oil 12 and no temperature zone for obtaining fractions has been defined, as one of the temperature zones. In this disclosure, "temperature zone" may be simply referred to as "zone".

[0029] Examples of indicators that show the properties of the raw material oil 12 (hereinafter referred to as raw material oil property values) include yield (the yield of a specific fraction obtained from the raw material oil 12), density, sulfur content, nitrogen content, kinematic viscosity, calorific value, and various other indicators including component ratios.

[0030] Examples of indicators that show the properties of the fractions of the raw material oil 12 (hereinafter referred to as fraction property values) include various indicators such as density, sulfur content, nitrogen content, kinematic viscosity, calorific value, and component ratio. Here, the properties of the raw material oil 12 include at least one of the properties of the raw material oil 12 and the properties of the fractions of the raw material oil 12. Here, the property values ​​of the raw material oil 12 include at least one of the raw material oil property values ​​and fraction property values.

[0031] The analysis system 10 specifically comprises a property prediction device 14, an analysis device 16, a measurement terminal 18, and a file server 20. The property prediction device 14, the measurement terminal 18, and the file server 20 are configured to communicate with each other via the NT network.

[0032] The property prediction device 14 is a device that predicts the properties of the raw oil 12. The property prediction device 14 is a device that predicts the properties of the raw oil 12 or the properties of the fractions of the raw oil 12. The property prediction device 14 is a computer used by a user (e.g., an analyst) involved in predicting the properties of the raw oil 12 or the properties of the fractions of the raw oil 12. The property prediction device 14 consists of, for example, a stationary device including a personal computer, or a portable device including a tablet, laptop, or smartphone.

[0033] The analyzer 16 is, for example, a device for analyzing the raw material oil 12. The analyzer 16 is, for example, a chromatograph that uses a chromatographic method. Examples of chromatograph methods include gas chromatography, surface liquid chromatography, liquid chromatography, or size exclusion chromatography. In this embodiment, the analyzer 16 is a gas chromatograph that can separate and quantify the components contained in the raw material oil 12 by heating and vaporizing the raw material oil 12 in order to analyze the components containing boiling points contained in the raw material oil 12. When the raw material oil 12 is crude oil, it is preferable that the analyzer 16 is a high-temperature gas chromatograph, which is a type of gas chromatograph. By using a high-temperature gas chromatograph, for example, it is possible to analyze components with boiling points in the temperature range of 500°C to 800°C contained in crude oil, which is an example of the raw material oil 12, and the prediction range of the properties of the raw material oil 12 and the properties of the fractions of the raw material oil 12 can be broadened.

[0034] The measurement terminal 18 is a computer that controls the measurement operations of the analyzer 16. The measurement terminal 18 is a computer used by, for example, a user (e.g., an analyst) involved in the analysis of the raw material oil 12. The measurement terminal 18 acquires the measurement data output from the analyzer 16 and links it with various supplementary information related to the measurement (e.g., identification information of the raw material oil 12 and at least one of the measurement conditions) and supplies the measurement result information 22, including the measurement data, to the file server 20. The measurement data is acquired, for example, by analyzing the raw material oil 12 using the analyzer 16. The measurement data is, for example, a collection of information about the components contained in the raw material oil 12 and the measured values ​​corresponding to the information about the components. The measurement data includes, for example, qualitative information about the components contained in the raw material oil 12 and quantitative information about the components contained in the raw material oil 12.

[0035] If the analyzer 16 is a gas chromatograph, the information regarding the components is, for example, the retention time. If the analyzer 16 is a gas chromatograph, the qualitative information regarding the components contained in the raw oil 12 is, for example, the retention time. The retention time is the time from when the sample (raw oil 12) is injected into the gas chromatograph until each component contained in the raw oil 12 is detected. The retention time is also called, for example, the retention time. Each component contained in the sample injected into the gas chromatograph moves through the column of the chromatograph together with the mobile phase (for example, helium gas), and as a result, differences in the speed at which they move through the column occur, resulting in separation. Therefore, there is a correlation between the retention time and the types of components contained in the raw oil 12. Thus, the types of components contained in the raw oil 12 can be determined based on the retention time.

[0036] If the analyzer 16 is a gas chromatograph, the measured value corresponding to the information about the components is, for example, the signal intensity value. If the analyzer 16 is a gas chromatograph, the quantitative information about the components contained in the raw oil 12 is, for example, the signal intensity value. The detector of the gas chromatograph is, for example, a flame ionization detector (hereinafter referred to as FID). The FID is a detector that can detect organic compounds, which are an example of components, and is commonly used as a detector for gas chromatographs. The FID sequentially detects the components that come out of the column of the gas chromatograph. The signal intensity value (measured value) of the FID correlates with the concentration of the organic compound (component). Therefore, the amount of components contained in the raw oil 12 can be determined based on the signal intensity value corresponding to a certain retention time. Hence, the signal intensity value is an example of quantitative information about the components contained in the raw oil 12.

[0037] As described above, if the analyzer 16 is a gas chromatograph and the measurement data includes retention time and signal intensity values, the measurement data will contain information that allows for the analysis of the types of components contained in the raw oil 12 and the amounts of each component.

[0038] The property data is, for example, data representing the properties of the raw material oil 12. The property data includes, for example, property values ​​for the raw material oil 12. The property values ​​for the raw material oil include, for example, at least one of the raw material oil property values ​​and the fraction property values. The raw material oil property values ​​include, for example, at least one of the yield, density, sulfur content, nitrogen content, kinematic viscosity, calorific value, and component ratio. The fraction property values ​​include, for example, at least one of the density, sulfur content, nitrogen content, kinematic viscosity, calorific value, and component ratio. The property values ​​for the raw material oil included in the property data can be obtained by using a measurement method corresponding to the desired properties for the raw material oil 12 or the fraction of the raw material oil 12. The property data is, for example, data in which each property value is associated with a temperature section, as described later, for each raw material oil 12.

[0039] The yield can be obtained, for example, by following JIS K2254 (2018) "Petroleum Products - Method for Determining Distillation Properties". The yield is, for example, the ratio of the amount of distillate of the fraction in the temperature section of interest to the sum of the residual amount and the total distillate of the raw material oil 12 being measured. The yield may be calculated, for example, based on volume fraction % or based on mass fraction %.

