Analytical device, analytical method, and program

Abstract

To provide an analyzing apparatus that makes it easier to reduce energy consumption.SOLUTION: Disclosed is an analyzing apparatus 10 for executing the analysis of consumption information as information related to energy consumed in an analysis object, which comprises: a cluster unit 13 for categorizing plural pieces of consumption information into a preset number of clusters based on similarity to each other; a model unit 15 for creating each model with plural different feature quantities related to at least one of the analysis object and the consumption information as an explanatory variable and the categorized clusters as an objective variable; and an analysis unit 16 for obtaining importance as a degree of contribution to model construction about each of the plural feature quantities used in each model.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an analysis device, an analysis method, and a program. [Background technology]

[0002] In order to achieve carbon neutrality of consumed energy, various analytical methods have been proposed for use in reducing the amount of consumed energy (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2015-231328 A Summary of the Invention [Problem to be solved by the invention]

[0004] In the analysis of energy consumption in a building, data on the energy consumption is collected, and methods for reducing the amount of energy consumption are analyzed based on the collected data. In the analysis, specialized engineers grasp the characteristics of the building and perform tasks such as collating data and drawing graphs.

[0005] However, when analyzing ways to reduce the energy consumption of a building, there are dozens of items to analyze, and the analysis requires a lot of man-hours. In addition, because there are so many items to analyze, there are cases where the analysis results contain errors, making it difficult to propose appropriate reduction methods.

[0006] Therefore, in order to propose an appropriate reduction method, analysis work is required to be performed by engineers with various specialized knowledge such as knowledge of data analysis, knowledge of buildings, knowledge of equipment installed in buildings, knowledge of energy, etc. However, it is difficult to secure engineers with such various specialized knowledge, and there has been a problem in that it is difficult to propose an appropriate reduction method.

[0007] The present invention has been made to solve the above-mentioned problems, and has an object to provide an analysis device and a program that can facilitate reduction in energy consumption. [Means for solving the problem]

[0008] In order to achieve the above object, the present invention provides the following means. An analysis device according to a first aspect of the present invention is an analysis device that analyzes consumption information, which is information regarding energy consumed by an analysis target, and includes: a cluster unit that classifies a plurality of pieces of the consumption information into a preset number of clusters based on their mutual similarities; a model unit that creates a model using a plurality of different feature quantities related to at least one of the analysis target and the consumption information as explanatory variables and the classified clusters as objective variables; and an analysis unit that calculates the importance, which is the degree to which each of the feature quantities used in the model contributes to the construction of the model.

[0009] A program according to a second aspect of the present invention is a program for causing a computer to analyze consumption information, which is information regarding energy consumed by an analysis target, and causes the computer to realize a cluster classification function for classifying a plurality of the consumption information into a preset number of clusters based on their mutual similarities, a model creation function for creating a model in which a plurality of different feature quantities related to at least one of the analysis target and the consumption information are used as explanatory variables and the classified clusters are used as objective variables, and an analysis function for determining the importance, which is the degree to which each of the feature quantities used in the model contributes to the construction of the model.

[0010] According to the analysis device according to the first aspect of the present invention and the program according to the second aspect, a model is created for each cluster categorized based on the similarity of consumption information, and the importance of features, which are explanatory variables of the model, is calculated. Since the importance of features is calculated, it is possible to know features that have a large effect on reducing energy consumption. This makes it easier to reduce energy consumption.

[0011] In the first aspect of the invention, it is preferable that a reclustering unit is further provided which arranges the clusters in order of the number of pieces of consumption information contained therein, and categorizes each of the clusters from the cluster with the largest number of pieces of consumption information to a predetermined rank into at least two or more divided clusters based on the mutual similarities of the plurality of pieces of consumption information contained in the cluster, and the model unit also creates a model for the divided clusters.

[0012] By classifying the divided clusters by the reclustering unit in this way, the characteristics of the created model become clearer than when the number of pre-set values ​​in the clustering unit is increased. This makes it easier to understand the features that have a large effect on reducing energy consumption, making it easier to reduce energy consumption.

