Non-deterministic energy consumption prediction method and related device under the condition of missing input variables

Abstract

The present application discloses a non-deterministic energy consumption prediction method and related devices in the case of missing input variables, including: analyzing whether there is an association rule relationship between each known input variable and an unknown input variable in a database according to the association rules defined by the user through the apriori algorithm; when there is an association rule relationship between the known input variable and the unknown input variable, the distribution prediction of the unknown input variable is performed according to the association rules between the two, otherwise, a uniform distribution is established according to the value range of the unknown input variable, so as to obtain the distribution interval of the unknown input variable; sampling the unknown input variable based on the distribution interval to obtain a set of unknown input variables; inputting the set of unknown input variables into an energy consumption model for batch calculation to obtain the distribution range of the energy consumption prediction value; thereby solving the technical problem of low energy consumption prediction accuracy in the prior art when the input variables are missing.

Description

Technical Field

[0001] The present application relates to the technical field of energy consumption prediction, and in particular to a method and related device for predicting non-deterministic energy consumption in the absence of input variables. Background Art

[0002] Energy consumption simulation can help users simulate and analyze the energy consumption of building construction. There are two main methods to establish the model, namely forward model and data-driven model. The forward model, also known as the white box model, is established based on the physical characteristics and physical laws of the object system (such as heat and mass balance, conservation of momentum and mass, etc.). Establishing a forward model requires an in-depth understanding of the object characteristics, starting from simple relationships to build a complex model that describes the entire system. The data-driven model is also called a black box model. It is based on a large amount of test data generated during the operation of the object system. It uses statistical methods to find the mapping relationship between variables and outputs without understanding the working principle of the system. Therefore, it is also called an inverse model.

[0003] However, in actual energy consumption prediction scenarios, the values ​​of all input variables cannot be accurately obtained in many cases. For example, building a white box model requires some input parameters, and it is difficult to fully obtain relevant input parameters such as convergence tolerance, solar radiation intensity, material roughness, etc. from drawings and design information even in the design stage, not to mention the completed buildings. Because the construction and design information has been lost over the years, these data are basically difficult to obtain or determine, resulting in the lack of input variable information, which in turn leads to inaccurate energy consumption prediction. There is still a lack of more scientific means to improve the problem of missing eigenvalues ​​when using energy consumption software, but most of them are determined based on literature or manual experience. However, literature and manual experience have limitations and cannot be applied to all situations, resulting in inaccurate filling data and reduced energy consumption prediction accuracy. Summary of the invention

[0004] The present application provides a non-deterministic energy consumption prediction method and related devices in the case of missing input variables, which are used to solve the technical problem of low energy consumption prediction accuracy in the prior art in the case of missing input variables.

[0005] In view of this, the first aspect of the present application provides a method for predicting non-deterministic energy consumption when input variables are missing, the method comprising:

[0006] Through the apriori algorithm, according to the association rules defined by the user, analyze whether there is an association rule relationship between each known input variable and the unknown input variable in the database;

[0007] When there is an association rule relationship between the known input variables and the unknown input variables, the distribution prediction of the unknown input variables is performed according to the association rule between the two. Otherwise, a uniform distribution is established according to the value range of the unknown input variables to obtain the distribution interval of the unknown input variables.

[0008] Sampling the unknown input variables based on the distribution interval to obtain a set of unknown input variables;

[0009] The unknown input variable set is input into the energy consumption model for batch calculation to obtain the distribution range of the energy consumption prediction value.

[0010] Optionally, the method of analyzing whether there is an association rule relationship between each known input variable and an unknown input variable in the database according to an association rule defined by a user by using an apriori algorithm also includes: discretizing continuous variables in the database.

[0011] Optionally, the analyzing whether there is an association rule relationship between each known input variable and the unknown input variable in the database according to the association rule defined by the user by using the apriori algorithm specifically includes:

[0012] All frequent item sets that meet the preset threshold in the database are found through iteration, and the association rules that meet the user's minimum trust are constructed using the frequent item sets, so as to analyze whether there is an association rule relationship between each known input variable and the unknown input variable in the database.