[0040] Density can be obtained, for example, by conforming to JIS K2249 "Crude oil and petroleum products - Method for determining density". When conforming to JIS K2249 "Crude oil and petroleum products - Method for determining density", the density (g / cm3) is the density at 15°C. Sulfur content (ppm) can be obtained, for example, by conforming to JIS K2541 "Crude oil and petroleum products - Test method for sulfur content". Nitrogen content (ppm) can be obtained, for example, by conforming to JIS K2609 "Crude oil and petroleum products - Test method for nitrogen content". Kinematic viscosity can be obtained, for example, by conforming to JIS K2283 "Crude oil and petroleum products - Test method for kinematic viscosity and method for calculating viscosity index". When conforming to JIS K2283 "Crude oil and petroleum products - Test method for kinematic viscosity and method for calculating viscosity index", the kinematic viscosity is the kinematic viscosity at 30°C. The calorific value may include at least one of the total calorific value and the net calorific value. The calorific value can be obtained, for example, by conforming to JIS K2279 "Crude oil and petroleum products - Test method and calculation method for calorific value." The component ratios (volume %) may include saturated content, olefin content, total aromatic content, mono-ring aromatic content, di-ring aromatic content, and tri-ring or more (tri-ring+) aromatic content. These component ratios can be obtained, for example, by conforming to the Japan Petroleum Institute method JPI-5S-49-97 "Petroleum products - Hydrocarbon type test method - High-performance liquid chromatography."

[0041] The measurement method and units for the property values ​​representing the properties of the raw material oil 12 are not limited to the examples given above and can be freely changed. For example, they can be changed based on the type of raw material oil 12 and fraction, the user's purpose for using the property prediction device 14, etc.

[0042] File server 20 is an on-premise server computer that manages data files related to the analysis processing of raw oil 12. File server 20 may be a cloud-based server computer instead of an on-premise one. In Figure 1, file server 20 is shown as a single computer, but file server 20 may be a group of computers that make up a distributed system.

[0043] In the example shown in Figure 1, the file server 20 has databases for property analysis tables (hereinafter referred to as "Analysis Table DB24"), measurement result information 22 (hereinafter referred to as "Measurement Result DB26"), and property prediction information (hereinafter referred to as "Prediction Information DB28").

[0044] The property analysis table includes, for example, at least one of the following: [1] "Sample Information" including the name and type of the raw material oil 12; [2] "Property Information" including the analysis items, whether or not results were obtained, the units of the indicators, and the numerical values; [3] "Judgment Information" including the type of judgment criteria applied and whether or not the judgment criteria were met; and [4] Evidence Information including the source name and calculation method of the numerical values.

[0045] The measurement result information 22 includes, for example, at least one of the following: [1] raw values ​​of measurement data, [2] "sample information" including the name and type of raw material oil 12, [3] "analysis work information" including the name of the analyst, date and time of analysis, location of analysis, and model of the analytical instrument 16, and [4] "measurement information" including value definitions of measurement data and measurement conditions.

[0046] The prediction information includes, for example, at least one of the following: [1] a group of model parameters 56 used for predicting characteristics (Figure 2), [2] prediction result information 60 showing the prediction results of characteristics, and [3] "prediction evaluation information" including the accuracy and error of the prediction.

[0047] <Hardware configuration of the property prediction device 14> Figure 2 is a block diagram showing an example of the hardware configuration of the property prediction device 14 according to this embodiment.

[0048] The property prediction device 14 includes a processor 101, a non-volatile memory 102, a volatile memory 103, an interface (hereinafter also referred to as "IF") 104, an input device 105, and a display device 106. The processor 101, the non-volatile memory 102, the volatile memory 103, the IF 104, the input device 105, and the display device 106 are each connected to one another by a bus (data bus). The processor 101, the non-volatile memory 102, the volatile memory 103, the IF 104, the input device 105, and the display device 106 can each transmit data to one another via the bus. The property prediction device 14 functions in various functional configurations described later by the processor 101 executing a predetermined program stored in the volatile memory 103 or the non-volatile memory 102.

[0049] The volatile memory 103 is, for example, a semiconductor memory such as SRAM (Static Random Access Memory) or DRAM (Dynamic Random Access Memory). The volatile memory 103 stores programs and various data necessary for executing the processing in the property prediction device 14. The non-volatile memory 102 is, for example, a rewritable storage device such as flash memory or a hard disk. The non-volatile memory 102 stores programs and various data necessary for executing the processing in the property prediction device 14.

[0050] The processor 101 is, for example, a CPU (Central Processing Unit). However, the processor 101 is not limited to a CPU. The processor 101 may also be a GPU (Graphics Processing Unit). In a specific example, the processor 101 is a multi-core processor. The processor 101 may also be a single-core processor. The processor 101 may include multiple processors or cores and be capable of performing parallel processing. The processor 101 is configured to execute computer programs. The processor 101 may include, for example, an ASIC (Application Specific Integrated Circuit) as part, or programmable hardware such as an FPGA (Field Programmable Gate Array) or a CPLD (Complex Programmable Logic Device) as part.

[0051] IF104 is a communication interface for communication with the measurement terminal 18 or the file server 20. For example, IF104 is an Ethernet interface ("Ethernet" is a registered trademark).

[0052] The input device 105 is, for example, an input device that receives input from an external source. The input device 105 receives user operations and inputs those operations to the property prediction device 14. The input device 105 consists of, for example, a mouse, keyboard, touch sensor, and microphone.

[0053] The display device 106 is comprised of a device including, for example, a liquid crystal display or an organic EL (Electro-Luminescence) display.

[0054] The property prediction device 14 may consist of a single information processing device or multiple information processing devices. Furthermore, Figure 2 only shows a part of the main hardware configuration of the property prediction device 14, and the property prediction device 14 may have other configurations. In addition, the property prediction device 14 may omit some of the hardware exemplified; for example, the input device 105 and the display device 106 may be integrated as a touch panel. The measurement terminal 18 and the file server 20 have hardware components similar to those of the property prediction device 14.

[0055] <Functional configuration of the property prediction device 14> Figure 3 is a block diagram showing an example of the functional configuration of the property prediction device 14 shown in Figure 2. Specifically, this property prediction device 14 is a computer comprising a communication unit 32, an input unit 34, a display unit 36, a control unit 38, and a storage unit 40. Furthermore, the functions of each functional block shown in Figure 2 may be executed by a single computer or by multiple computers in a distributed manner. When the functions of each functional block shown in Figure 3 are executed by multiple computers in a distributed manner, these multiple computers may send and receive data via a communication network including a LAN (Local Area Network), a WAN (Wide Area Network), or the Internet.