[0013] In the first aspect of the above invention, it is preferable that a planning unit is further provided that determines a method of reducing the energy consumption in the analysis target based on the feature and the importance of the feature.

[0014] By providing the planning unit in this manner, a method for reducing the amount of energy consumption is proposed. By proposing a method for reducing the amount of energy consumption, it becomes easier to reduce the amount of energy consumption.

[0015] In the first aspect of the above invention, it is preferable that the system further includes an index setting unit that determines an index indicating the energy consumption tendency of the analysis subject based on the feature amount and the importance of the feature amount.

[0016] By providing the index setting unit in this manner, an index showing the energy consumption tendency is presented. By presenting an index showing the energy consumption tendency, it becomes easier to reduce the amount of energy consumption. Effect of the Invention

[0017] The analysis device according to the first aspect of the present invention and the program according to the second aspect of the present invention have the effect of making it easier to reduce energy consumption by making it possible to know features that have a large impact on reducing energy consumption. [Brief description of the drawings]

[0018] [Figure 1] FIG. 1 is a schematic diagram illustrating a configuration of an analysis device according to an embodiment of the present invention. [Diagram 2] 10 is a flowchart illustrating the contents of a calculation process in the analysis device. [Diagram 3] 11 is a table explaining input data items used for format conversion of consumption information. [Figure 4] 1 is a table illustrating feature amounts that are explanatory variables used in modeling processing. [Diagram 5] 13 is a graph illustrating the energy consumption pattern of cluster No. 0. [Figure 6] 13 is a graph illustrating the energy consumption pattern of cluster No. 1. [Figure 7] 13 is a graph illustrating the energy consumption pattern of cluster No. 2. [Figure 8] 13 is a graph illustrating the energy consumption pattern of cluster No. 1a. [Figure 9] 13 is a graph illustrating the energy consumption pattern of cluster No. 1b. [Figure 10] FIG. 10(a) is a graph showing the analysis results of the feature importance of cluster No. 0, and FIG. 10(b) is a graph explaining the magnitude of the standardized feature of cluster No. 0. [Figure 11] FIG. 11(a) is a graph showing the analysis results of the feature importance of cluster No. 1, and FIG. 11(b) is a graph explaining the magnitude of the standardized feature of cluster No. 1. [Figure 12] FIG. 12(a) is a graph showing the analysis results of the feature importance of cluster No. 2, and FIG. 12(b) is a graph explaining the magnitude of the standardized feature of cluster No. 2. [Figure 13] FIG. 13(a) is a graph showing the analysis results of the feature importance of cluster No. 1a, and FIG. 13(b) is a graph explaining the magnitude of the standardized feature of cluster No. 1a. [Figure 14] FIG. 14(a) is a graph showing the analysis results of the feature importance of cluster No. 1b, and FIG. 14(b) is a graph explaining the magnitude of the standardized feature of cluster No. 1b. [Figure 15] 13 is a graph illustrating the energy consumption pattern of Comparison No. 0. [Figure 16] 13 is a graph illustrating the energy consumption pattern of Comparative Example No. 1. [Figure 17] 13 is a graph illustrating an energy consumption pattern of Comparative Example No. 2. [Figure 18] 13 is a graph illustrating the energy consumption pattern of Comparative Example No. 3. [Figure 19] FIG. 19(a) is a graph showing the analysis results of feature importance for Comparison No. 0, and FIG. 19(b) is a graph illustrating the magnitude of the standardized feature for Comparison No. 0. [Figure 20] FIG. 20(a) is a graph showing the analysis results of feature importance for Comparison No. 1, and FIG. 20(b) is a graph illustrating the magnitude of the standardized feature for Comparison No. 1. [Figure 21] FIG. 21(a) is a graph showing the analysis results of feature importance for Comparison No. 2, and FIG. 21(b) is a graph illustrating the magnitude of the standardized feature for Comparison No. 2. [Figure 22] FIG. 22(a) is a graph showing the analysis results of feature importance for Comparison No. 3, and FIG. 22(b) is a graph illustrating the magnitude of the standardized feature for Comparison No. 3. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] An analysis device 10 according to an embodiment of the present invention will be described below with reference to Fig. 1 to Fig. 22. The analysis device 10 of this embodiment is a device that analyzes consumption information, which is information about energy consumed in an analysis target. Furthermore, the analysis device 10 may plan measures to reduce the energy consumption amount in the analysis target and generate a benchmark for energy consumption trends.