[0013] Optionally, sampling the unknown input variables based on the distribution interval to obtain the unknown input variable set specifically includes:

[0014] Based on the distribution interval, unknown input variables are sampled by Morris analysis to obtain a set of unknown input variables.

[0015] A second aspect of the present application provides a non-deterministic energy consumption prediction system in the case of missing input variables, the device comprising:

[0016] An analysis unit, used to analyze whether there is an association rule relationship between each known input variable and an unknown input variable in the database according to the association rule defined by the user through the apriori algorithm;

[0017] Establishing a unit, used for predicting the distribution of the unknown input variable according to the association rule between the known input variable and the unknown input variable when there is an association rule relationship between the two, otherwise, establishing a uniform distribution according to the value range of the unknown input variable, thereby obtaining the distribution interval of the unknown input variable;

[0018] A sampling unit, used for sampling the unknown input variables based on the distribution interval to obtain a set of unknown input variables;

[0019] The prediction unit is used to input the unknown input variable set into the energy consumption model for batch calculation to obtain the distribution range of the energy consumption prediction value.

[0020] Optionally, it further includes: a pre-processing unit;

[0021] The preprocessing unit is used to discretize the continuous variables in the database.

[0022] Optionally, the analysis unit is specifically used to:

[0023] All frequent item sets that meet the preset threshold in the database are found through iteration, and the association rules that meet the user's minimum trust are constructed using the frequent item sets, so as to analyze whether there is an association rule relationship between each known input variable and the unknown input variable in the database.

[0024] Optionally, the sampling unit is specifically used to:

[0025] Based on the distribution interval, unknown input variables are sampled by Morris analysis to obtain a set of unknown input variables.

[0026] A third aspect of the present application provides a non-deterministic performance consumption prediction device in the case of missing input variables, the device comprising a processor and a memory:

[0027] The memory is used to store program codes and transmit the program codes to the processor;

[0028] The processor is used to execute the steps of the non-deterministic performance consumption prediction method in the case of missing input variables as described in the first aspect according to the instructions in the program code.

[0029] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store program code, and the program code is used to execute the non-deterministic performance consumption prediction method in the case of missing input variables described in the first aspect above.

[0030] It can be seen from the above technical solutions that this application has the following advantages:

[0031] The energy consumption prediction method of the present application first discretizes the continuous variables, then extracts the association rules according to the above algorithm steps, establishes the association between the known variables and the unknown variables, and after the item set of unknown features is inferred from the known variables, the continuous value range of the unknown variables is restored, so that the originally missing unknown variable data can be obtained through abstract mining methods; and when there is no association rule relationship between the known input variables and the unknown input variables, due to the lack of distribution information of the unknown features, it is assumed that it is uniformly distributed, and then a number of sampling points of unknown variables are obtained by random sampling from the distribution, and all sampling results are substituted into the energy consumption model for calculation and prediction, and the distribution range of the energy consumption prediction value can be obtained. Different from conventional energy consumption simulation, the energy consumption prediction method of the present application determines the distribution interval of the unknown variables, and the energy consumption prediction result given at the end is not a certain value, but an interval. It has great reference significance for energy consumption simulation, especially energy consumption simulation with unknown input features, and has a wider applicability. The prediction results are rich in more information and are more conducive to helping users make decisions. Thereby solving the technical problem of low energy consumption prediction accuracy in the prior art when input variables are missing. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A schematic diagram of a flow chart of an embodiment of a method for predicting non-deterministic energy consumption in the case of missing input variables provided in an embodiment of the present application;

[0033] Figure 2 This is a structural diagram of an embodiment of a non-deterministic energy consumption prediction system in the case of missing input variables provided in an embodiment of the present application. DETAILED DESCRIPTION

[0034] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0035] The following is an explanation of the inventive principle of this application:

