CN115456115B - Cold station operation and maintenance multilayer energy-saving potential diagnosis method based on actual measurement subentry measurement data - Google Patents
Cold station operation and maintenance multilayer energy-saving potential diagnosis method based on actual measurement subentry measurement data Download PDFInfo
- Publication number
- CN115456115B CN115456115B CN202211416842.6A CN202211416842A CN115456115B CN 115456115 B CN115456115 B CN 115456115B CN 202211416842 A CN202211416842 A CN 202211416842A CN 115456115 B CN115456115 B CN 115456115B
- Authority
- CN
- China
- Prior art keywords
- data
- energy consumption
- cold
- energy
- maintenance
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000003745 diagnosis Methods 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 65
- 238000012423 maintenance Methods 0.000 title claims abstract description 32
- 238000005259 measurement Methods 0.000 title abstract description 18
- 238000005265 energy consumption Methods 0.000 claims abstract description 93
- 238000004140 cleaning Methods 0.000 claims abstract description 20
- 230000015556 catabolic process Effects 0.000 claims abstract description 15
- 238000006731 degradation reaction Methods 0.000 claims abstract description 15
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 68
- 238000004458 analytical method Methods 0.000 claims description 25
- 239000000498 cooling water Substances 0.000 claims description 14
- 238000002790 cross-validation Methods 0.000 claims description 12
- 238000009826 distribution Methods 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 11
- 238000001816 cooling Methods 0.000 claims description 10
- 238000004088 simulation Methods 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 6
- 230000003068 static effect Effects 0.000 claims description 6
- 238000010845 search algorithm Methods 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 238000007781 pre-processing Methods 0.000 claims description 4
- 238000005057 refrigeration Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 238000012795 verification Methods 0.000 claims description 4
- 230000002068 genetic effect Effects 0.000 claims description 3
- 239000002245 particle Substances 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 2
- 238000007710 freezing Methods 0.000 description 7
- 230000008014 freezing Effects 0.000 description 7
- 238000004378 air conditioning Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000010438 heat treatment Methods 0.000 description 2
- 238000002156 mixing Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- UCTWMZQNUQWSLP-VIFPVBQESA-N (R)-adrenaline Chemical compound CNC[C@H](O)C1=CC=C(O)C(O)=C1 UCTWMZQNUQWSLP-VIFPVBQESA-N 0.000 description 1
- 241000254032 Acrididae Species 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 238000012550 audit Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 238000005293 physical law Methods 0.000 description 1
- 238000010926 purge Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000005406 washing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Air Conditioning Control Device (AREA)
Abstract
The invention provides a cold station operation and maintenance multilayer energy-saving potential diagnosis method based on actual measurement subentry measurement data, which diagnoses the operation and maintenance energy efficiency problem of a cold station in a shallow diagnosis and deep diagnosis combined mode, wherein the shallow diagnosis utilizes the total energy consumption data of a cold machine, and analyzes the problem of the performance degradation of the cold machine by a modeling and variable deduction mode. Deep diagnostics analyze the energy efficiency of the equipment and system after the cleaning of the item measurement data. According to the invention, by the method for analyzing the operation, maintenance and energy-saving potential of the cold station based on the actually measured subentry metering data, shallow diagnosis and deep diagnosis are carried out according to different data conditions, so that the difference of different building energy consumption data can be adapted, and the common energy efficiency problem of the cold station can be effectively diagnosed and analyzed.
Description
Technical Field
The invention belongs to the technical field of building energy consumption analysis, and particularly relates to a cold station operation and maintenance multilayer energy-saving potential diagnosis method based on actual measurement subentry measurement data.
Background
Energy-saving potential analysis aiming at the existing buildings is an important ring for completing a double-carbon target, rapid development of a national public building energy consumption monitoring platform and deep application of a data-driven algorithm in the whole industry are beneficial to application of a data mining technology to the energy-saving potential analysis of the buildings, so that the period of traditional building energy audit is shortened, and technical personnel are assisted to rapidly position energy-saving potential points.
The blank of real data is filled with simulation data, and the simulation data is used for granularity conversion, so that the data fusion method can only fill up 'macroscopic' data such as total energy consumption of heating, ventilating and air conditioning buildings at present. In the case that no subentry electric meter exists in the existing building, a non-intrusive load splitting method based on event driving is adopted, and the core of the method is that the electrical signals of starting and stopping of different devices and conversion of working states need to be known.
