Intelligent laboratory carbon emission monitoring and early warning method

Abstract

The invention discloses an intelligent laboratory carbon emission monitoring and early warning method, which is used for realizing early warning on a carbon emission abnormal state caused by laboratory abnormity and improving the risk management and control level of intelligent laboratory carbon emission. The method comprises the steps of collecting equipment energy consumption data based on an intelligent laboratory carbon emission monitoring network, and determining an intelligent laboratory carbon emission monitoring database; determining a to-be-processed data set according to the intelligent laboratory carbon emission monitoring database; preprocessing the to-be-processed data set to obtain a carbon emission monitoring feature library; constructing a cascade depth radial network based on the carbon emission monitoring feature library; based on the carbon emission monitoring feature library, training the cascade depth radial network until the network error is less than a preset target error, and obtaining a carbon emission monitoring early warning model; based on a carbon emission monitoring and early warning model, determining a carbon emission value according to the actual carbon emission monitoring feature vector; judging whether the carbon emission value exceeds an early warning standard or not; and if the carbon emission exceeds the early warning standard, carrying out carbon emission standard exceeding early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing and pattern recognition applications, and in particular to an intelligent laboratory carbon emission monitoring and early warning method. Background Art

[0002] As an important carrier of high-tech and high-tech industries, smart laboratories play an important role in serving the country's high-quality development strategy. The construction and operation of low-carbon smart laboratories has become an important demand. Due to the particularity of their work, laboratories are large users of water, energy and waste. According to statistics, the electricity consumed by laboratories is 10 times that of ordinary office buildings, and the water is 4 times that of ordinary office buildings, and a large amount of waste is also generated. Building "green smart laboratories and low-carbon smart laboratories" has become a new path for the development of laboratories. The carbon emission index operation monitoring standard for smart laboratories aims to promote the total energy consumption and intensity of smart laboratories to be better than conventional laboratories through a series of measures and management requirements, and achieve the goal of low-carbon emissions. At the same time, improve the energy utilization efficiency of smart laboratories, reduce energy consumption and operating costs, improve the competitiveness of laboratories, reduce negative impacts on the environment by reducing greenhouse gas emissions and improving energy utilization efficiency, and achieve more efficient and environmentally friendly operations.

[0003] It can be seen from this that how to achieve early warning of abnormal carbon emission conditions caused by laboratory abnormalities, and thus effectively improve the risk control level of carbon emissions in smart laboratories has become one of the urgent issues to be solved. Summary of the invention

[0004] The present invention provides a smart laboratory carbon emission monitoring and early warning method for realizing early warning of abnormal carbon emission status caused by laboratory abnormalities, thereby effectively improving the risk control level of carbon emissions in smart laboratories.

[0005] In a first aspect, a smart laboratory carbon emission monitoring and early warning method is provided, comprising:

[0006] Based on the smart laboratory carbon emission monitoring network, collect equipment energy consumption data and determine the smart laboratory carbon emission monitoring database;

[0007] Determine the data set to be processed based on the smart laboratory carbon emission monitoring database;

[0008] Preprocess the data set to be processed to obtain a carbon emission monitoring feature library;

[0009] Based on the carbon emission monitoring feature library, a cascade deep radial network is constructed;

[0010] Based on the carbon emission monitoring feature library, the cascade deep radial network is trained until the network error is less than the preset target error, and the training stops to obtain the carbon emission monitoring and early warning model.

[0011] Based on the carbon emission monitoring and early warning model, the carbon emission value is determined according to the actual carbon emission monitoring feature vector.

[0012] In one embodiment, the equipment energy consumption data includes server energy consumption, air conditioning facility energy consumption, lighting facility energy consumption, elevator energy consumption, fume hood energy consumption, test box energy consumption, switch energy consumption, router energy consumption, water energy consumption, heat energy consumption, greenhouse gas emissions, waste emissions, and vehicle energy consumption.