[0020] In this embodiment, the analysis targets multiple buildings (in other words, a group of buildings), and each individual building in the group of buildings is treated as one unit in the analysis. The consumption information includes at least one of the energy consumption consumed by each individual building as a whole, the ratio between office and commercial use, the energy consumption ratio between summer, mid-season, and winter, the base power ratio, the total floor area, the energy consumption ratio between electricity, gas, and DHC, the year since construction, the air conditioning system, and the lighting system. In this embodiment, a case will be described where the consumption information is the energy consumption consumed by each individual building as a whole.

[0021] Energy consumption is a value obtained by integrating at least two of the electricity, gas, and oil consumed in a building into one data value. Energy consumption may be calculated as a value of the amount of electricity, such as watt-hours, watt-seconds, or joules, or may be calculated as a value using other known conversion methods, or may include both a value of the amount of electricity and a value using a known conversion method.

[0022] The gas includes at least one of city gas (LNG) and propane gas (LPG). The oil includes at least one of heavy oil A, heavy oil B, heavy oil C, kerosene, and diesel. The objects to be integrated may include substances that generate energy consumed in the building, and may include substances other than the above substances.

[0023] The energy consumption trend is information showing a change in energy consumption over time, and includes waveform information that visualizes the energy consumption over time.

[0024] 1 analyzes consumption information, which is information about energy consumed by an object to be analyzed. The analysis device 10 is an information processing device such as a server having a CPU (central processing unit), a ROM, a RAM, an input / output interface, etc.

[0025] The analysis device 10 is communicably connected to an information processing device that stores consumption information of the building to be analyzed. In this embodiment, the analysis device 10 is communicably connected to a building energy management system 51 (hereinafter also referred to as "BEMS 51") so as to communicate consumption information.

[0026] The analysis device 10 may acquire the consumption information from the BEMS 51, or may acquire the consumption information from a device other than the BEMS 51. For example, the consumption information may be acquired by manual import, from a BAS (Building Automation System: central monitoring device), or from a cloud of another company.

[0027] Manual importing includes a method of transferring consumption information from a building management device to the analysis device 10 using a portable storage medium, and a method of inputting consumption information into the analysis device 10 by typing the consumption information into a keyboard.

[0028] Obtaining consumption information from another company's cloud includes obtaining consumption information for a building managed by a company other than the company that owns the analytical device 10, and includes methods such as obtaining consumption information from a BEMS 51 or BAS owned by another company, or using a portable storage medium to transfer the information from a management device managed by another company to the analytical device 10.

[0029] The BEMS 51 may be provided one-to-one with respect to a building, or one BEMS 51 may be provided for a plurality of buildings. In this embodiment, an example in which one BEMS 51 is provided for one building will be described. Also, an example in which the analysis device 10 is connected to 27 BEMS 51 provided in 27 buildings so as to be able to communicate consumption information will be described.

[0030] The program stored in the storage device such as the ROM mentioned above causes the CPU, ROM, RAM, and input / output interface to work together to function as at least an acquisition unit 11, a preprocessing unit 12, a cluster unit 13, a reclustering unit 14, a model unit 15, an analysis unit 16, a planning unit 17, and an index setting unit 18.

[0031] The acquisition unit 11 has a configuration for acquiring consumption information from a plurality of BEMS 51. The consumption information acquired from the BEMS 51 is created in a format defined for each BEMS 51. The consumption information acquired from the plurality of BEMS 51 may all be created in the same format, or may include consumption information created in different formats.