[0056] The communication unit 32 is an interface for sending and receiving electrical signals to and from external devices. This allows the property prediction device 14 to, for example, acquire prediction target data 58 from the file server 20 and supply the prediction result information 60 it has generated to the file server 20.

[0057] The input unit 34 accepts information input from the user, for example. The display unit 36 ​​provides the user with, for example, prediction result information 60 generated by the property prediction device 14 in a visually recognizable manner. The property prediction device 14 constructs a graphical user interface (GUI) by, for example, combining the input function of the input unit 34 and the display function of the display unit 36.

[0058] The control unit 38 reads and executes the programs and data stored in the memory unit 40, thereby functioning as a data acquisition unit 42, a data processing unit 44, a feature generation unit 46, a learning processing unit 48, a characteristic prediction unit 50, and an output processing unit 52.

[0059] The data acquisition unit 42 acquires various data related to the learning process or prediction process. For example, the data acquisition unit 42 acquires a learning dataset 54, a group of model parameters 56, and data to be predicted 58. The data acquisition unit 42 may acquire data via communication from an external device including a file server 20, or it may acquire data via various parts of the characteristic prediction device 14 (for example, the input unit 34 or the storage unit 40).

[0060] The data processing unit 44 performs processing to modify the data acquired by the data acquisition unit 42. This processing includes, for example, at least one of the following: thinning, integration, and splitting.

[0061] "Sampling" is an information processing technique used to reduce the number of samples that make up measurement data. For example, by performing sampling every n (n≧1) data points, the number of data points is reduced to 1 / n.

[0062] "Integrated processing" is information processing for integrating a plurality of data with different sources prior to learning processing. This integrated processing includes [1] "aggregation processing" for aggregating a plurality of measurement data at one location or [2] association processing for associating the input / output relationships of learning data. As an example of aggregation processing, [1] aggregating N (N≧2) data files into n (1≦n<N) data files, [2] aggregating a plurality of data sheets into one data sheet, etc. can be cited. As an example of association processing, arranging a series of learning target data 62 and one or more property values (that is, correct values 64) corresponding to the learning target data 62 in a predetermined positional relationship, etc. can be cited. Through this integrated processing, multi-dimensional data (integrated data D3 in FIG. 8) in which an aggregate of measurement data (measurement data group D1 in FIG. 8) and an aggregate of property values regarding the raw material oil 12 (property data group D2 in FIG. 8) are associated is generated. This multi-dimensional data has measurement values and property values arranged in a matrix form.

[0063] "Division processing" is information processing for dividing the multi-dimensional data obtained through integrated processing into predetermined groups. This group is classified, for example, for each property regarding the raw material oil 12 within a section (hereinafter, also simply referred to as "property within a section"). Here, "property within a section" corresponds to a category consisting of a combination of a section and a property. For example, "yield at a temperature section of 0°C - 80°C" and "yield at a temperature section of 80°C - 150°C" are classified into different categories. Through this division processing, a learning data set 54 is generated for each property within a section.

[0064] The feature quantity generation unit 46 performs generation processing for generating a set of feature quantities (hereinafter referred to as "feature quantity set") regarding measurement values corresponding to the above-described sections from the data processed by the data processing unit 44 (for example, prediction target data 58 or learning target data 62). This generation processing includes [1] arithmetic processing for calculating an arithmetic value corresponding to a measurement value according to a predetermined arithmetic rule, [2] statistical processing for calculating a statistic regarding the population of measurement values using a statistical method, or [3] dimensionality reduction processing for generating M (1 ≦ M < N) feature quantities from N (N ≧ 2) measurement values or arithmetic values.

[0065] As an example of the arithmetic rule, there are four arithmetic operations of addition, subtraction, multiplication, and division, function operations, or LUT (Lookup Table) operations. As an example of the statistic, there are maximum value, minimum value, average value, median value, standard deviation, variance, etc. As an example of the dimensionality reduction processing, there are principal component analysis (PCA), independent component analysis (ICA), latent semantic analysis (LSA), or linear discriminant analysis (LDA).

[0066] When there are a plurality of prediction models PM, the feature quantity generation unit 46 may generate different feature quantity sets for each prediction model PM, or may generate a feature quantity set common to two or more prediction models PM. For example, when at least a part of the input / output characteristics is common among a plurality of prediction models PM, the feature quantity generation unit 46 may generate at least one of [1] a feature quantity set common to two or more sections, [2] a feature quantity set common to two or more properties, and [3] a feature quantity set common to two or more combinations of sections and properties.

[0067] When the prediction target data 58 or the learning target data 62 includes metadata regarding measurement, the feature quantity generation unit 46 may generate a feature quantity set using the metadata, or may generate feature quantities excluding the metadata. As an example of the metadata, there are [1] measurement conditions of the raw material oil 12, [2] type of the raw material oil 12, or [3] device information regarding the analyzer 16.

[0068] The learning processing unit 48 performs learning processing on the learner using the training dataset 54. Specifically, the learning processing unit 48 takes a set of features generated from the training data 62 as input values ​​and performs supervised learning with one or more correct values ​​64 as output values. Examples of rules for updating the learning parameters include stochastic gradient descent, momentum method, AdaGrad method, or Adam method. The learning error may be either the mean absolute value error (MAE) or the root mean square error (RMSE).

[0069] Through this learning process, the values ​​of each of the model parameter group 56 are determined, thereby generating a predictive model PM regarding the properties of the raw oil 12 or the properties of the fractions of the raw oil 12. The predictive model PM is a mathematical model that, when input from the feature set generated by the feature generation unit 46, outputs property values ​​related to the raw oil 12. The predictive model PM is, for example, a regression model that uses the feature set as explanatory variables and the raw oil property values ​​or fraction property values ​​as the dependent variable. Examples of regression models include linear regression, ridge regression, lasso regression, elastic network regression, logistic regression, random forest, gradient boosting decision tree, support vector machine, or neural network regression.

[0070] A predictive model PM is provided for each property within a section, or for each type of raw oil 12. In this case, the input / output characteristics of the learner may be the same (or uniform) regardless of the property within the section, or they may be different. Similarly, the input / output characteristics of the learner may be the same regardless of the type of raw oil 12, or they may be different. The learning processing unit 48 generates a predictive model PM for each property within a section by, for example, applying a learning process to learners with common input / output characteristics using a learning dataset 54 generated for each property within the section.