[0036] In order to obtain the non-deterministic interval of the energy consumption prediction value, this application mainly supplements the information of the unknown input variables based on the association rule mining algorithm (Apriori) to determine the unknown information. The basis of algorithm design is to consider the association rules between building features, such as the relatively high lighting power density of buildings built in the early stage, and the low efficiency of equipment aging. Then after data mining, the year of building completion will be associated with the lighting power density. The association here exists in the form of probability. The following is a detailed introduction to the apriori algorithm; the performance of buildings that have undergone energy-saving renovations has been improved in all aspects. If they have undergone energy-saving renovations, we can infer that their building performance belongs to a relatively good range under certain circumstances; the low efficiency of equipment such as chillers and water pumps indicates a low level of operation and maintenance management, so it can be inferred that the efficiency of other equipment is also relatively low, etc.; Therefore, such rules can be packaged into an association rule library, and owners and managers can add appropriate association rules according to their own circumstances.

[0037] See also Figure 1 , a non-deterministic performance consumption prediction method in the case of missing input variables provided in an embodiment of the present application includes:

[0038] Step 101: Analyze whether there is an association rule relationship between each known input variable and the unknown input variable in the database according to the association rule defined by the user through the apriori algorithm;

[0039] The present application first uses the apriori algorithm to screen out strong association rule relationships between the complete input variable data already in the database and the unknown input variables according to the association rules defined by the owner or manager.

[0040] It should be noted that the Apriori algorithm is a method proposed by R. Agrawal and R. Srikant to find frequent itemsets of Boolean association rules in a data set. It can use prior knowledge to predict the association rules of data, and its name Apriori also comes from this.

[0041] First, we need to define several basic concepts of association analysis:

[0042] (1) k-itemset: An event containing K elements that satisfies a certain minimum support threshold is called a frequent k-itemset.

[0043] (2) Support: The support of association rule A→B refers to the probability that events A and B (the intersection of events A and B is empty) occur at the same time, support = P(AB);

[0044] (3) Confidence: refers to the probability that event B will occur based on event A (the intersection of events A and B is empty), confidence = P(B|A) = P(AB) / P(A);

[0045] Briefly speaking, the Apriori algorithm is divided into two steps:

[0046] 1. Through iteration, find all frequent item sets in the database, and the threshold for determining the frequent item sets is set by the user himself;

[0047] 2. Use the frequent item sets to construct rules that meet the minimum confidence of the user.

[0048] The specific steps include: First, scan the data set once to generate the 1-item set C1 from it. Then call the Scan function to scan C1 and filter out the item sets that do not meet the minimum support. The finally remaining item sets are the frequent item sets L1. In the second round of iteration, only new combinations need to be made for the frequent item sets generated in the previous round of iteration, and then call the Scan function to check whether the support of the new combinations meets the minimum support requirement, and filter out the new combinations that do not meet. Loop like this until no new combinations can be generated. Among them, C1, C2, …, Ck represent the 1-item set, 2-item set, …, k-item set respectively; L1, L2, …, Lk represent the frequent item sets after “filtering” of the corresponding item sets; Scan: represents the data item scanning function, which filters out the item sets that do not meet the minimum support. The joining step: There are two cases. The first is to generate C1 from the data set, and the second is to generate Ck according to Lk-1. Briefly speaking, the joining step is the process of generating item sets. The pruning step: Eliminate the item sets that do not meet the minimum support.

The effectiveness of this process is because, if P(A) >= P(AB), if P(A) < t, then there must be P(AB) < t, where t is the support threshold

[0049] Step 102: When there is an association rule relationship between the known input variable and the unknown input variable, predict the distribution of the unknown input variable according to the association rule between the two. Otherwise, establish a uniform distribution according to the value range of the unknown input variable, so as to obtain the distribution interval of the unknown input variable;

[0050] It can be understood that if there is a certain association between the unknown variable and the known variable in the association rule library, predict the distribution of the unknown feature according to the rule between the two. Otherwise, establish a uniform distribution according to the value range of the unknown variable in the database, so as to obtain the distribution interval of the unknown input variable.

[0051] Step 103: Sample the unknown input variable based on the distribution interval to obtain the unknown input variable set;

[0052] In this embodiment, specifically, sample the unknown input variable based on the distribution interval through the Morris analysis method to obtain the unknown input variable set.