[0032] The pre-processing unit 12 has a configuration for processing the consumption information acquired by the acquisition unit 11. In this embodiment, a process for converting the format of the consumption information from a format determined for each BEMS 51 to a format that can be processed by the cluster unit 13, the recluster unit 14, the model unit 15, and the analysis unit 16 is performed. The details of the process will be described later.

[0033] The cluster unit 13 has a configuration for categorizing the plurality of pieces of consumption information processed by the pre-processing unit 12 into clusters. In this embodiment, the plurality of pieces of consumption information are categorized into a preset number of clusters based on the mutual similarity of the pieces of consumption information. The details of categorization will be described later.

[0034] The reclustering unit 14 has a configuration for classifying the clusters classified by the clustering unit 13 into at least two or more divided clusters. The classification into the divided clusters is performed based on the similarity of a plurality of pieces of consumption information included in the cluster.

[0035] The model unit 15 has a configuration for creating a model. The model is created using a plurality of different feature quantities related to at least one of the analysis target and the consumption information as explanatory variables and a categorized cluster as a response variable.

[0036] The model is created using a plurality of different feature quantities related to at least one of the analysis target and the consumption information as explanatory variables, and using the divided clusters categorized by the reclustering unit 14 as objective variables. The details of model creation will be described later.

[0037] The analysis unit 16 has a configuration for calculating the importance of a feature, which is the degree of contribution to the construction of a model. The importance is calculated for each of a plurality of feature values ​​used in the model. The method of calculating the importance will be described later in detail.

[0038] The planning unit 17 has a configuration for determining a method for reducing the amount of energy consumption in the analysis target. The reduction method is determined based on the feature amount and the importance of the feature amount determined by the analysis unit 16. The method for determining the reduction method will be described in detail later.

[0039] The index setting unit 18 has a configuration for obtaining an index (also referred to as a benchmark) indicating the energy consumption tendency of the analysis target. The index is obtained based on the feature amount and the importance of the feature amount obtained by the analysis unit 16. The method of obtaining the index will be described later in detail.

[0040] Next, categorization, model creation, analysis, planning, and index setting in the analysis device 10 configured as described above will be described with reference to FIG. When the calculation process in the analysis device 10 is started, the acquisition unit 11 performs a process of acquiring consumption information (S10) as shown in Fig. 2. Specifically, the acquisition unit 11 performs a process of acquiring consumption information from the BEMS 51 connected so as to be capable of communicating information. The acquired consumption information is collected at predetermined intervals for a predetermined period of time. For example, the information is collected at one-hour intervals for a one-year period of time.

[0041] When the acquisition unit 11 acquires the consumption information, the pre-processing unit 12 performs a process of converting the format of the acquired consumption information (S20). The conversion process includes a process of extracting measurement points required for processing by the clustering unit 13, the reclustering unit 14, the modeling unit 15, and the analyzing unit 16. It also includes a name changing process of unifying the names of the required measurement points. It also includes a process of unifying the format of the consumption information into a time series on the vertical axis and a measurement point on the horizontal axis.

[0042] In this embodiment, for each building, a process is performed to convert the consumption information into 24-hour data with the average value calculated by time for the five input data items shown in Fig. 3. Furthermore, a normalization process is performed for each building by dividing the average value by the maximum energy consumption.

[0043] After the format of the consumption information is converted, the clustering unit 13 performs a clustering process to categorize the plurality of pieces of consumption information into clusters (S30). The categorization process is performed based on the mutual similarity between the plurality of pieces of consumption information.

[0044] In this embodiment, the description will be given by applying it to an example where clusters are categorized into three clusters, cluster No. 0, cluster No. 1, and cluster No. 2. The description will be given by applying it to an example where the number of buildings belonging to cluster No. 0 is 1, the number of buildings belonging to cluster No. 1 is 24, and the number of buildings belonging to cluster No. 2 is 2.