[0071] Examples of conditions for terminating the learning process (hereinafter referred to as "learning termination conditions") include: [Condition 1] the learning error becoming smaller than a threshold, [Condition 2] the number of learning iterations reaching the upper limit, or [3] a combination of (Condition 1) and (Condition 2) being met. The learning termination conditions can be set by the user as appropriate. The learning termination conditions may be the same regardless of the properties within a section, or they may differ depending on the section or properties. Specifically, the learning termination conditions for a first property value (e.g., yield) may be set to be stricter than the learning termination conditions for a second property value (e.g., density or sulfur content).

[0072] The property prediction unit 50 performs prediction processing to predict the properties of the raw oil 12 or the properties of the fractions of the raw oil 12 using the prediction model PM that has been trained by the learning processing unit 48. Prior to this prediction processing, the property prediction unit 50 selects the corresponding model parameter group 56 from among multiple types of model parameter groups 56 and sets it in the learner, thereby making the prediction model PM, which corresponds to the properties within the section, available. Based on the feature set and the prediction model PM, the property prediction unit 50 generates prediction result information 60 that includes predicted values ​​of the properties.

[0073] The output processing unit 52 performs output processing to cause the prediction results from the property prediction unit 50 (for example, at least one of the raw oil property values ​​and the fraction property values) to be output to the output means. The output means may be provided in the property prediction device 14 or in an external device different from the property prediction device 14. Examples of output include displaying visual information, outputting sound, or transmitting signals. Specifically, the output processing unit 52 either [1] supplies display data showing the analysis result screen 70 (Figure 12) to the display unit 36, or [2] transmits data including the prediction result information 60 to the file server 20.

[0074] The memory unit 40 stores the programs and data necessary for the control unit 38 to control each component. In the example shown in Figure 3, the memory unit 40 stores the training dataset 54, the model parameter group 56, the data to be predicted 58, and the prediction result information 60.

[0075] The training dataset 54 is a collection of training data used for training the learner. Each training dataset consists of a set of training data 62 and ground truth values ​​64. The training data 62 corresponds to the data used for training from the measurement data output from the analyzer 16. In addition to this measurement data, the training data 62 may also include the metadata described above (e.g., measurement conditions, type of raw oil 12, etc.). The training data 62 may also include, for example, a set of features generated by the feature generation unit 46. The ground truth values ​​64 correspond to the property values ​​of one or more raw oils 12 that correspond to the training data 62.

[0076] The model parameter group 56 is a collection of model parameters identified through the learning process of the predictive model PM. This model parameter group 56 is defined for each property within a section or for each type of raw oil 12. Examples of model parameters include [1] "hyperparameters" for identifying the model structure of the predictive model PM, or [2] "variable parameters" whose optimal values ​​change depending on the population of training data. Examples of variable parameters include weight coefficients between computational units and threshold values ​​for activation functions.

[0077] The data to be predicted 58 corresponds to the data used for predicting properties among the measurement data obtained through the measurement of the analytical device 16. After the properties have been predicted, the data to be predicted 58 may be used as the data to be trained 62. In addition to this measurement data, the data to be predicted 58 may also include the metadata described above (for example, measurement conditions, type of raw oil 12, etc.). The data to be predicted 58 may also include, for example, a set of features generated by the feature generation unit 46.

[0078] The prediction result information 60 includes at least one of the following: information related to the property analysis table described above, for example, [1] "sample information" including the name or type of the raw material oil 12, [2] "prediction information" including predicted values ​​for each analysis item, [3] "judgment information" including the type of judgment criteria applied and whether the judgment criteria are passed or failed, and [4] evidence information such as the method for predicting properties.

[0079] <Structure of the Predictive Model PM> Figure 4 shows an example of a combination of section and target species for prediction. More specifically, Figure 4 shows a level table combining 16 types of sections and 3 types of target species for prediction. Of the 48 total levels, 45 levels marked with a checkmark have been selected as targets for prediction.

[0080] In the example in Figure 4, the compartments include "Crude", "0(-)", "0-80", "80-100", "80-150", "150-200", "200-250", "150-250", "250-300", "300-350", "350-380", "250-380", "380-500", "380-600", "380(+)", "500(+)", and "600(+)". The symbol "XY" indicates a temperature compartment where the lower limit is X (°C) and the upper limit is Y (°C). The symbol "Z(+)" indicates a temperature compartment where the temperature is Z (°C) or higher. The symbol "Z(-)" indicates a temperature compartment where the temperature is Z (°C) or lower. "Crude" indicates the state of raw material oil 12 (crude oil) itself. In other words, "Crude" indicates a temperature zone where no distillation operation is performed on the raw oil 12, and "no temperature zone is defined." For the predicted species, yield (mass%) and density (g / cm³) are provided. 3 It contains ), and sulfur (mass%).

[0081] In the example in Figure 4, the 15 levels where the predicted type is "yield" and the categories range from "0 (-)" to "600 (+)" represent the properties of the raw oil 12. Furthermore, the 1 level where the predicted type is "density" and the category is "Crude" also represents the properties of the raw oil 12. In addition, the 1 level where the predicted type is "sulfur content" and the category is "Crude" also represents the properties of the raw oil 12. In other words, in the example in Figure 4, 17 levels of raw oil 12 properties are shown as predictable. In the example in Figure 4, the 17 levels enclosed by solid lines represent the properties of the raw oil 12.

[0082] In the example in Figure 4, the 14 levels where the predicted type is "density" and the categories range from "0-80" to "600(+)" represent the properties of the fractions of the raw oil 12. Furthermore, the 14 levels where the predicted type is "sulfur content" and the categories range from "0-80" to "600(+)" also represent the properties of the fractions of the raw oil 12. In other words, in the example in Figure 4, 28 levels of the properties of the fractions of the raw oil 12 are shown as prediction targets. In the example in Figure 4, the 28 levels enclosed by the dotted lines represent the properties of the fractions of the raw oil 12.

[0083] In this embodiment, since the raw material oil 12 is crude oil, for example, section "Crude" corresponds to crude oil. Section "0(-)" corresponds to liquefied petroleum gas (LPG). Sections "0-80" and "80-150" correspond to naphtha or gasoline. Section "150-250" corresponds to jet fuel or kerosene. Section "250-380" corresponds to diesel fuel. Section "380-600" corresponds to vacuum diesel fuel or lubricating oil. Section "380(+)" corresponds to atmospheric residual oil or heavy fuel oil C. Section "600(+)" corresponds to vacuum residual oil or asphalt.