[0053] It should be noted that the Morris method, also known as the meta-effect method, is a simple and effective method that can screen out some important input factors from the many input factors contained in the model. The definition of a meta-effect is as follows:

[0054] Assume that a model has k independent inputs Xi, i = 1, ..., k, and these independent variables change in a k-dimensional unit body with a change level of p, that is, the input variable space is discretized into a p-level grid space Ω. For a given input variable X, the element effect of the i-th dimension is defined as:

[0055]

[0056] Among them, the value of Δ is taken from {1p-1,…,1-1p-1}, X+eiΔ is still in Ω, and ei is a vector that is 1 in the i-th dimension and 0 in the other dimensions.

[0057] The meta-effect distribution of the i-th dimension input variable is obtained by random sampling of X in the Ω space, which is represented by Fi, that is, EEi~Fi. The two sensitivity indicators proposed by Morris, μ and σ, are the mean and standard deviation of the Fi distribution. μ is an estimate of the overall impact of the variable on the result, while σ is an estimate of the comprehensive impact of the variable. The larger the σ, the greater the impact of the variable on the output is affected by the values ​​of other variables. Since some effects will produce positive and negative offsets when calculating the μ value when Fi contains positive and negative values, thereby underestimating the sensitivity of an important factor, the μ* proposed by Campolongo is often used instead of μ. μ* is the mean of the absolute value distribution of the meta-effect. This meta-effect distribution is defined as Gi, |EEi|~Gi.

[0058] Step 104: Input the unknown input variable set into the energy consumption model for batch calculation to obtain the distribution range of the energy consumption prediction value.

[0059] Finally, all sampling results are substituted into the energy consumption model for batch calculation to obtain the distribution range of the energy consumption prediction value.

[0060] Furthermore, in an optional embodiment, considering that the existing Apriori algorithm is generally only applicable to discrete variables, in order to be applicable to continuous variables, the present application first discretizes the continuous variables, and then extracts association rules according to the above algorithm steps, establishes the association between known variables and unknown variables, and after inferring the item set of unknown features from the known variables, restores the continuous value range of the unknown variables, so that the originally missing unknown variable data can be obtained through abstract mining methods. When there is no association rule relationship between the known input variables and the unknown input variables, due to the lack of distribution information of the unknown features, it is assumed to be uniformly distributed, and then a number of sampling points of the unknown variables are obtained by randomly sampling from the distribution, and all sampling results are substituted into the energy consumption model for calculation and prediction, and the distribution range of the energy consumption prediction value can be obtained.

[0061] For example, through association rule mining, it is found that there is a correlation between material thickness and thermal conductivity. When the material thickness is a, the probability that the thermal conductivity is p is b. Then, when sampling, there is a probability of p that the thermal conductivity = b, and the probability of 1-p is not b.

[0062] The non-deterministic energy consumption prediction method provided by the present application in the case of missing input variables first discretizes the continuous variables, then extracts the association rules according to the above algorithm steps, establishes the association between the known variables and the unknown variables, and after inferring the item set of unknown features from the known variables, restores the continuous value range of the unknown variables, so that the originally missing unknown variable data can be obtained through abstract mining methods; and when there is no association rule relationship between the known input variables and the unknown input variables, due to the lack of distribution information of the unknown features, it is assumed to be uniformly distributed, and then a number of sampling points of the unknown variables are obtained by random sampling from the distribution, and all the sampling results are substituted into the energy consumption model for calculation and prediction, so that the distribution range of the energy consumption prediction value can be obtained. This solves the technical problem of low energy consumption prediction accuracy in the case of missing input variables in the prior art.

[0063] The above is a non-deterministic energy consumption prediction method in the case of missing input variables provided in an embodiment of the present application, and the following is a non-deterministic energy consumption prediction system in the case of missing input variables provided in an embodiment of the present application.