However, if the existing buildings have different informatization degrees, it is relatively difficult to perform an omnidirectional and non-differential analysis on the energy-saving potential of the buildings with the same data structure, because the data of different buildings are different. Filling and mixing with simulation data is a 'temporary solution and non-permanent solution' method, which has a certain tolerance to the problem of the existing building energy consumption metering platform or the problem of poor BAS data, and if the filling method is used for completing the data which can be actually monitored, the existing emerging Internet of things platform related to building operation and maintenance does not play the due value. Most of the non-intrusive load splitting methods are focused on the research on residential buildings mainly because the operating characteristics of main electric equipment in the residential buildings are relatively regular, for example, a refrigerator works intermittently, while relatively few commercial buildings adopt the non-intrusive load splitting methods for splitting, mainly because the electric equipment is complex and various, and the uncertainty of the use of the equipment is increased due to the uncertainty of personnel.
Disclosure of Invention
In view of the above, the present invention is directed to solve the above problems of the existing analysis method when analyzing the operation, maintenance and energy saving potential of a building cold station in consideration of the difference of energy consumption data of different buildings.
In order to solve the technical problems, the invention provides the following technical scheme:
a cold station operation and maintenance multilayer energy-saving potential diagnosis method based on measured subentry metering data comprises the following steps:
determining the energy consumption data condition of a target analysis building, and performing shallow diagnosis and/or deep diagnosis on the operation and maintenance energy efficiency of the cold station according to the energy consumption data condition;
for total energy consumption data of the cold machine group, a shallow diagnosis method is adopted, and a key variable is conjectured for the energy consumption of the cold machine group through an optimal black box model so as to diagnose the energy efficiency problem of the cold station, wherein the optimal black box model is a pre-trained model used for obtaining the building load distribution of a target analysis building, and the key variable is a key parameter influencing the energy consumption of the cold machine group;
and for the subentry metering data, a deep diagnosis method is adopted, the subentry metering data are subjected to data cleaning based on a cross validation mode, and the cleaned data are used for diagnosing the energy efficiency problem of the cold station.
Further, a shallow diagnosis method is adopted to diagnose the energy efficiency problem of the cold station, and the method specifically comprises the following steps:
acquiring total energy consumption data of a cold machine group of a target analysis building;
preprocessing each item of energy consumption data in the total energy consumption data of the cold machine group;
based on the preprocessed energy consumption data, the key energy consumption variable of the cold machine group is presumed through the optimal black box model, and the load distribution of the target analysis building is obtained;
and determining a cold machine model selection scheme optimally matched with the target analysis building load distribution by using a global search algorithm so as to complete the diagnosis of the cold station energy efficiency problem.
Further, the key variables specifically include:
the method comprises the steps of judging the refrigerating performance of a refrigerator, judging the type of an air system and the type of a water system, wherein the type of the air system and the type of the water system are used as static variable information to be processed, the refrigerating performance of the refrigerator is divided into a refrigerating coefficient of the refrigerator and a degradation coefficient of the refrigerator, and the divided two coefficient values are used as key variables needing to be presumed in shallow diagnosis.
Further, the process of establishing the optimal black box model specifically includes:
acquiring historical key variable information influencing the total energy consumption of the refrigerator;
sampling other related variable information except static variable information in the historical key variable information;
building an energy consumption model by combining the static variable information and other sampled key variable information;
training a black box model by using a building-cooler energy consumption database as a training data source;
and carrying out hyper-parameter optimization on the trained black box model to obtain an optimal black box model.
Furthermore, the method for predicting the energy consumption key variable of the cold group through the optimal black box model specifically comprises the following steps:
and speculating the key energy consumption variable of the cold machine group by adopting a particle swarm algorithm or a genetic algorithm, wherein when the deviation between the energy consumption values calculated according to the actual value of the key variable and the key variable value obtained by sampling is measured, the deviation is measured by adopting a simulation calibration guide rule index or a dynamic time warping index.
Further, a deep diagnosis method is adopted to diagnose the energy efficiency problem of the cold station, and the method specifically comprises the following steps:
acquiring the sub-item metering data of different buildings;
performing data cross validation on each subentry measurement data, and performing data cleaning in the process of validating single equipment and system level data;
and diagnosing the cold station energy efficiency problem after the verification is passed.