[0084] Here, the division of the compartments is not limited to the example in Figure 4. For example, a compartment "0-150" may be added, or compartments "0-80" and "80-150" may be omitted instead of a newly created compartment "0-150". If a compartment "0-150" exists, it may be used as a compartment corresponding to naphtha or gasoline instead of "0-80" and "80-150". Furthermore, the user may divide the compartments to distinguish between light and heavy components, for example, by assigning compartment "0-150" to naphtha, compartment "0-80" to light naphtha, and compartment "80-150" to heavy naphtha. Also, the lower or upper limit of a compartment corresponding to a certain fraction, i.e., the range of distillation temperatures required to obtain a certain fraction, may vary depending on the distillation apparatus owned by the user. Therefore, the compartments can be adjusted as appropriate by the user. The user can make adjustments as appropriate based on factors such as the type of raw material oil they provide, the type of fraction they desire, and the specifications of the distillation equipment (equipment for obtaining fractions) they own.

[0085] Figure 4 shows an example of the structure of a predictive model PM for predicting the properties of raw oil 12. This predictive model PM consists of a learner that takes a set of features as input and outputs a single property value (or a predicted property value). When the predictive model PM receives a set of features as input values, it outputs a property value as an output value. In the example in Figure 4, multiple types of predictive models PM have a common model structure. By selectively setting multiple types of model parameter groups 56 in a predetermined memory area (indicated as "LP"), at least one predictive model PM for each property within a section is constructed. For example, a predictive model PM for predicting the "yield in temperature sections from 0 to 80°C" is constructed.

[0086] In this disclosure, the prediction model PM may be referred to as a raw oil prediction model when it takes a set of features as input and outputs raw oil property values. Furthermore, the prediction model PM may be referred to as a fraction prediction model when it takes a set of features as input and outputs fraction oil property values.

[0087] For example, a raw material oil prediction model is provided for each level (combination of section and properties) related to the properties of the raw material oil 12. For example, if all 17 levels exemplified in Figure 4 are to be predicted, 17 raw material oil prediction models are provided.

[0088] For example, a fraction prediction model is provided for each level (combination of section and properties) related to the properties of the fractions of the raw oil 12. For example, if all 28 levels exemplified in Figure 4 are to be predicted, 28 fraction prediction models are provided.

[0089] [Operation of Analysis System 10] The analysis system 10 in this embodiment is configured as described above. Next, the operation of the analysis system 10 (in particular, the property prediction device 14) will be explained with reference to Figures 6 to 12.

[0090] <Learning operation by the property prediction device 14> Figure 6 is a flowchart showing an example of the learning operation by the property prediction device 14 shown in Figures 1 to 3.

[0091] In step SP10, the data acquisition unit 42 refers to the analysis table DB24 and measurement results DB26 of the file server 20 and acquires the data necessary for the learning process (in this case, the learning target data 62 and the correct answer value 64). This acquires the learning dataset 54.

[0092] Figure 7 shows an example of measurement data obtained by high-temperature gas chromatography. The horizontal axis of the graph represents retention time (in minutes), and the vertical axis represents signal intensity (in dimensionless). This measurement data consists of signal intensities sequentially arranged at predetermined sampling intervals (e.g., 1 second). "Retention time" corresponds to the time from injection of the sample until each component is detected (or the time it remains in the column). Signal intensity corresponds to an index proportional to the concentration of the component. This measurement data is used as training data 62 or prediction data 58 (Figure 3).

[0093] In step SP12 of Figure 6, the data processing unit 44 performs decimation on the training dataset 54 (in this case, the training target data 62) acquired in step SP10, as needed.

[0094] In step SP14, the data processing unit 44 performs a formatting process on the training dataset 54 that was thinned in step SP12.

[0095] Figure 8 shows an example of a method for formatting the training dataset 54. The measurement data group D1 is a collection of data where information about the raw oil 12 (hereinafter referred to as raw oil information) and measurement data are associated. The property data group D2 is a collection of data where raw oil information and property values ​​are associated. The raw oil information is, for example, the metadata described above. First, by associating the measurement data group D1 and the property data group D2 using the raw oil information as a key, a single integrated data D3 is generated in which the measurement data and property values ​​are integrated. Next, by dividing the integrated data D3 into sections and into properties, a divided data group D4 consisting of multiple (45 in the example in Figure 4) divided data is generated. Each divided data consists of a common explanatory variable and one type of target variable.

[0096] In step SP16 of Figure 6, the feature generation unit 46 generates feature sets for each property within a section from the training dataset 54 that has been formatted in step SP14.

[0097] In step SP18, the learning processing unit 48 performs learning processing on the learner using the feature set generated in step SP16.

[0098] Figure 9 shows an example of how to generate the prediction model PM in Figure 5. The feature generation unit 46 selects the first divided data from the divided data group D4 and generates the first set of features using principal component analysis (PCA). The learning processing unit 48 performs learning on the learner using the combination of the first set of features and property values. This generates the first prediction model PM. The same operation is repeated thereafter to sequentially generate the nth (1 ≤ n ≤ N) prediction model PMs. In this way, N prediction model PMs are generated.

[0099] In step SP20 of Figure 6, the learning processing unit 48 checks whether the termination condition for the learning process is met. If the termination condition is not met (step SP20: NO), the control unit 38 returns to step SP10 and repeats steps SP10 to SP20 sequentially until the termination condition is met. On the other hand, if the termination condition is met (step SP20: YES), the learning processing unit 48 proceeds to the next step SP22.

[0100] In step SP22, the learning processing unit 48 saves the values ​​of the model parameter group 56 at the time the termination condition in step SP20 was met.

[0101] In this way, the property prediction device 14 completes the learning operation shown in the flowchart of Figure 6. As a result, the property prediction unit 50 can utilize the prediction model PM (Figure 5) through the set of model parameter group 56.

[0102] <Predictive operation by the property prediction device 14> Figure 10 is a functional block diagram relating to the predictive operation by the control unit 38 in Figure 3. Figure 11 is a flowchart showing an example of the predictive operation by the property prediction device 14 in Figures 1 to 3. The following explanation of this predictive operation will refer to Figures 10 and 11 together.

[0103] In step SP30, the data acquisition unit 42 acquires information designated as the target of prediction (i.e., designated information) via the input unit 34 through input operations by the analyst. This designated information includes, for example, the sample name, section type, or property type of the raw material oil 12.

[0104] In step SP32, the data acquisition unit 42 refers to the specified information acquired in step SP30 and acquires the measurement data to be predicted (i.e., the prediction target data 58) from the measurement result DB26.