[0013] In one embodiment, based on the smart laboratory carbon emission monitoring network, equipment energy consumption data is collected to determine the smart laboratory carbon emission monitoring database, including:

[0014] Based on the intelligent laboratory carbon emission monitoring network, collecting the energy consumption data of the first preset number of groups of equipment;

[0015] The energy consumption data of each group of equipment is structured and the carbon emission observation vector is determined by the following formula:

[0016] SO=(pm,pd,pt,pl,bdsl,sp,pi,po,ti,to,sr,sa)

[0017] Among them, SO is the carbon emission observation vector, pm, pd, pt, pl, bd, sl, sp, pi, po, ti, to, sr, and sa are server energy consumption, air conditioning facility energy consumption, lighting facility energy consumption, elevator energy consumption, fume hood energy consumption, laboratory box energy consumption, switch energy consumption, router energy consumption, water energy consumption, heat energy consumption, greenhouse gas emissions, waste emissions, and vehicle energy consumption, respectively;

[0018] Based on a first preset number of carbon emission observation vectors, a smart laboratory carbon emission monitoring database is determined.

[0019] In one implementation, the row vector of the data set to be processed is a first preset number of carbon emission observation vectors.

[0020] In one embodiment, the column vector of the data set to be processed is a value set of carbon emission observation components, which is 13 sets in total.

[0021] In one embodiment, the data set to be processed is preprocessed to obtain a carbon emission monitoring feature library, including:

[0022] Each column vector of the data set to be processed is dedimensionalized, and the processed column vector is determined by the following formula:

[0023] Mc′=(Mc-mean(Mc)). / std(Mc)

[0024] Among them, Mc′ is the processed column vector, Mc is the column vector of the data set to be processed, mean() is the vector mean function, std() is the vector standard deviation function, and . / is the vector dot division operation;

[0025] Each row vector of the data set to be processed is centralized, and the processed row vector is determined by the following formula:

[0026] Mr′=(Mr-min(Mr)). / (max(Mr)-min(Mr))

[0027] Among them, Mr′ is the processed row vector, Mr is the row vector of the data set to be processed, min() is the minimum function in the column direction, and max() is the maximum function in the column direction;

[0028] Based on the first preset number of processed row vectors, the carbon emission monitoring feature library is determined by the following formula:

[0029] SF=Mr′

[0030] MF={SF k} k＝1,…,N

[0031] Among them, SF is the carbon emission monitoring feature vector, MF is the carbon emission monitoring feature library, and N is the first preset number.

[0032] In one embodiment, based on the carbon emission monitoring feature library, a cascade deep radial network is constructed, including:

[0033] The carbon emissions corresponding to the first preset number of carbon emission monitoring feature vectors are standardized respectively, and the carbon emission feature vector is determined by the following formula:

[0034] DF=stdnorml(DQ,[0,1]

[0035] Among them, DF is the carbon emission feature vector, DQ is the value set of carbon emission, and stdnorml() is the vector standard normalization function. According to the carbon emission feature vector and the carbon emission monitoring feature library, the single-layer radial network is determined by the following formula:

[0036] NS = newrb(MF, DF, es)

[0037] Among them, NS is a single-layer radial network, newrb() is a single-layer radial network construction function, and es is the network fitting error;

[0038] According to the single-layer radial network, a cascaded deep radial network is constructed by the following formula:

[0039] NC=casrb(NS,ec,gl)

[0040] Among them, NC is the cascaded deep radial network, casrb() is the cascaded deep radial network construction function, ec is the deep training step, and gl is the target error.

[0041] In one embodiment, based on the carbon emission monitoring feature library, the cascade deep radial network is trained until the network error is less than the preset target error and the training stops, thereby obtaining a carbon emission monitoring and early warning model, including:

[0042] Dividing the first preset number group of carbon emission monitoring feature vectors and their corresponding carbon emission feature vectors into a second preset number group of training data sets and a third preset number group of testing data sets;

[0043] Based on the second preset number of training data sets and the third preset number of test data sets, the cascaded deep radial network is trained until the network error is less than the preset target error, and the training stops to obtain a carbon emission monitoring and early warning model.