[0064] See also Figure 2 , a non-deterministic energy consumption prediction system in the case of missing input variables provided in an embodiment of the present application includes:

[0065] An analysis unit 201 is used to analyze whether there is an association rule relationship between each known input variable and an unknown input variable in the database according to the association rule defined by the user through the apriori algorithm;

[0066] Establishing unit 202, used for predicting the distribution of the unknown input variable according to the association rule between the known input variable and the unknown input variable when there is an association rule relationship between the known input variable and the unknown input variable; otherwise, establishing a uniform distribution according to the value range of the unknown input variable, thereby obtaining the distribution interval of the unknown input variable;

[0067] A sampling unit 203, used for sampling unknown input variables based on a distribution interval to obtain a set of unknown input variables;

[0068] The prediction unit 204 is used to input the unknown input variable set into the energy consumption model for batch calculation to obtain the distribution range of the energy consumption prediction value.

[0069] Furthermore, an embodiment of the present application also provides a non-deterministic performance consumption prediction device in the case of missing input variables, the device comprising a processor and a memory:

[0070] The memory is used to store program codes and transmit the program codes to the processor;

[0071] The processor is used to execute the non-deterministic performance consumption prediction method in the case of missing input variables described in the above method embodiment according to the instructions in the program code.

[0072] Furthermore, a computer-readable storage medium is provided in an embodiment of the present application, and the computer-readable storage medium is used to store program code, and the program code is used to execute the non-deterministic performance consumption prediction method in the absence of input variables described in the above method embodiment.

[0073] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0074] The terms "first", "second", "third", "fourth", etc. in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein, for example. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0075] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0076] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0077] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0078] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0079] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (full name in English: Read-Only Memory, English abbreviation: ROM), random access memory (full name in English: Random Access Memory, English abbreviation: RAM), disk or optical disk and other media that can store program codes.

[0080] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for predicting non-deterministic energy consumption in the absence of input variables, characterized in that: include: All frequent item sets that meet the preset threshold in the database are found through iteration, and association rules that meet the user's minimum trust are constructed using the frequent item sets, so as to analyze whether there is an association rule relationship between each known input variable and the unknown input variable in the database; When there is an association rule relationship between the known input variables and the unknown input variables, the distribution prediction of the unknown input variables is performed according to the association rule between the two. Otherwise, a uniform distribution is established according to the value range of the unknown input variables to obtain the distribution interval of the unknown input variables. Based on the distribution interval, sampling the unknown input variables by Morris analysis method to obtain a set of unknown input variables; The unknown input variable set is input into the energy consumption model for batch calculation to obtain the distribution range of the energy consumption prediction value.

2. The method for predicting non-deterministic energy consumption in the case of missing input variables according to claim 1, characterized in that: The apriori algorithm is used to analyze whether there is an association rule relationship between each known input variable and the unknown input variable in the database according to the association rules defined by the user. Previously, it also includes: discretizing the continuous variables in the database.

3. A non-deterministic energy consumption prediction system in the case of missing input variables, characterized in that: include: An analysis unit is used to find all frequent item sets that meet a preset threshold in the database through iteration, and use each of the frequent item sets to construct an association rule that meets the user's minimum trust, so as to analyze whether there is an association rule relationship between each known input variable and an unknown input variable in the database; Establishing a unit, used for predicting the distribution of the unknown input variable according to the association rule between the known input variable and the unknown input variable when there is an association rule relationship between the two, otherwise, establishing a uniform distribution according to the value range of the unknown input variable, thereby obtaining the distribution interval of the unknown input variable; A sampling unit, used for sampling unknown input variables by Morris analysis method based on the distribution interval to obtain a set of unknown input variables; The prediction unit is used to input the unknown input variable set into the energy consumption model for batch calculation to obtain the distribution range of the energy consumption prediction value.

4. The non-deterministic energy consumption prediction system in the case of missing input variables according to claim 3, characterized in that: Also includes: Pre-processing unit; The preprocessing unit is used to discretize the continuous variables in the database.

5. A non-deterministic energy consumption prediction device in the case of missing input variables, characterized in that: The device comprises a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the non-deterministic performance consumption prediction method in the case of missing input variables as described in any one of claims 1-2 according to the instructions in the program code.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store program code, and the program code is used to execute the non-deterministic performance consumption prediction method in the case of missing input variables as described in any one of claims 1-2.

Images