[0105] In step SP34, the property prediction unit 50 refers to the specified information obtained in step SP30 and retrieves a group of model parameters 56 to be used for prediction from the prediction information DB28. As a result, one or more prediction models PM are selected.

[0106] In step SP36, the control unit 38 performs preprocessing on the prediction target data 58 acquired in step SP32. This preprocessing includes [1] decimation by the data processing unit 44 (SP36A), [2] formatting by the data processing unit 44 (SP36B), and [3] generation by the feature generation unit 46 (SP36C). Each information processing is performed in the same manner as in the learning operation described above (Figures 6 to 9).

[0107] In step SP38, the property prediction unit 50 uses the prediction model PM selected in step SP34 to predict the properties of the raw oil 12 from the set of features generated through the preprocessing in step SP36. This yields one or more predicted values.

[0108] In step SP40, the output processing unit 52 outputs prediction result information 60, which includes the property values ​​predicted in step SP38. For example, the output processing unit 52 generates display data for displaying the analysis result screen 70 and supplies the display data to the display unit 36. As a result, the analysis result screen 70 is displayed on the display unit 36 ​​of the property prediction device 14.

[0109] Figure 12 shows an example of the analysis results screen 70 displayed on the display unit 36 ​​in Figure 2. The analysis results screen 70 is provided with a sample name field 72 showing the sample name of the raw material oil 12, and an analysis results field 74 showing the analysis results of the raw material oil 12. The analysis results field 74 is provided with a property analysis table showing the correspondence between the analysis items and the measurement conditions. Predicted property values ​​76 are entered in the corresponding fields of the property analysis table. Through this property analysis table, the analyst can grasp the properties of the raw material oil 12 at a glance.

[0110] In this way, the property prediction device 14 completes the prediction operation shown in the flowchart of Figure 11. As a result, users, including the analysis operator, can use the property analysis table that reflects the prediction results by updating the analysis table DB 24.

[0111] [Summary of Embodiments] As described above, the analysis system 10 in this embodiment includes an analysis device 16 that outputs measurement data (here, prediction target data 58 or learning target data 62), which is a collection of measured values ​​relating to the raw material oil 12, and a property prediction device 14 that performs analytical processing on the measurement data output from the analysis device 16 to predict the properties of the raw material oil 12 or the properties of the fractions of the raw material oil 12.

[0112] The property prediction device 14 according to this embodiment includes a data acquisition unit 42 that acquires measurement data which is a collection of information about the components contained in the raw oil 12 and measurement values ​​corresponding to the information about the components; a feature quantity generation unit 46 that generates a feature quantity set which is a collection of feature quantities related to the measurement values ​​from the measurement data acquired by the data acquisition unit 42; and a property prediction unit 50 that predicts the properties of the raw oil 12. The property prediction unit 50, upon input of the feature quantity set generated by the feature quantity generation unit 46, includes a raw oil prediction model that outputs the yield of fractions that can be obtained from the raw oil 12 according to predetermined temperature zones. There are multiple temperature zones, and the number of raw oil prediction models that output yields is equal to the number of temperature zones.

[0113] This configuration includes a predictive model PM (raw oil prediction model) that, upon input of a feature set, outputs the yield of fractions obtainable from the raw oil 12 according to predetermined temperature zones. There are multiple temperature zones, and the number of raw oil prediction models that output yields is equal to the number of temperature zones. Therefore, by using different predictive models PM for each fraction yield corresponding to the temperature zone, the accuracy of predicting the properties of the raw oil 12 can be improved. Furthermore, the user can accurately formulate production plans for products or base materials to be produced from the raw oil 12 based on the prediction results regarding the yield of fractions obtainable from the raw oil 12.

[0114] Furthermore, the properties prediction unit 50 predicts at least one of the properties of the raw oil 12, in addition to the yield, from the density of the raw oil 12, the sulfur content of the raw oil 12, and the nitrogen content of the raw oil 12. A raw oil prediction model is provided for each type of property of the raw oil 12 predicted by the properties prediction unit 50.

[0115] In this configuration, the properties prediction unit 50 further predicts at least one of the properties of the raw oil 12, in addition to the yield, from the density of the raw oil 12, the sulfur content of the raw oil 12, and the nitrogen content of the raw oil 12. A raw oil prediction model is provided for each type of raw oil property predicted by the properties prediction unit 50. Therefore, since the raw material prediction model (prediction model PM) is used according to the type of raw oil property predicted by the properties prediction unit 50, the accuracy of predicting the properties of the raw oil 12 can be improved. Furthermore, since each property of the raw oil 12 can be predicted from measurement data, it is possible to omit performing predetermined analyses on the raw oil 12 for each property. As a result, the user can more efficiently plan production plans for products or base materials to be produced from the raw oil 12 based on the predicted properties. For example, when predicting sulfur content, since sulfur content is corrosive and accelerates the deterioration of production equipment, this can be used to limit the amount of raw oil 12 processed or to select the production equipment itself that processes the raw oil 12. For example, when predicting sulfur or nitrogen content, sulfur and nitrogen are types of impurities that have adverse effects on the environment and on catalysts used in production facilities. Therefore, in order to mitigate adverse effects on production facilities or the environment during the production or use of products or base materials that can be produced from the raw oil 12, the necessity of removing these impurities in the production process of products and base materials, and the conditions for such treatment (for example, controlling the sulfur content in the raw oil 12 using a desulfurization unit) can be considered. For example, when predicting density, density changes depending on the types of components contained in the raw oil 12 (for example, the number of carbon atoms) and the amount of each component, so it can be used to understand the properties of the raw oil 12 and the volume of fractions that can be obtained from the raw oil 12. In particular, when the raw oil 12 is crude oil, crude oil can be classified into ultra-light crude oil, light crude oil, medium crude oil, heavy crude oil, and ultra-heavy crude oil based on density or specific gravity or API degree converted from density. Since the fraction yield can change depending on the type of crude oil classified, density is important in production planning for petroleum products or base materials at refineries.

[0116] Furthermore, the property prediction device 14 includes a property prediction unit 50 that further predicts the properties of the fraction and, upon input of a set of features generated by the feature generation unit 46, outputs a fraction property value indicating the properties of the fraction according to the temperature zone.

[0117] This configuration further includes a fraction prediction model (prediction model PM) that, when given a set of features as input, outputs fraction property values ​​indicating the properties of the fraction according to the temperature zone. Therefore, it is possible to predict the properties of fractions obtainable from the raw oil 12 while reducing the impact on the prediction accuracy of the properties of the raw oil 12.