[0044] In one implementation, the sum of the second preset number and the third preset number is the first preset number.

[0045] In one embodiment, based on the carbon emission monitoring and early warning model, the carbon emission value is determined according to the actual carbon emission monitoring feature vector by the following formula:

[0046] rd = ssim(NC′, rf)

[0047] Among them, rd is the carbon emission value, NC′ is the carbon emission monitoring and early warning model, rf is the actual carbon emission monitoring feature vector, and ssim() is the deep network prediction function.

[0048] The embodiment of the present invention provides a smart laboratory carbon emission monitoring and early warning method, which combines the methods of IoT intelligent perception, big data analysis and machine learning to establish a smart laboratory carbon emission monitoring and early warning model, monitor and analyze the smart laboratory carbon emission in real time, and issue early warnings for abnormal carbon emission situations. Through the process of building a smart laboratory carbon emission monitoring network, data collection, model building, training and verification, and actual deployment and application, a scientific, effective and accurate smart laboratory carbon emission monitoring and early warning method is realized.

[0049] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0051] Figure 1 The present invention is a flowchart of a smart laboratory carbon emission monitoring and early warning method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0052] In order to provide early warning for abnormal carbon emission status caused by laboratory abnormalities and effectively improve the risk management level of carbon emissions in smart laboratories, an embodiment of the present invention provides an intelligent laboratory carbon emission monitoring and early warning method.

[0053] It should be noted that the terms "first", "second", etc. in the description and claims of the embodiments of the present invention and the above 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 terms used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein.

[0054] The preferred embodiments of the present invention are described below in conjunction with the drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention. In addition, the embodiments of the present invention and the features in the embodiments may be combined with each other if there is no conflict.

[0055] Embodiment 1

[0056] An intelligent laboratory carbon emission monitoring and early warning method provided by an embodiment of the present invention is shown in Figure 1 , including the following steps:

[0057] S11. Based on the smart laboratory carbon emission monitoring network, collect equipment energy consumption data and determine the smart laboratory carbon emission monitoring database;

[0058] S12. Determine the data set to be processed according to the smart laboratory carbon emission monitoring database;

[0059] S13, preprocessing the data set to be processed to obtain a carbon emission monitoring feature library;

[0060] S14. Construct a cascade deep radial network based on the carbon emission monitoring feature library;

[0061] S15. Based on the carbon emission monitoring feature library, the cascade deep radial network is trained until the network error is less than the preset target error, and the training stops to obtain a carbon emission monitoring and early warning model;

[0062] S16. Based on the carbon emission monitoring and early warning model, determine the carbon emission value according to the actual carbon emission monitoring feature vector;

[0063] S17, judging whether the carbon emission value exceeds the warning standard;

[0064] S18. If the warning standard is exceeded, a carbon emission exceeding standard warning will be issued.

[0065] In the specific implementation, first, we will build a smart laboratory carbon emission monitoring network, collect data, and establish a smart laboratory carbon emission monitoring database.

[0066] Specifically, the working area of ​​the smart laboratory is selected to build a smart laboratory carbon emission monitoring network. In order to use big data analysis and machine learning methods to comprehensively evaluate the carbon emission intensity of the smart laboratory, the carbon emission intensity and operating status of the smart laboratory are monitored in real time based on the deployment of relevant IoT smart sensor devices for the monitored energy-consuming objects. The monitored quantities include: energy consumption of electrical equipment, energy consumption of communication facilities, as well as water consumption, heat consumption, greenhouse gas emissions, waste emissions, etc.

[0067] Electrical equipment and communication equipment can also be subdivided into: server energy consumption, lighting facility energy consumption, elevator energy consumption, fume hood energy consumption, test box energy consumption, switch energy consumption, router energy consumption, etc.