[0045] The categorization process may be performed using cluster analysis, which is a machine learning method. A preset number is used as the number of clusters to be categorized. The preset number may be determined using the elbow method for evaluating the number of clusters. The elbow method is a method that considers the number of clusters that does not change the SSE value (sum of squared error within clusters) even if the number of clusters is increased to be the appropriate number of clusters.

[0046] After the clustering process, the reclustering unit 14 performs a reclustering process in which the cluster is divided into at least two or more clusters and classified into smaller divided clusters (S40).

[0047] After the clustering process, the reclustering process S40 may be skipped and the analysis process S41 described later may be performed. The SSE value described above may be used to determine whether to skip the reclustering process S40 and perform the analysis process S41 or not, or the result of an analysis of feature importance described later or the size of a standardized feature may be used.

[0048] The clusters to which the reclustering process is performed are the clusters that are arranged in descending order of the number of pieces of consumption information contained therein, and are ranked from the cluster with the largest number of pieces of consumption information to a predetermined rank. In this embodiment, an example is described in which the predetermined rank is 1st. Note that the predetermined rank may be 2nd or lower.

[0049] The reclustering process is performed based on the mutual similarity between multiple pieces of consumption information, similar to the clustering process. The categorization process may be performed using cluster analysis, which is a machine learning method. A preset number is used as the number of clusters to be categorized.

[0050] In this embodiment, the present invention will be described by applying it to an example where re-clustering processing is performed on cluster No. 1, which has 24 buildings. The present invention will be described by applying it to an example where cluster No. 1 is categorized into two clusters, cluster No. 1a and cluster No. 1b. The present invention will be described by applying it to an example where cluster No. 1a has 12 buildings, and cluster No. 1b has 12 buildings.

[0051] After the reclustering process is performed, the model unit 15 performs a modeling process to create a model (S50). As the modeling process, a machine learning method may be used to model and reproduce the clusters (cluster No. 0, cluster No. 1a, cluster No. 1b, and cluster No. 2 in this embodiment) that are the objective variables using explanatory variables. As the machine learning method, a random forest may be used.

[0052] In this embodiment, the modeling process is performed by using the feature values ​​shown in Fig. 4 as explanatory variables. The feature values ​​shown in Fig. 4 can be created from data obtained from building specifications and energy consumption data. In Fig. 4, district heating and cooling is also referred to as "DHC (District Heating and Cooling)".

[0053] After the modeling process, the analysis unit 16 performs an analysis process to determine the importance of the feature (S60). In the analysis process, the degree to which the feature, which is an explanatory variable used in constructing the model, contributes to the construction of the model is determined. The feature that contributes more is determined to have a higher importance.

[0054] The analysis process may be a calculation process that outputs the importance as a numerical value ranging from 0 to 1. For example, the importance increases as the numerical value approaches 1. A known method may be used as the specific method of the calculation process.

[0055] After the analysis process, the planning unit 17 performs a planning process to determine a method for reducing energy consumption (S70). The reduction method is determined based on the importance of the determined feature. For example, a reduction method linked to a feature with a high importance among the multiple feature amounts used to construct the model may be planned. Alternatively, a combination of multiple reduction methods linked to each of the multiple feature amounts with a high importance may be planned. The linked reduction methods may be stored in advance in the planning unit 17, or may be stored in an external storage device.

[0056] When the planning process is performed, the index setting unit 18 performs a setting process to obtain an index indicating the energy consumption tendency (S80). The index is set based on the importance of the obtained feature. For example, an index linked to a feature with a high importance among the multiple feature amounts used to construct the model may be set. In addition, the planning may be performed by combining multiple indexes linked to multiple feature amounts with high importance. The linked indexes may be stored in advance in the index setting unit 18, or may be stored in an external storage device.

[0057] Next, a description will be given of each cluster classified by the clustering unit 13 and the reclustering unit 14. First, the energy consumption pattern of each cluster will be described with reference to Figs.