Further, the data cross validation specifically includes:
the method comprises the steps of taking the itemized energy consumption data of various equipment groups of different buildings as a starting point of cross validation, firstly validating data related to single equipment respectively, and then validating data of a system level.
Further, verifying data associated with a single device specifically includes:
for the cooling water pump, sequentially verifying the energy consumption of a single cooling water pump, the frequency of the single cooling water pump and the flow of main pipe cooling water;
for the cooling tower, sequentially verifying the energy consumption of a single cooling tower fan and the frequency of the single cooling tower fan;
and for the cold machine, sequentially verifying the energy consumption of the single cold machine and the starting and stopping states of the single cold machine.
Furthermore, when the starting and stopping states of the single cold machine are verified, the starting and stopping state data of the single cold machine are obtained by mapping the energy consumption data of the single cold machine.
Further, the system-level data specifically includes:
the water supply temperature of the main pipe chilled water, the return water temperature and the cold load of the main pipe chilled water, the water supply of the single-unit cold machine chilled water, the return water of the single-unit cold machine chilled water, the water supply temperature of the single-unit cold machine cooling water and the return water temperature of the single-unit cold machine cooling water.
In conclusion, the invention provides a cold station operation and maintenance multilayer energy-saving potential diagnosis method based on actually measured subentry metering data, which diagnoses the operation and maintenance energy efficiency problem of a cold station in a shallow diagnosis and deep diagnosis combined mode, wherein the shallow diagnosis utilizes the total energy consumption data of a cold machine, and analyzes the problem of the performance degradation of the cold machine by a modeling and variable inference mode. Deep diagnostics analyze the energy efficiency of the equipment and system after the cleaning of the item measurement data. According to the invention, by the method for analyzing the operation, maintenance and energy-saving potential of the cold station based on the actually measured subentry metering data, shallow diagnosis and deep diagnosis are carried out according to different data conditions, so that the difference of different building energy consumption data can be adapted, and the common energy efficiency problem of the cold station can be effectively diagnosed and analyzed.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a technical route diagram of a cold station operation and maintenance multi-layer energy-saving potential diagnosis method based on measured subentry measurement data according to an embodiment of the present invention;
fig. 2 is a flow chart of data cleansing for cold station energy consumption data according to an embodiment of the present invention.
Detailed Description
In order to make the objects, features and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
The energy-saving potential analysis is carried out on the existing buildings, and due to the fact that the existing buildings are different in informatization degree, the energy-saving potential analysis of the buildings in all aspects and without difference is relatively difficult if the same data structure is used for carrying out the energy-saving potential analysis on the buildings in all aspects, and the difference of the data of the buildings is large. Filling and mixing with simulation data is a 'temporary solution and non-permanent solution' method, which has a certain tolerance to the problem of the existing building energy consumption metering platform or the problem of poor BAS data, and if the filling method is used for completing the data which can be actually monitored, the existing emerging Internet of things platform related to building operation and maintenance does not play the due value. Most of the non-intrusive load splitting methods are focused on the research on residential buildings mainly because the operating characteristics of main electric equipment in the residential buildings are relatively regular, for example, a refrigerator works intermittently, while relatively few commercial buildings adopt the non-intrusive load splitting methods for splitting, mainly because the electric equipment is complex and various, and the uncertainty of the use of the equipment is increased due to the uncertainty of personnel.
Based on the method, the invention provides a cold station operation and maintenance multi-layer energy-saving potential diagnosis method based on actual measurement subentry measurement data.
The following provides a detailed description of an embodiment of the cold station operation and maintenance multi-layer energy-saving potential diagnosis method based on the measured subentry metering data.
The embodiment provides a cold station operation and maintenance multilayer energy-saving potential diagnosis method based on measured subentry metering data, which comprises the following steps:
the method comprises the following steps: and determining the energy consumption data condition of the target analysis building, and performing shallow diagnosis and/or deep diagnosis on the operation and maintenance energy efficiency of the cold station according to the energy consumption data condition.
Step two: and for the total energy consumption data of the cold machine group, a shallow diagnosis method is adopted, and a key variable is conjectured for the energy consumption of the cold machine group through an optimal black box model so as to diagnose the energy efficiency problem of the cold station, wherein the optimal black box model is a pre-trained model used for obtaining the building load distribution of the target analysis building, and the key variable is a key parameter influencing the energy consumption of the cold machine group.