[0068] By collecting data on energy consumption of electrical equipment, water consumption, greenhouse gas emissions, etc., a laboratory carbon emission monitoring database is established. A single data entry is composed of labels such as server energy consumption, air-conditioning facility energy consumption, lighting facility energy consumption, elevator energy consumption, fume hood energy consumption, laboratory box energy consumption, switch energy consumption, router energy consumption, water consumption, heat consumption, greenhouse gas emissions, waste emissions, and vehicle energy consumption. After structuring, the carbon emission observation vector SO is obtained.

[0069] SO=(pm,pd,pt,pl,bd,sl,sp,pi,po,ti,to,sr,sa) (1)

[0070] Wherein, PM, PD, PT, PL, BD, SL, SP, PI, PO, TI, TO, SR, and SA are server energy consumption, air-conditioning facility energy consumption, lighting facility energy consumption, elevator energy consumption, fume hood energy consumption, laboratory box energy consumption, switch energy consumption, router energy consumption, water consumption, heat consumption, greenhouse gas emissions, waste emissions, and vehicle energy consumption, respectively.

[0071] In the actual test environment, N groups of SO are collected to form a sedimentation safety monitoring database, namely the smart laboratory carbon emission monitoring database {SO k} k＝1,…,N , typically, N=1000.

[0072] On this basis, the carbon emission observation vector is preprocessed to obtain the carbon emission monitoring feature vector.

[0073] Specifically, since the dimensions and value ranges of each component in the carbon emission observation vector SO are different, in order to facilitate unified processing and use in machine learning, SO must first be preprocessed.

[0074] Construct the data vector DV to be processed:

[0075] DV=(pm,pd,pt,pL,bd,sl,sppi,po,ti,yo,sr,sa) (3)

[0076] Since there is a one-to-one correspondence between the data vector DV to be processed and the carbon emission observation vector SO, the carbon emission monitoring database corresponds to the data set M = {DV k} k＝1,…,N M has a matrix form, with its row vectors corresponding to each set of actual carbon emission observation vectors, denoted as Mr, with a total of N row vectors; its column vectors correspond to the actual value set of each observation component, denoted as Mc, with a total of 13 column vectors.

[0077] Since each component in Mr has a different dimension, Mc is dedimensionalized to facilitate unified data processing.

[0078] Mc′=(Mc-mean(Mc)). / std(Mc) (4)

[0079] In the formula, mean() represents the vector mean function, std() represents the vector standard deviation function, and . / represents the vector dot division operation.

[0080] Since each Mc has a different numerical distribution, Mr is centralized to facilitate joint comparison and calculation of data.

[0081] Mr′=(Mr-min(Mr)). / (max(Mr)-min(Mr)) (5)

[0082] In the formula, min() represents the minimum function in the column direction, max() represents the maximum function in the column direction, and . / represents the vector dot division operation.

[0083] After the discrete data joint unit coding, de-dimensionalization and centralization of the carbon emission monitoring database, the carbon emission monitoring feature database MF = {SF k} k＝1,…,N , where the carbon emission monitoring feature vector SF = Mr′, there are N SFs in total.

[0084] Furthermore, a cascaded deep radial network is constructed using the carbon emission monitoring feature library to establish a carbon emission monitoring and early warning model.

[0085] The N carbon emissions (dq) corresponding to the N SFs actually observed are normalized from 0 to 1 to obtain the carbon emission feature vector DF:

[0086] DF=stdnorml(DQ,[0,1]) (6)

[0087] Where DQ is the value set of dq, and stdnorml() represents the vector standard normalization function.

[0088] Define a single-layer radial network as follows:

[0089] NS=newrb(MF,DF,es) (7)

[0090] Where MF is the carbon emission monitoring feature library composed of N carbon emission monitoring feature vectors, DF is the corresponding carbon emission feature vector, newrb() is the single-layer radial network construction function, es represents the network fitting error, and NS is the obtained single-layer radial network.