[0058] In Figures 5 to 9, the vertical axis represents normalized primary energy consumption, and the horizontal axis represents time. In Figures 5 to 9, the graphs plotted with open circles are graphs for holidays, and the graphs plotted with open squares are graphs for intermediate periods. The graphs plotted with filled circles are graphs for summer, the graphs plotted with filled squares are graphs for winter, and the graphs plotted with crosses (x) are graphs for peak days.

[0059] From Fig. 5, we can see that Cluster No. 0 is a cluster with a pattern in which there is little change in energy consumption depending on the time of day. From Fig. 6, we can see that Cluster No. 1 is a cluster with a pattern in which there is a large difference in energy consumption between day and night. From Fig. 7, we can see that Cluster No. 2 is a cluster with a pattern in which energy consumption is particularly large in the summer. From Figs. 8 and 9, we can see that Cluster No. 1a and Cluster No. 1b have differences in base power and energy consumption in winter.

[0060] Next, the analysis results of feature importance (top 5 items) in each cluster and the size of standardized features will be explained with reference to Figures 10 to 14. In Figures 10(a) to 14(a), the vertical axis shows the top 5 feature items in each cluster, and the horizontal axis shows the importance of each feature.

[0061] In Figures 10(b) to 14(b), the vertical axis shows the top five feature values ​​in each cluster, and the horizontal axis shows the standardized feature value size. In Figures 10(b) to 14(b), the average for all buildings is shown with a white circle (◯), and the average for each cluster is shown with a white square (□).

[0062] Figure 10(a) and (b) show that buildings classified as cluster No. 0 have higher base power ratios and mid-season consumption rates than the average for all buildings. They also have smaller summer consumption rates and total floor area, and an average proportion of offices.

[0063] From Figures 12(a) and (b), we can see that buildings classified into Cluster No. 2 have high summer consumption rates and DHC usage rates. We can also see that the electricity consumption ratio, winter consumption ratio, and intermediate season consumption ratio are small.

[0064] 11(a) and (b) show that buildings classified into cluster No. 1 show values ​​close to the average for all buildings in terms of all feature quantities. Cluster No. 1a and cluster No. 1b, which are subdivisions of cluster No. 1, show differences in the office ratio, intermediate season energy consumption ratio, and winter energy consumption ratio, as shown in Figs. 13 and 14.

[0065] Next, a case where comparison targets are classified into four clusters by the cluster unit 13 will be described with reference to Fig. 15 to Fig. 22. Here, the clusters classified into four by the cluster unit 13 are denoted as Comparison No. 0, Comparison No. 1, Comparison No. 2, and Comparison No. 3, respectively.

[0066] The number of buildings belonging to comparison No. 0 is 1, and the number of buildings belonging to comparison No. 1 is 14. The number of buildings belonging to comparison No. 2 is 11, and the number of buildings belonging to comparison No. 3 is 1.

[0067] Comparing the energy consumption patterns of Comparison No. 1 and Comparison No. 2, they show similar energy consumption patterns, as shown in Figures 16 and 17. In addition, comparing the analysis results of the feature importance and the magnitude of the standardized feature, as shown in Figures 20 and 21, no outstanding feature importance was observed, and the feature magnitude also shows a value close to the average for all buildings.

[0068] In other words, in comparison with the case of Cluster No. 1a and Cluster No. 1b of this embodiment, Comparison No. 1 and Comparison No. 2 are similar clusters in terms of energy consumption characteristics and building characteristics. Furthermore, in Comparison No. 1 and Comparison No. 2, the explanatory factors that caused the buildings to be classified into the respective clusters could not be identified.

[0069] According to the analysis device 10 having the above configuration, a model is created for each cluster categorized based on the similarity of consumption information, and the importance of features, which are explanatory variables of the model, is calculated. Since the importance of features is calculated, it is possible to know features that have a large impact on reducing energy consumption. This makes it easier to plan methods for reducing energy consumption and to set indicators that show energy consumption trends.

[0070] When trying to decarbonize (energy-efficient) multiple buildings owned by a company, it is possible to devise energy-saving measures without analyzing each individual building. In other words, by using a model created for a cluster of the same type as the target building, it is possible to devise methods for reducing energy consumption without analyzing the target building.