Step three: and for the subentry metering data, a deep diagnosis method is adopted, the subentry metering data are subjected to data cleaning based on a cross validation mode, and the cleaned data are used for diagnosing the energy efficiency problem of the cold station.
The embodiment provides a cold station operation and maintenance multilayer energy-saving potential diagnosis method based on measured subentry metering data, which diagnoses the operation and maintenance energy efficiency problem of a cold station in a shallow diagnosis and deep diagnosis combined mode, wherein the shallow diagnosis utilizes the total energy consumption data of a cold machine, and analyzes the problem of the performance degradation of the cold machine by a modeling and variable deduction mode. Deep diagnostics analyze the energy efficiency of the equipment and system after the cleaning of the item measurement data. According to the invention, by the method for analyzing the operation, maintenance and energy-saving potential of the cold station based on the actually measured subentry metering data, shallow diagnosis and deep diagnosis are carried out according to different data conditions, so that the difference of different building energy consumption data can be adapted, and the common energy efficiency problem of the cold station can be effectively diagnosed and analyzed.
The shallow diagnosis method and the deep diagnosis method for cold station energy efficiency diagnosis in the present embodiment are described in detail below with reference to fig. 1 and 2.
1. Superficial diagnosis
The shallow diagnosis is to analyze the problem of the performance degradation of the refrigerator by modeling and variable inference by using the total energy consumption data of the refrigerator.
As shown in fig. 1, the shallow diagnosis process includes obtaining total energy consumption data of the cold cluster of the target analysis building; preprocessing each sub-item energy consumption data in the total energy consumption data of the cold machine group; based on the preprocessed energy consumption data, the key energy consumption variable of the cold machine group is presumed through the optimal black box model, and the load distribution of the target analysis building is obtained; and determining a cold machine model selection scheme which is optimally matched with the target analysis building load distribution by using a global search algorithm so as to finish the diagnosis of the cold station energy efficiency problem. The superficial diagnosis will be described in detail below.
(1) Selection of key variables
The key parameters influencing the energy consumption of the refrigerator are the refrigeration performance of the refrigerator, the type of an air system and the type of a water system. The last two parameters can not be changed, the refrigerating performance of the refrigerator is divided into a refrigerator COP and a refrigerator degradation coefficient, and the two parameters are key parameters required to be inferred. Part of the reason for chiller degradation is deposits on the condenser surfaces, which are often a result of poor water quality. Scaling can cause certain problems in heat exchange in a water chilling unit, indirectly cause parameter indexes such as exhaust temperature or exhaust pressure to be incapable of meeting set requirements, and further cause performance degradation of different degrees under different PLRs (partial load ratios, which are often used for describing the performance of air conditioning equipment under non-design working conditions). The scaling problem can not occur in the absorption water chilling unit which does not need water as a heat exchange medium, the degradation coefficient of the chiller can reduce the nominal refrigerating capacity of the water chilling unit, and when the degradation problem is serious, the water chilling unit can not process the design building load. The effect of the cold machine degradation coefficient is shown in the following equation:
wherein,
the nominal refrigerating capacity of the refrigerator under the degradation condition is referred to;
(2) Optimal black box model formation to assist in inferring key variables
In order to fill the database under the condition of ensuring the data richness, the key variables are sampled by an LHS (hyper-Laval Ding Chouyang) method, and then the idf files are processed in batches by a Grasshopper and an Eppy toolkit to form the building-cooler energy consumption database. Because the black box model has a lower professional knowledge threshold compared with the white box model, and many tools are sourced and maintained and updated regularly, the black box model is a popular trend for analyzing the energy-saving potential of the building. The invention utilizes a regression tree model of a decision tree in machine learning, adopts two methods of XGboost (eXtreme Gradient Boosting) and LightGBM, and trains a black box model (an optimized black box model which is investigated in advance or is subjected to parameter adjustment in a standard library) according to a building-cold machine energy consumption database as training data. Building load distribution of a target analysis building can be obtained according to the black box model, and then whether the matching degree of the cold machine selection scheme and the building load distribution can be further improved or not is analyzed by adopting a global search algorithm.