[0091] Define a cascaded deep radial network as follows:

[0092] NC=casrb(NS,ec,gl) (8)

[0093] Where casrb() represents the cascaded deep radial network construction function, ec represents the deep training step, gl represents the target error, and NC is the obtained cascaded deep radial network.

[0094] Divide N groups of SF and their corresponding DF into n groups of training data sets and Nn groups of test data sets, train the cascaded deep radial network NC until the network error is less than the target gl, then stop the training and obtain the required intelligent laboratory carbon emission monitoring and early warning model. Typically, N = 1000, n = 800, gl = 0.001.

[0095] Furthermore, an early warning of excessive carbon emissions is issued based on the smart laboratory carbon emission monitoring and early warning model.

[0096] In actual carbon emission monitoring, assuming that the actual carbon emission monitoring feature vector is rf, the corresponding carbon emission value rd can be obtained according to the smart laboratory carbon emission monitoring and early warning model:

[0097] rd=ssim(NC′,rf) (9)

[0098] Where NC is the cascaded deep radial network corresponding to the carbon emission monitoring and early warning model, and ssim() represents the deep network prediction function.

[0099] According to the carbon emission value predicted in Formula 9, determine whether it exceeds the standard; if it exceeds the standard, issue a carbon emission exceeding standard warning; if it does not exceed the standard, continue to collect actual carbon emission monitoring feature vectors to predict carbon emission values ​​and determine whether it exceeds the standard.

[0100] Since there is a corresponding relationship between different carbon emission values ​​and their carbon emission operating status, the carbon emissions can be divided into two intervals: those that meet emission standards and those that exceed the standards. When carbon emissions exceed the standards, an early warning can be issued and professionals can be notified to carry out emergency maintenance of laboratory equipment in a timely manner.

[0101] Embodiment 1 provides a smart laboratory carbon emission monitoring and early warning method. This method combines advanced technologies such as Internet of Things intelligent perception, big data analysis, and machine learning. By building a smart laboratory carbon emission monitoring network, collecting carbon emission observation data, and performing data structured feature extraction, deep neural network training, and establishing a smart laboratory carbon emission monitoring and early warning model, it can achieve early warning of abnormal carbon emission states caused by laboratory abnormalities, and can effectively improve the risk management level of carbon emissions in smart laboratories.

[0102] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0103] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A smart laboratory carbon emission monitoring and early warning method, characterized in that: include: Based on the smart laboratory carbon emission monitoring network, collect equipment energy consumption data and determine the smart laboratory carbon emission monitoring database; Determine the data set to be processed based on the smart laboratory carbon emission monitoring database; Preprocess the data set to be processed to obtain a carbon emission monitoring feature library; Based on the carbon emission monitoring feature library, a cascade deep radial network is constructed; Based on the carbon emission monitoring feature library, the cascade deep radial network is trained until the network error is less than the preset target error, and the training stops to obtain the carbon emission monitoring and early warning model. Based on the carbon emission monitoring and early warning model, the carbon emission value is determined according to the actual carbon emission monitoring feature vector; Determine whether the carbon emission value exceeds the warning standard; If the warning standard is exceeded, a carbon emission excess warning will be issued.

2. The method according to claim 1, characterized in that The equipment energy consumption data includes server energy consumption, air-conditioning facility energy consumption, lighting facility energy consumption, elevator energy consumption, fume hood energy consumption, test box energy consumption, switch energy consumption, router energy consumption, water energy consumption, heat energy consumption, greenhouse gas emissions, waste emissions, and vehicle energy consumption.