[0071] In addition, a group of buildings may include buildings with detailed consumption information and buildings with insufficient (or no) consumption information. In this case, by using the analysis results performed on buildings with detailed consumption information, it is possible to plan energy consumption reduction methods for buildings with insufficient consumption information.

[0072] By linking the classified clusters with detailed consumption information of a specific building (also referred to as a representative building) that belongs to the cluster, it is possible to identify the cluster to which a building with insufficient consumption information belongs and to plan methods for reducing energy consumption.

[0073] The efficiency of data analysis can be improved even for individual buildings, and the number of cases can be significantly increased. In other words, if the cluster to which a building belongs is known, the feature values ​​that have a large impact on reducing energy consumption can be identified. Therefore, data analysis can be performed on the feature values ​​that have a large impact, and data analysis can be made more efficient than when data analysis is performed on all feature values. In addition, the necessary analysis can be performed even without data measurement.

[0074] By using the analysis device 10 of this embodiment, it is possible to reduce the cost of analysis, and it is also possible to expand into the small and medium-sized building market. In conventional technology, the cost of analysis was high and economical, making it difficult to propose measures such as reducing energy consumption to small and medium-sized buildings. In contrast, the analysis device 10 of this embodiment can reduce the cost of analysis, making it easier to propose measures such as reducing energy consumption to small and medium-sized buildings.

[0075] In addition, in the conventional technology, a lot of man-hours were required for the analytical work of finding out the feature quantities that have a large impact on reducing energy consumption. In the energy management of buildings, there are several tens of items to be analyzed. The number of graphs created in the analytical work alone exceeded 100, requiring a lot of man-hours. By using the analysis device 10 of the present embodiment, it is possible to reduce the man-hours required for the analytical work of finding out the importance of such feature quantities.

[0076] In conventional techniques, analysis work requires various types of specialized knowledge, such as data analysis, architecture, facilities, and energy, and there are insufficient human resources to handle the work. By using the analysis device 10 of the present embodiment, the need for various types of specialized knowledge required for analysis work is reduced, making it easier to secure the necessary human resources.

[0077] Categorizing the divided clusters by the reclustering unit 14 makes it easier to clarify the features of the created model compared to increasing the number preset in the clustering unit 13. It becomes easier to understand the features that have a large effect on reducing energy consumption, making it easier to reduce energy consumption.

[0078] A method for reducing the amount of energy consumption is proposed by providing the planner 17. By proposing a method for reducing the amount of energy consumption, it becomes easier to reduce the amount of energy consumption.

[0079] An index showing the energy consumption tendency is presented by providing the index setting unit 18. By presenting an index showing the energy consumption tendency, it becomes easier to reduce the amount of energy consumption.

[0080] In the above embodiment, the analysis target is a group of buildings, and each individual building in the group of buildings is treated as a single unit in the analysis. However, the analysis target may be multiple floors in a building. The multiple floors may belong to a single building, or may be distributed across multiple buildings.

[0081] When the analysis targets multiple floors, each floor is treated as a single unit of analysis. The multiple floors may all belong to one building, or may be distributed across different buildings.

[0082] In this case, the consumption information includes at least one of the energy consumption on individual floors, the set temperature and operating time of the air conditioning equipment on individual floors, data on the illuminance and operating time of the lighting equipment on individual floors, the consumption ratio of energy consumption by the lighting equipment, the consumption ratio of energy consumption by night-time lighting, the consumption ratio of energy consumption by the air conditioning equipment, the consumption ratio of energy consumption by night-time air conditioning, and the number of occupants.

[0083] The energy consumption is a value obtained by integrating at least two of the electricity, gas, oil, and DHC hot and cold energy consumed by the floor into one data value. The energy consumption includes the value of the amount of electricity such as watt-hours, watt-seconds, and joules, the energy value calculated using a known conversion method, and the value of the amount of heat of hot and cold energy.