(3) Preprocessing of itemized energy consumption data
When the energy consumption of the chiller group is used for conjecturing the key variable, attention needs to be paid to calculating the subentry metering data in the strict sense of the energy consumption of the chiller group, and compared with the total energy consumption, the subentry metering data is easier to find the data quality problem, such as mean shift; bulk missing values; the signal to noise ratio is relatively small. The glitch in the building energy consumption data curve is caused by the quality problem of the sensor, which needs to be observed and investigated in the field, and the quality of data transmitted by the sensor is measured by using the concept of Signal-to-Noise Ratio (SNR), which is used to compare the Signal strength with the environmental Noise strength, SNR is defined as the Ratio of the Signal strength to the Noise strength, and the unit of the Signal strength is usually expressed in decibels (dB), and the calculation formula is as follows. Signal strength is greater than noise strength when the strength is greater than 0dB and the ratio is greater than 1:1. Lower SNR indicates a more similar noise level and signal level, and thus a less useful signal; the higher the SNR value, the smaller the noise intensity, and the general case is that the SNR is poor in the range of [5,20 ].
Where P is power and A is amplitude.
(4) Shallow diagnostic energy efficiency diagnostic analysis based on key variable speculation
And (3) inferring key variables according to the black box model and the energy consumption data of the cold machine group after noise reduction, wherein the global search algorithm and the index for measuring the inferred deviation are subject to the total energy consumption of the cold machine group after noise reduction.
The key variable speculation can be considered as a set-constrained optimization problem, with the formula expressed in mathematical language as follows.
Wherein the function
The objective function or the cost function is a real-valued function, and the potential meaning of the optimization problem is that we need to find a suitable functionxMake a functionf(x)To a minimum.xIs onenA dimension vector expressed as->
Independent of each other and are commonly referred to as decision variables. The set Ω isnDimension real number space->
A subset of (a) is called a constraint set or a feasible set.
After the nature of the problem is determined, the problem that we visualize is abstracted to mathematical language,f(x)and if the deviation is smaller, the value of the sampled key variable is close to the real value of the key variable. The value ranges of the two key variables in the omega set need to be determined, because the key variables presumed in the method have practical significance, and the constraint of the value ranges can reduce the dilemma of falling into local optimum. Finally, we need to determine which optimization method can be used to deduce the value of the key variable in a short time with an error within an acceptable range.
The invention adopts a more general Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) for speculation. When measuring the deviation between the energy consumption values calculated according to the actual values of the key variables and the sampled key variable values, the evaluation may be performed by using indexes provided in a Simulation Calibration guide (Simulation Calibration guides), such as Mean Bias Error (MBE) and variance Coefficient Root Mean Square Error (CV-RMSE) in ASHRAE guide 14, which are indexes commonly used by professional researchers for measuring the difference between the analog values and the actual values, and may also consider another index because the building energy consumption data is a Time series in nature, and another index commonly used for judging the similarity of different Time series is Dynamic Time Warping (DTW), which is particularly suitable for Time series of different tempos, and the DTW measures the similarity between two Time series by using a normalized Path Distance (Warp Path Distance). The two measures are adopted by the invention of the specimen, and are combined with the two global random algorithms in pairs.
The above is a detailed description of the superficial diagnosis, and the following is a detailed description of the deep diagnosis.
2. Deep diagnosis
Deep diagnostics are an energy efficiency problem for analyzing equipment and systems after cleaning the item measurement data.
As shown in fig. 1, the process of deep diagnosis includes acquiring various metering data of different buildings; performing data cross validation on each subentry measurement data, and performing data cleaning in the process of validating single equipment and system level data; and diagnosing the cold station energy efficiency problem after the verification is passed. The process of deep diagnosis is described in detail below.
(1) Data quality problem of subentry measurement data
In engineering, not all the subentry data are equally important, for example, the power consumption data of the equipment group may be relatively more important, because it is an important basis for payment. It is thus readily apparent that the data quality may vary depending on the importance, and that relatively minor data quality may be relatively poor when there are fewer missing values in the important data. The data cleaning idea of the invention is as follows: and selecting the data with the multiplied quality as calibration, and correcting other data by using a physical rule. The quality problem of the subentry data of each building can be various and eight doors, and whether the commonness can be found out in the characteristics becomes the key for building a general subentry data cleaning flow. The cleaned data can reflect the physical rule of the heating, ventilating and air conditioning system and can be directly used for deep diagnosis.