[0118] Furthermore, in the property prediction device 14, the properties of the fraction include at least one of density, sulfur content, and nitrogen content, and the number of fraction prediction models is provided according to the combination of the number of temperature compartments and the number of fraction properties predicted by the property prediction unit 50.

[0119] This configuration provides a number of fraction prediction models corresponding to the number of temperature zones and the number of fraction properties. Therefore, by using different prediction models depending on the combination of fraction properties and temperature zones in the raw oil 12, the accuracy of predicting the fraction properties of the raw oil 12 can be improved. Furthermore, since each fraction property obtainable from the raw oil 12 can be predicted from the measurement data of the raw oil 12, it is possible to omit performing predetermined analyses on each fraction property. As a result, users can more efficiently plan production for products or base materials produced from the raw oil 12 based on the predicted fraction properties. For example, when predicting the sulfur or nitrogen content of a fraction, sulfur and nitrogen are types of impurities that have adverse effects on the environment. This can be used to consider the necessity of processing to remove these impurities from the fraction and the conditions for such processing (e.g., controlling the sulfur content in the fraction using a desulfurization unit) in order to keep the content of these impurities within the regulatory limits based on laws and regulations that must be met when using the fraction as a product or base material. For example, when predicting the density of a fraction, the density changes depending on the type of components contained in the fraction (e.g., the number of carbon atoms) and the amount of each component, so this can be used to understand the properties and volume of the fraction.

[0120] Furthermore, in the property prediction device 14, the measurement data is data measured by a gas chromatograph, the measured value is signal intensity, and the information regarding the components is retention time.

[0121] In this configuration, the measurement data is data measured by a gas chromatograph. The gas chromatograph separates and qualitatively and quantitatively analyzes each component contained in the raw material oil 12 by heating and vaporizing the raw material oil 12. Therefore, the gas chromatograph can be considered a device that simulates distillation of the raw material oil 12 and measures the boiling point distribution of the raw material oil 12. Consequently, the measurement data from the gas chromatograph is preferable for predicting the yield of fractions that can be obtained from the raw material oil 12 according to predetermined temperature zones.

[0122] Furthermore, in the properties prediction device 14, the temperature zones are determined based on the type of fraction desired by the user for the raw material oil 12. The raw material oil 12 is crude oil, and its type includes at least one of LPG, naphtha, gasoline, kerosene, jet fuel, diesel fuel, lubricating oil, heavy oil, residual oil, and asphalt.

[0123] In this configuration, temperature zones are determined based on the type of fraction the user desires from the crude oil. Therefore, the user can define temperature zones as the temperature range in which the fractions related to the petroleum product for which they wish to develop a production plan can be obtained. This allows the user to accurately plan the production of petroleum products and other products to be produced from crude oil, based on the predicted yield of each fraction.

[0124] Furthermore, in the property prediction device 14, the feature generation unit 46 compresses the number of dimensions of the measurement data to generate a set of features.

[0125] This configuration allows for the extraction of features with a higher correlation to the properties of the raw material oil 12 within a section, while reducing the amount of information contained in the measurement data.

[0126] Furthermore, in the property prediction device 14, the raw oil prediction model is a mathematical model in which a learning process is performed using different learning datasets 54 for each property of the raw oil 12 within a section, with a learner having common input / output characteristics.

[0127] This configuration increases the regularity of the learning process by standardizing the input / output characteristics, thereby reducing the amount of computation required.

[0128] Furthermore, in the property prediction device 14, the feature quantity generation unit 46 generates a set of feature quantities common to the properties of two or more temperature zones or two or more raw material oils 12.

[0129] This configuration allows for a reduction in the computational load during training or prediction processes through the commonization of feature sets.

[0130] Furthermore, the property prediction device 14 is further equipped with a data processing unit 44 that generates integrated data D3 by associating a collection of measurement data (measurement data group D1) with a collection of property values ​​related to the raw oil 12 (property data group D2), and then divides the integrated data D3 according to the properties of the raw oil 12 in each temperature zone to generate a learning dataset 54.

[0131] This configuration allows for the more efficient generation of the training dataset 54 through data integration and splitting.

[0132] Furthermore, the property prediction device 14 is further equipped with a learning processing unit 48 that generates a raw oil prediction model or a fraction prediction model for each property of the raw oil 12 in a temperature zone by performing learning processing on a learner with common input / output characteristics using a learning dataset 54 generated by the data processing unit 44 for each property of the raw oil 12 in a temperature zone.

[0133] This configuration increases the regularity of the learning process by standardizing the input / output characteristics, thereby reducing the amount of computation required.

[0134] Furthermore, the property prediction device 14 is further equipped with an output processing unit 52 that instructs the output means to output a property analysis table including the raw oil property values ​​predicted by the property prediction unit 50.

[0135] This configuration allows for the property analysis table to be supplemented by predicting the properties of the raw material oil 12.

[0136] [Differentiation] This disclosure is not limited to the embodiments described above, and can be freely modified without departing from the spirit of this disclosure. Alternatively, the respective configurations may be combined as they see fit, without creating any technical inconsistencies. Alternatively, the execution status or execution order of each step constituting the flowchart may be changed, without creating any technical inconsistencies.

[0137] In the embodiments described above, the case in which the raw material oil 12 is crude oil or petroleum products was used as an example, but the type of raw material oil 12 is not limited to these. For example, the raw material oil 12 may be [1] industrial mineral oils used as insulating oil, refrigeration oil, or lubricating oil, [2] extracts from environmental samples using oily solvents, [3] animal fats and oils derived from edible animals, or [4] vegetable fats and oils derived from plants.

[0138] In the embodiments described above, the case where the analyzer 16 is a gas chromatograph was used as an example, but the type of analyzer 16 is not limited to this. For example, the analyzer 16 may be an "infrared spectrometer" (hereinafter referred to as an IR analyzer) that uses infrared spectroscopy, or a "nuclear magnetic resonance analyzer" (hereinafter referred to as an NMR analyzer) that uses nuclear magnetic resonance. The analyzer 16 is not particularly limited as long as it is a device that can acquire measurement data which is a collection of information about the components contained in the raw oil and measurement values ​​corresponding to the information about the components. For example, when using an IR analyzer instead of a gas chromatograph, "absorbance" may be used instead of "signal intensity" as the measurement value, and "wavenumber" may be used instead of "retention time" as the information about the components.