[0084] Furthermore, the analysis target may be multiple air conditioning equipment in a building. Also, the analysis target may be any of the facility equipment including multiple air conditioning equipment installed in a building. When the analysis target is multiple air conditioning equipment, each individual air conditioning equipment in the multiple air conditioning equipment is treated as one unit in the analysis. The multiple air conditioning equipment may all belong to one building, or may be distributed across different buildings.

[0085] In this case, the consumption information includes at least one of the energy consumption consumed by individual air conditioning equipment, the set temperature, the indoor temperature which is the indoor unit suction temperature of the air conditioning equipment, operating time, coefficient of performance (COP), load factor, heating and cooling capacity (amount of heat processed by heating and cooling), hot and cold water temperature, refrigerant temperature data, indoor carbon dioxide concentration, hot and cold water valve opening, VAV (Variable air Volume system) required air volume, inverter frequency, and the number of people present in the room.

[0086] Energy consumption is a value calculated by integrating at least two of the following energy sources consumed in a building: electricity, gas, oil, and DHC heating and cooling. Air conditioning equipment includes individual PAC (packaged) air conditioners, air handling units, and fan coil units.

[0087] In the classification by the cluster unit 13 and the recluster unit 14, classification may be performed based on performance such as the load rate and efficiency of the air conditioning equipment and facility devices. In this case, classification into a cluster with a high load rate, a cluster with a standard load rate, and a cluster with a low load rate, or into a cluster with high efficiency, a cluster with standard efficiency, and a cluster with low efficiency, may be performed.

[0088] Furthermore, by providing items that affect the load factor and efficiency in the feature quantities, it is possible to use the analysis device 10 to study the optimization of the capacity of air conditioning facilities and facility devices.

[0089] The technical scope of the present invention is not limited to the above-mentioned embodiment, and various modifications can be made without departing from the spirit of the present invention. For example, the present invention is not limited to the above-mentioned embodiment, and may be applied to an embodiment in which these embodiments are appropriately combined, and is not particularly limited. [Explanation of symbols]

[0090] 10...analysis device, 13...clustering unit, 14...reclustering unit, 15...modeling unit, 16...analysis unit, 17...planning unit, 18...index setting unit

Claims

1. An analysis device that analyzes consumption information, which is information regarding energy consumed by an analysis subject, comprising: A clustering unit that classifies the plurality of pieces of consumption information into a predetermined number of clusters based on the similarity between the pieces of consumption information; a model unit that creates a model using a plurality of different feature quantities related to at least one of the analysis target and the consumption information as explanatory variables and the classified cluster as a response variable; an analysis unit that calculates an importance level, which is a degree of contribution to construction of the model, for each of the plurality of feature amounts used in the model; An analytical device comprising:

2. a reclustering unit that arranges the clusters in order of the number of pieces of consumption information included therein, and categorizes each of the clusters in a predetermined order from the cluster with the largest number of pieces of consumption information into at least two or more divided clusters based on the mutual similarity of the pieces of consumption information included in the cluster, 2. The analysis device according to claim 1, wherein the model unit also creates a model for the divided clusters.

3. 3. The analysis device according to claim 1, further comprising a planning unit that determines a method of reducing the amount of energy consumed in the analysis target based on the characteristic amount and the importance of the characteristic amount.

4. 3. The analysis device according to claim 1, further comprising an index setting unit that determines an index indicating the energy consumption tendency of the analysis target based on the feature amount and the importance of the feature amount.

5. A program for causing a computer to analyze consumption information, which is information regarding energy consumed by an analysis target, comprising: The computer includes: A cluster classification function for classifying a plurality of pieces of consumption information into a predetermined number of clusters based on the similarity between the pieces of consumption information; a model creation function for creating a model using a plurality of different feature quantities related to at least one of the analysis target and the consumption information as explanatory variables and using the classified cluster as a response variable; an analysis function for determining the importance of each of the plurality of feature quantities used in the model, the importance being the degree to which the feature quantities contribute to the construction of the model; A program for realizing the above.

Images