(2) Data cleaning framework for cross-validation-based itemized metering data
FIG. 2 is a flow diagram of data cleansing in a data cleansing framework. The framework adopts physical laws to perform cross validation on data. And taking the itemized energy consumption data of each equipment group as a starting point of cross validation, then validating data related to single equipment respectively, and finally validating data of a system level. The data cleaning process provides reliable data basis for deep diagnosis. The dashed arrows in fig. 2 indicate that it is likely that a partial flow will go wrong due to a complete missing of data.
The data cleaning process is explained in detail by using total energy consumption of a chilled water pump, energy consumption of a single chilled water pump, frequency of the single chilled water pump, water flow of a main pipe, water temperature of the main pipe and cold load:
1) Total energy consumption of the chilled water pump: and (4) removing zero values and null values, and judging whether the histogram of the total energy consumption of the chilled water pump is reasonable or not according to expert experience.
2) Energy consumption of a single refrigerating water pump: and (4) performing data cleaning according to kirchhoff's law and the total energy consumption of the processed chilled water pumps, namely, the energy consumption of a single chilled water pump is equal to the total energy consumption of the chilled water pumps, and the formula is as follows.
Wherein
For total energy consumption of the freeze pump>
The energy consumption of a single refrigerating water pump is reduced,nthe total number of the chilled water pumps.
3) Frequency of a single refrigerating water pump: and performing data cleaning according to similar criteria to correct the problem of non-correspondence between the data and the label, wherein the formula is as follows.
Wherein
For the energy consumption of a single freezing water pump, is adjusted>
Rated energy consumption for a single freezing water pump>
For a single chilled water pump frequency, is selected>
The rated frequency of a single refrigerating water pump is adopted,nthe total number of the chilled water pumps. The points to be explained here are: in order to ensure the safe operation of the water pump, the lowest frequency of the variable-frequency water pump is set at 30Hz, so that the situation that the low-speed condition of the water pump possibly does not meet the similar criterion can be avoided.
4) Flow rate of chilled water of the dry pipe: and performing data cleaning according to the similarity criterion and the frequency of the single processed freezing water pump, wherein the formula is as follows, and then comparing the calculated main pipe freezing water flow with the monitored main pipe freezing water flow by using expert experience and judging whether the water flow is reasonable or not.
Wherein
For a calculated dry tube freeze water flow rate, <' >>
For a single chilled water pump frequency, is selected>
For a single chilled water pump rated frequency->
The rated flow rate of a single refrigerating water pump,nthe total number of the chilled water pumps.
Dry pipe water temperature, cold load: and analyzing whether the water temperature and the cold load of the main pipe are reasonable or not according to the law of conservation of energy and the corrected data, and rejecting the data with overlarge deviation according to the following formula.
Wherein
For a modified dry tube chilled water flow rate, <' > or>
For a modified dry tube cooling water flow, <' > or>
Freezing water return temperature for the dry pipe>
Freezing water return temperature for the dry pipe>
Supplying water temperature for the cooling water of the dry pipe>
For cooling the return water temperature of the drying pipe>
Is water with constant pressure and specific heat capacity>
Is the density of water.
The water pump and cooling tower fan related data purge are similar, while the chiller related variables are slightly different: in fig. 2, the cold start-stop state is at the same level as the water pump frequency and the fan frequency, and the cleaning of the cold start-stop state is additionally described here: the start-stop state of the cold machine theoretically corresponds to the energy consumption of the corresponding cold machine, so that the energy consumption data of the single cold machine is directly mapped to the start-stop state of the corresponding cold machine.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (7)
1. The cold station operation and maintenance multilayer energy-saving potential diagnosis method based on the measured subentry metering data is characterized by comprising the following steps of:
determining the energy consumption data condition of a target analysis building, and performing shallow diagnosis and/or deep diagnosis on the operation and maintenance energy efficiency of the cold station according to the energy consumption data condition;
for the total energy consumption data of the cold machine group, the shallow diagnosis method is adopted, and a key variable is conjectured for the energy consumption of the cold machine group through an optimal black box model so as to diagnose the energy efficiency problem of the cold station, wherein the optimal black box model is a pre-trained model used for obtaining the building load distribution of the target analysis building, and the key variable is a key parameter influencing the energy consumption of the cold machine group;
for the subentry metering data, the deep diagnosis method is adopted, the subentry metering data are subjected to data cleaning based on a cross validation mode, and the cleaned data are used for diagnosing the energy efficiency problem of the cold station;
the method for diagnosing the energy efficiency problem of the cold station specifically comprises the following steps:
acquiring each sub-item metering data of different buildings;
performing data cross validation on the subentry metering data, and performing data cleaning in the process of validating single equipment and system level data;
diagnosing the cold station energy efficiency problem after the verification is passed;
the data cross validation specifically comprises:
respectively verifying data related to single equipment by taking the itemized energy consumption data of each equipment group of different buildings as a starting point of cross verification, and then verifying system level data;
verifying data relating to a single device, including in particular:
for the cooling water pump, sequentially verifying the energy consumption of a single cooling water pump, the frequency of the single cooling water pump and the flow of main pipe cooling water;
for the cooling tower, sequentially verifying the energy consumption of a single cooling tower fan and the frequency of the single cooling tower fan;
and for the cold machine, sequentially verifying the energy consumption of the single cold machine and the starting and stopping states of the single cold machine.