[0139] In the embodiment described above, the case in which the property prediction device 14 performs a learning process was used as an example, but the device configuration is not limited to this. For example, the property prediction device 14 does not need to be provided with a learning processing unit 48. In this case, an external device other than the property prediction device 14 stores the model parameter group 56 obtained through the learning process in the prediction information DB 28, and then the property prediction device 14 can perform property prediction processing by reading and using the model parameter group 56.

[0140] In this specification or in the claims, information, physical quantities, features, sample values, indicators, parameters, etc., may be expressed using absolute values, relative values ​​from a given value, or corresponding other information.

[0141] Where expressions such as "acquiring / setting information / using / based on / with / as input" (including similar expressions) are used in this specification or claims, unless otherwise specified, this includes using the information itself or using information that has been processed in some way (e.g., noise-added, normalized, features extracted from the information, intermediate representation of the information, etc.). Furthermore, where it is stated that some result is obtained by "acquiring / setting information / using / based on / as input" (including similar expressions), unless otherwise specified, this includes cases where the result is obtained based solely on the information in question or where the result is influenced by other information, factors, conditions, and / or states other than the information in question. Furthermore, where it is stated that "information is output" (including similar expressions), unless otherwise specified, this includes cases where the information itself is used as output or where information that has been processed in some way (e.g., noise-added, normalized, features extracted from the information, intermediate representation of various types of information, etc.) is used as output. [Explanation of Symbols]

[0142] 10...Analysis system, 12...Raw oil, 14...Property prediction device, 16...Analysis device, 22...Measurement result information, 32...Communication unit (output means), 34...Input unit, 36...Display unit (output means), 38...Control unit, 40...Storage unit, 42...Data acquisition unit, 44...Data processing unit, 46...Feature generation unit, 48...Learning processing unit, 50...Property prediction unit, 52...Output processing unit, 54...Training dataset, 56...Model parameter group, 58...Prediction target data (measurement data), 60...Prediction result information, 62...Training target data (measurement data), 64...Correct value, 70...Analysis result screen, D1...Measurement data group, D2...Property data group, D3...Integrated data, D4...Divided data group, PM...Prediction model

Claims

1. A data acquisition unit acquires measurement data which is a collection of information about the components contained in the raw material oil and measurement values ​​corresponding to the information about the components, A feature generation unit generates a feature set, which is a collection of features related to the measured values, from the measurement data acquired by the data acquisition unit, The system includes a property prediction unit that predicts the properties of the raw material oil, The property prediction unit includes a raw oil prediction model that, upon input of the feature set generated by the feature generation unit, outputs the yield of fractions obtainable from the raw oil according to a predetermined temperature zone. There are multiple temperature zones, The number of feed oil prediction models that output the yield is equal to the number of temperature zones. Property prediction device.

2. The property prediction unit further predicts, in addition to the yield, at least one of the properties of the raw material oil: the density of the raw material oil, the sulfur content of the raw material oil, and the nitrogen content of the raw material oil. The aforementioned raw material oil prediction model is provided for each type of raw material oil property predicted by the property prediction unit. The property prediction device according to claim 1.

3. The property prediction unit further predicts the properties of the fraction, The fraction prediction model further includes, upon input of the feature set generated by the feature generation unit, outputting fraction property values ​​indicating the properties of the fraction according to the temperature section. The property prediction device according to claim 1.

4. The properties of the said fraction include at least one of density, sulfur content, and nitrogen content. The fraction prediction model is provided with a number corresponding to the combination of the number of temperature compartments and the number of fraction properties predicted by the property prediction unit. The property prediction device according to claim 3.

5. The aforementioned measurement data is data measured by a gas chromatography device. The aforementioned measurement is signal intensity. The information relating to the aforementioned component is the retention time. The property prediction device according to claim 1.

6. The temperature zones are determined based on the type of fraction desired by the user for the raw material oil. The aforementioned raw material oil is crude oil, The aforementioned types include at least one of LPG, naphtha, gasoline, kerosene, jet fuel, diesel fuel, lubricating oil, heavy oil, residual oil, and asphalt. The property prediction device according to claim 1.

7. The feature generation unit compresses the number of dimensions of the measurement data to generate the feature set. The property prediction device according to claim 3.

8. The aforementioned raw material oil prediction model is a mathematical model in which a learning process is performed using different learning datasets for each property of the raw material oil within the temperature zone, with a learner having common input / output characteristics. The property prediction device according to claim 7.

9. The feature generation unit generates a set of feature quantities that are common to two or more temperature zones or two or more properties of the raw material oil. The property prediction device according to claim 8.

10. The system further includes a data processing unit that generates integrated data by associating the collection of measurement data with the collection of property values ​​related to the raw material oil, and then divides the integrated data according to the properties of the raw material oil within the temperature compartment to generate the learning dataset. The property prediction device according to claim 9.

11. The learning processing unit further comprises a learning unit that performs learning processing on a learning device having common input / output characteristics, using the learning dataset generated by the data processing unit for each property of the raw material oil within the temperature compartment, thereby generating a raw material oil prediction model or a fraction prediction model for each property of the raw material oil within the temperature compartment. The property prediction device according to claim 10.

12. The system further includes an output processing unit that instructs an output means to output a property analysis table including the raw oil property values ​​predicted by the property prediction unit. The property prediction device according to claim 1.

13. An acquisition process that acquires information about the components contained in the raw material oil and measurement data which is a collection of measurement values ​​corresponding to the information about the components, A generation process that generates a feature set, which is a collection of features related to the measured values, from the measurement data obtained through the acquisition process, A property prediction process for predicting the properties of the aforementioned raw material oil, Run this on one or more computers. The property prediction process is performed using a raw oil prediction model that, upon inputting the set of features generated through the generation process, outputs the yield of fractions obtainable from the raw oil according to a predetermined temperature zone. There are multiple temperature zones, The number of feed oil prediction models that output the yield is equal to the number of temperature zones. Property prediction program.

14. An analytical device that outputs information about the components contained in the raw material oil and measurement data which is a collection of measured values ​​corresponding to the information about the components, A property prediction device that performs analytical processing on the measurement data output from the analytical device and predicts the properties of the raw material oil, Equipped with, The aforementioned property prediction device, A generation process that generates a feature set, which is a collection of features related to the measured values, from the aforementioned measurement data, The process involves performing a property prediction process to predict the properties of the aforementioned raw material oil, The property prediction process is performed using a raw oil prediction model that, upon inputting the set of features generated through the generation process, outputs the yield of fractions obtainable from the raw oil according to a predetermined temperature zone. There are multiple temperature zones, The number of feed oil prediction models that output the yield is equal to the number of temperature zones. Analysis system.

Images