3. The method according to claim 2, characterized in that Based on the smart laboratory carbon emission monitoring network, equipment energy consumption data is collected to determine the smart laboratory carbon emission monitoring database, including: Based on the intelligent laboratory carbon emission monitoring network, collecting the energy consumption data of the first preset number of groups of equipment; The energy consumption data of each group of equipment is structured and the carbon emission observation vector is determined by the following formula: so=(pm, pd, pt, pl, bd, sl, sp, pi, po, ti, to, sr, sa) Among them, SO is the carbon emission observation vector, pm, pd, pt, pl, bd, sl, sp, pi, po, ti, to, sr, and sa are server energy consumption, air conditioning facility energy consumption, lighting facility energy consumption, elevator energy consumption, fume hood energy consumption, laboratory box energy consumption, switch energy consumption, router energy consumption, water energy consumption, heat energy consumption, greenhouse gas emissions, waste emissions, and vehicle energy consumption, respectively; Based on a first preset number of carbon emission observation vectors, a smart laboratory carbon emission monitoring database is determined.

4. The method according to claim 3, characterized in that: The row vectors of the data set to be processed are a first preset number of carbon emission observation vectors.

5. The method according to claim 4, characterized in that The column vector of the data set to be processed is a value set of carbon emission observation components, which is 13 sets in total.

6. The method according to claim 5, characterized in that Preprocess the data set to be processed to obtain the carbon emission monitoring feature library, including: Each column vector of the data set to be processed is dedimensionalized, and the processed column vector is determined by the following formula: Mc′=(Mc-mean(Mc)). / std(Mc) Among them, Mc′ is the processed column vector, Mc is the column vector of the data set to be processed, mean() is the vector mean function, std() is the vector standard deviation function, and . / is the vector dot division operation; Each row vector of the data set to be processed is centralized, and the processed row vector is determined by the following formula: Mr′=(Mr-min(Mr)). / (max(Mr)-min(Mr)) Among them, Mr′ is the processed row vector, Mr is the row vector of the data set to be processed, min() is the minimum function in the column direction, and max() is the maximum function in the column direction; Based on the first preset number of processed row vectors, the carbon emission monitoring feature library is determined by the following formula: SF=Mr′ MF={SF k } k＝1,…,N Among them, SF is the carbon emission monitoring feature vector, MF is the carbon emission monitoring feature library, and N is the first preset number.

7. The method according to any one of claims 1 to 6, characterized in that: Based on the carbon emission monitoring feature library, a cascade deep radial network is constructed, including: The carbon emissions corresponding to the first preset number of carbon emission monitoring feature vectors are standardized respectively, and the carbon emission feature vector is determined by the following formula: DF=stdnorml(DQ,[0,1]) Among them, DF is the carbon emission feature vector, DQ is the value set of carbon emission, and stdnorml() is the vector standard normalization function; According to the carbon emission feature vector and the carbon emission monitoring feature library, the single-layer radial network is determined by the following formula: NS = newrb(MF, DF, es) Among them, NS is a single-layer radial network, newrb() is a single-layer radial network construction function, and es is the network fitting error; According to the single-layer radial network, a cascaded deep radial network is constructed by the following formula: NC=casrb(NS,ec,gl) Among them, NC is the cascaded deep radial network, casrb() is the cascaded deep radial network construction function, ec is the deep training step, and gl is the target error.

8. The method according to claim 7, characterized in that Based on the carbon emission monitoring feature library, the cascade deep radial network is trained until the network error is less than the preset target error, and the training stops. The carbon emission monitoring and early warning model is obtained, including: Dividing the first preset number group of carbon emission monitoring feature vectors and their corresponding carbon emission feature vectors into a second preset number group of training data sets and a third preset number group of testing data sets; Based on the second preset number of groups of training data sets and the third preset number of groups of test data sets, the cascaded deep radial network is trained until the network error is less than the preset target error, and the training stops to obtain a carbon emission monitoring and early warning model.

9. The method according to claim 8, characterized in that The sum of the second preset number and the third preset number is the first preset number.

10. The method according to claim 9, characterized in that Based on the carbon emission monitoring and early warning model, the carbon emission value is determined by the following formula according to the actual carbon emission monitoring feature vector: rd=ssim(NC,rf) Among them, rd is the carbon emission value, NC′ is the carbon emission monitoring and early warning model, rf is the actual carbon emission monitoring feature vector, and ssim() is the deep network prediction function.

Images