2. The cold station operation and maintenance multilayer energy-saving potential diagnosis method based on measured subentry metering data according to claim 1, wherein the shallow diagnosis method is adopted to diagnose the cold station energy efficiency problem, and specifically comprises the following steps:
acquiring total energy consumption data of the cold machine group of the target analysis building;
preprocessing each sub-item energy consumption data in the total energy consumption data of the cold machine group;
based on the preprocessed energy consumption data, the key energy consumption variable of the cold machine group is presumed through the optimal black box model, and the load distribution of the target analysis building is obtained;
and determining a cold machine model selection scheme which is optimally matched with the target analysis building load distribution by using a global search algorithm so as to finish the diagnosis of the cold station energy efficiency problem.
3. The cold station operation and maintenance multilayer energy-saving potential diagnosis method based on measured subentry metering data according to claim 2, wherein the key variables specifically include:
the system comprises a refrigerator refrigeration performance, an air system type and a water system type, wherein the air system type and the water system type are used as static variable information to be processed, the refrigerator refrigeration performance is divided into a refrigerator refrigeration coefficient and a refrigerator degradation coefficient, and the divided two coefficient values are used as the key variables needing to be presumed in the shallow diagnosis.
4. The cold station operation and maintenance multilayer energy-saving potential diagnosis method based on the measured subentry metering data as claimed in claim 2, wherein the process of establishing the optimal black box model specifically comprises:
acquiring historical key variable information influencing the total energy consumption of the refrigerator;
sampling other related variable information except static variable information in the historical key variable information;
building an energy consumption model of the building by combining the static variable information and other sampled key variable information;
training a black box model by using a building-cooling machine energy consumption database as a training data source;
and carrying out hyper-parameter optimization on the trained black box model to obtain the optimal black box model.
5. The multi-layer energy-saving potential diagnosis method for operation and maintenance of a cold station based on measured subentry metering data according to claim 2, wherein the step of inferring a key cold-cluster energy consumption variable through the optimal black box model specifically comprises the following steps:
and speculating the key energy consumption variable of the cold machine group by adopting a particle swarm algorithm or a genetic algorithm, wherein when deviation between energy consumption values calculated according to the actual value of the key variable and the key variable value obtained by sampling is measured, the deviation is measured by adopting a simulation calibration guide rule index or a dynamic time warping index.
6. The cold station operation and maintenance multilayer energy-saving potential diagnosis method based on measured subentry metering data according to claim 1, characterized in that when the start-stop state of the single cold machine is verified, the start-stop state data of the single cold machine is obtained by mapping energy consumption data of the single cold machine.
7. The cold station operation and maintenance multi-layer energy-saving potential diagnosis method based on the measured subentry metering data according to claim 1, wherein the system level data specifically comprises:
the system comprises a main pipe chilled water supply temperature, a main pipe chilled water return temperature and a cold load, as well as single-unit cold machine chilled water supply, single-unit cold machine chilled water return, single-unit cold machine cooling water supply temperature and single-unit cold machine cooling water return temperature.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211416842.6A CN115456115B (en) | 2022-11-14 | 2022-11-14 | Cold station operation and maintenance multilayer energy-saving potential diagnosis method based on actual measurement subentry measurement data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211416842.6A CN115456115B (en) | 2022-11-14 | 2022-11-14 | Cold station operation and maintenance multilayer energy-saving potential diagnosis method based on actual measurement subentry measurement data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115456115A CN115456115A (en) | 2022-12-09 |
| CN115456115B true CN115456115B (en) | 2023-04-07 |
Family
ID=84295624
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211416842.6A Active CN115456115B (en) | 2022-11-14 | 2022-11-14 | Cold station operation and maintenance multilayer energy-saving potential diagnosis method based on actual measurement subentry measurement data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115456115B (en) |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105627506A (en) * | 2015-12-25 | 2016-06-01 | 深圳达实智能股份有限公司 | Building cooling load prediction method and device of air conditioning refrigerator |
| CN105956752A (en) * | 2016-04-25 | 2016-09-21 | 天津城建大学 | Energy efficiency evaluation and diagnosis analysis method used for green public building operation phase |
| CN109583008B (en) * | 2018-10-18 | 2023-09-15 | 天津大学 | General modeling method for water chiller energy efficiency model based on stepwise regression |
| CN110805997A (en) * | 2019-11-14 | 2020-02-18 | 中金新源（天津）科技有限公司 | Energy-saving control method for central air-conditioning system |
| CN111486552B (en) * | 2020-04-24 | 2021-07-20 | 辽宁工程技术大学 | Method for identifying water supply temperature strategy of chilled water of air conditioner based on subentry metering data |
| CN113757944B (en) * | 2021-09-24 | 2022-11-11 | 广东电网有限责任公司 | Air conditioning system fault diagnosis method and system based on air conditioning system model |
| CN114880309A (en) * | 2022-04-15 | 2022-08-09 | 同济大学 | Data cleaning method for building cold station subentry metering data |
-
2022
- 2022-11-14 CN CN202211416842.6A patent/CN115456115B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115456115A (en) | 2022-12-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Guo et al. | Remaining useful life prediction for rolling bearings using EMD-RISI-LSTM | |
| US11436395B2 (en) | Method for prediction of key performance parameter of an aero-engine transition state acceleration process based on space reconstruction | |
| CN109240244B (en) | Data-driven equipment running state health degree analysis method and system | |
| CN108931387B (en) | Fault diagnosis method based on multi-sensor signal analysis and capable of providing accurate diagnosis decision | |
| Zhou et al. | Vibration-based fault diagnosis of the natural gas compressor using adaptive stochastic resonance realized by Generative Adversarial Networks | |
| CN109858140B (en) | Fault diagnosis method for water chilling unit based on information entropy discrete Bayesian network | |
| Xu et al. | An anomaly detection and dynamic energy performance evaluation method for HVAC systems based on data mining | |
| CN112199890B (en) | System-level fault diagnosis method for integrated nuclear power device | |
| CN114360703A (en) | Medical equipment fault early warning method | |
| CN109492790A (en) | Wind turbines health control method based on neural network and data mining | |
| Tian et al. | An unsupervised data mining-based framework for evaluation and optimization of operation strategy of HVAC system | |
| CN112700162B (en) | Method and device for evaluating running state of rail transit air conditioner | |
| CN109711664A (en) | A kind of power transmission and transforming equipment health evaluation system based on big data | |
| Xu et al. | Anomaly detection with gru based bi-autoencoder for industrial multimode process | |
| Liu et al. | Real-time comprehensive health status assessment of hydropower units based on multi-source heterogeneous uncertainty information | |
| CN115456115B (en) | Cold station operation and maintenance multilayer energy-saving potential diagnosis method based on actual measurement subentry measurement data | |
| CN116258467B (en) | Electric power construction management and control system | |
| Feng et al. | Distributed chaotic bat algorithm for sensor fault diagnosis in AHUs based on a decentralized structure | |
| CN117469105A (en) | Blade icing identification method and system based on condition variation self-encoder | |
| CN112462708B (en) | Remote diagnosis and optimized scheduling method and system for pump station | |
| CN115828744A (en) | White light LED fault on-line diagnosis and service life prediction method | |
| CN115310746A (en) | State evaluation method and system for main transmission system of wind generating set | |
| CN114114995A (en) | A debug system for data center air conditioning system | |
| TWI790917B (en) | Method for estimating power consumption of water chiller and cooling system | |
| CN118411156B (en) | Full life cycle health management method and system for large equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |