CN112364999B - Training method and device for water chiller adjustment model and electronic equipment - Google Patents
Training method and device for water chiller adjustment model and electronic equipment Download PDFInfo
- Publication number
- CN112364999B CN112364999B CN202011119619.6A CN202011119619A CN112364999B CN 112364999 B CN112364999 B CN 112364999B CN 202011119619 A CN202011119619 A CN 202011119619A CN 112364999 B CN112364999 B CN 112364999B
- Authority
- CN
- China
- Prior art keywords
- sample set
- water chiller
- samples
- sample
- missing
- 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
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 212
- 238000012549 training Methods 0.000 title claims abstract description 47
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000012545 processing Methods 0.000 claims description 28
- 238000012217 deletion Methods 0.000 claims description 25
- 230000037430 deletion Effects 0.000 claims description 25
- 238000004590 computer program Methods 0.000 claims description 17
- 238000005265 energy consumption Methods 0.000 claims description 17
- 238000004140 cleaning Methods 0.000 claims description 8
- 238000010801 machine learning Methods 0.000 claims description 8
- 238000005406 washing Methods 0.000 claims description 6
- 230000003750 conditioning effect Effects 0.000 claims 1
- 230000000875 corresponding effect Effects 0.000 description 12
- 238000010438 heat treatment Methods 0.000 description 11
- 238000004378 air conditioning Methods 0.000 description 10
- 238000010606 normalization Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000009423 ventilation Methods 0.000 description 6
- 238000005057 refrigeration Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000001816 cooling Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000001276 controlling effect Effects 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000002596 correlated effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000006467 substitution reaction 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
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
- F24F11/00—Control or safety arrangements
- F24F11/62—Control or safety arrangements characterised by the type of control or by internal processing, e.g. using fuzzy logic, adaptive control or estimation of values
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Signal Processing (AREA)
- Mechanical Engineering (AREA)
- Combustion & Propulsion (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Chemical & Material Sciences (AREA)
- Computing Systems (AREA)
- General Physics & Mathematics (AREA)
- Fuzzy Systems (AREA)
- Air Conditioning Control Device (AREA)
Abstract
The application provides a training method and a training device for a water chiller adjusting model and electronic equipment, and relates to the technical field of water chiller control, wherein the method comprises the following steps: the method comprises the steps of obtaining a first sample set and a second sample set, determining the similarity between samples in the first sample set and samples in the second sample set, and then training a water chiller regulation model of a target water chiller according to a plurality of samples with the similarity meeting a first preset condition in the second sample set, wherein the samples in the first sample set are generated according to historical operating data of the target water chiller, and the samples in the second sample set are generated according to historical operating data of the same type of water chiller of the target water chiller. The technical scheme provided by the application can select the sample with high similarity as the training sample in the second sample set similar to the first sample set, can effectively increase the training data of the target water chiller, and ensures that the water chiller adjusting model has good accuracy.
Description
Technical Field
The application relates to a water chiller technology, in particular to a training method and a training device for a water chiller adjusting model and electronic equipment, and belongs to the technical field of water chiller control.
Background
Research shows that in modern commercial buildings, the energy consumption of the heating, ventilating and air conditioning system accounts for about 40% of the energy consumption of the whole building, and the energy consumption of the water chilling unit accounts for about 40% of the energy consumption of the heating, ventilating and air conditioning system. Therefore, the energy consumption of the water chilling unit is reduced, and the energy consumption of the whole building can be effectively reduced.
In order to achieve the purposes of energy conservation and environmental protection, modern commercial buildings adopt a scheme of combining a building automatic control system and a heating ventilation air conditioning system to control the energy consumption of a water chiller in the building. And acquiring related data of the water chiller through the heating ventilation air conditioning system, processing the related data through the building automatic control system to obtain the control quantity of the water chiller, and finally controlling the water chiller to work according to the control quantity. As the state of the art has improved, the processing scheme of the relevant data has changed from an experience-based control scheme to a data-based control scheme. For example, a water chiller regulation model of the building is obtained through a machine learning algorithm and training data, and then relevant data is processed by the water chiller regulation model to obtain a control quantity of the water chiller. Data-based control schemes have higher accuracy and wider applicability than experience-based control schemes.
However, the data-based control scheme also has some problems, such as that in the case where the water chiller is just started to be used, there is insufficient training data for the water chiller adjustment model of the water chiller, and the accuracy of training out the water chiller adjustment model is not high.
Disclosure of Invention
In view of this, the present application provides a training method and apparatus for a water chiller adjustment model, and an electronic device, which can ensure that the water chiller adjustment model has good accuracy for a water chiller that is just started to be used.
In order to achieve the above object, in a first aspect, an embodiment of the present application provides a training method for a water chiller adjustment model, including:
acquiring a first sample set and a second sample set, wherein the samples in the first sample set are generated according to historical operating data of a target water chiller, the samples in the second sample set are generated according to historical operating data of the same type of water chiller of the target water chiller, and the data volume of the first sample set is smaller than that of the second sample set;
determining a similarity between samples in the first sample set and samples in the second sample set;
and training a water cooler adjusting model of the target water cooler according to a plurality of samples with the similarity meeting the first preset condition in the second sample set.
Optionally, after acquiring the first sample set and the second sample set, before determining the similarity between the samples in the first sample set and the samples in the second sample set, the method further includes:
and performing data cleaning on the first sample set and the second sample set.
Optionally, performing data washing on the first sample set and the second sample set includes:
and carrying out normalization processing on the first sample set and the second sample set.
Optionally, performing data washing on the first sample set and the second sample set includes:
and performing missing value processing on each missing sample in the first sample set and the second sample set, wherein the missing sample is a sample missing at least one item of data.
Optionally, performing missing value processing on each missing type sample in the first sample set and the second sample set, including:
for each deletion type sample, determining the similarity between the deletion type sample and each integral type sample according to non-deletion items in the deletion type sample, wherein the integral type sample is a sample in which all items of data in a sample set of the deletion type sample exist;
and according to the plurality of intact samples with the similarity meeting the second preset condition, carrying out missing value processing on the missing items of the missing samples.
Optionally, the performing missing value processing on the missing item of the missing type sample according to a plurality of intact type samples whose similarities meet a second preset condition includes:
for each missing item in the missing type sample, determining an average value of corresponding items of the missing item in a plurality of complete type samples;
the mean value is taken as the value of the missing term.
Optionally, the water chiller adjustment model is a supervised machine learning model.
In a second aspect, an embodiment of the present application provides a training device for a water chiller adjustment model, including:
the system comprises an acquisition module, a storage module and a processing module, wherein the acquisition module is used for acquiring a first sample set and a second sample set, samples in the first sample set are generated according to historical operating data of a target water chiller, samples in the second sample set are generated according to historical operating data of the same type of water chiller of the target water chiller, and the data volume of the first sample set is smaller than that of the second sample set;
the training module is used for determining the similarity between the samples in the first sample set and the samples in the second sample set, and training the water chiller regulation model of the target water chiller according to a plurality of samples of which the similarity in the second sample set meets a first preset condition.
Optionally, the apparatus further comprises:
and the data cleaning module is used for performing data cleaning on the first sample set and the second sample set.
Optionally, the data cleansing module is specifically configured to:
and carrying out normalization processing on the first sample set and the second sample set.
Optionally, the data cleansing module is specifically configured to:
and performing missing value processing on each missing sample in the first sample set and the second sample set, wherein the missing sample is a sample missing at least one item of data.
Optionally, the data cleansing module is specifically configured to:
for each deletion type sample, determining the similarity between the deletion type sample and each integral type sample according to non-deletion items in the deletion type sample, wherein the integral type sample is a sample in which all items of data in a sample set of the deletion type sample exist;
and according to the plurality of intact samples with the similarity meeting the second preset condition, carrying out missing value processing on the missing items of the missing samples.
Optionally, the data cleansing module is specifically configured to:
for each missing item in the missing type sample, determining an average value of corresponding items of the missing item in a plurality of complete type samples;
the mean value is taken as the value of the missing term.
Optionally, the water chiller adjustment model is a supervised machine learning model.
In a third aspect, an embodiment of the present application provides an electronic device, including: a memory for storing a computer program and a processor; the processor is configured to perform the method of the first aspect or any of the embodiments of the first aspect when the computer program is invoked.
In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the method according to the first aspect or any embodiment of the first aspect.
According to the training method and device for the water chiller regulation model and the electronic device, a first sample set and a second sample set can be obtained, the similarity between samples in the first sample set and samples in the second sample set is determined, then the water chiller regulation model of the target water chiller is trained according to a plurality of samples, the similarity of the samples in the first sample set and the similarity of the samples in the second sample set meets a first preset condition, wherein the samples in the first sample set are generated according to historical operation data of the target water chiller, the samples in the second sample set are generated according to historical operation data of the same type of water chiller of the target water chiller, and the data volume of the first sample set is smaller than that of the second sample set. The technical scheme that this application provided selects the sample that the similarity is high as the training sample through in the second sample set that is similar with first sample set, can effectively increase the training data of target cold water machine, guarantees that cold water machine regulation model has good rate of accuracy.
Drawings
Fig. 1 is a schematic flowchart of a training method of a water chiller adjustment model according to an embodiment of the present disclosure;
FIG. 2 is a schematic diagram of a system for use in an adjustment model of a water chiller according to an embodiment of the present disclosure;
fig. 3 is a schematic structural diagram of a training device of a water chiller adjustment model according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
In order to facilitate understanding of technical solutions in the embodiments of the present application, some terms referred to in the embodiments of the present application are first explained below:
heating, ventilating and air conditioning system: a system for indoor heating, ventilation and air conditioning can control the temperature and humidity of air and improve the indoor comfort level.
Building automatic control system: a comprehensive system composed of central computer and various control subsystems features use of sensing technology, computer and modern communication technology, and can perform full-automatic comprehensive management on heating, ventilating, elevator, air conditioner, water supply and drainage, electric energy, fire alarm and security.
The load sequence of the water chilling unit is as follows: a mode for controlling the working of the water chilling unit can distribute a refrigeration load to each water chilling unit according to the refrigeration requirement. Wherein, the refrigeration load is a percentage, which represents the percentage of the actual power of the water chiller in the rated maximum power. For example, [0.2, 0.7] indicates a chiller including two chillers, in which the first cooling load is 20% and the second cooling load is 70%.
Coefficient of performance efficiency: the physical quantity describing the performance of the water chiller is specifically defined as the ratio of the refrigerating capacity to the effective input power when the water chiller performs refrigeration under specified conditions.
The technical solution of the present application will be described in detail below with specific examples. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments.
The training method of the water chiller adjustment model provided by the embodiment of the application can be applied to electronic equipment such as a computer, a notebook or a workstation, and the embodiment of the application does not limit the specific type of the electronic equipment.
In order to reduce the energy consumption of a building, modern commercial buildings often adopt various control schemes to control a water chilling unit of a heating, ventilating and air conditioning system, and the control schemes are mainly divided into experience-based control schemes and data-based control schemes.
The experience-based control scheme is mainly to set a threshold value of a physical quantity related to the cooling load of the water chiller according to application experience, and to change the operating state of the water chiller when the physical quantity exceeds the threshold value. Typical experience-based control schemes include: sequence control based on chilled water return temperature, sequence control based on bypass water flow, sequence control based on direct energy consumption, and sequence control based on total refrigeration load. Taking the sequence control based on the return temperature of the chilled water as an example, when the automatic building control system monitors that the return temperature of the chilled water exceeds a set threshold of the water cooler, the water cooler is controlled to start working. However, the control accuracy of the experience-based control scheme is not high and depends very much on the working experience of the designer.
The data-based control scheme includes a part load ratio scheme and a machine learning scheme. The partial load ratio scheme can be used for modeling the water chilling unit, the input of the partial load ratio scheme is the load percentage of the water chilling unit, the output of the partial load ratio scheme is the energy consumption of the water chilling unit or the performance efficiency coefficient of the water chilling unit, and the purpose of saving energy of the water chilling unit is achieved by selecting the load percentage of the water chilling unit which enables the total energy consumption of the water chilling unit to be the lowest. Compared with an experience-based control scheme, the method can realize the control of the water chiller with higher control precision. However, the model of the partial load ratio scheme is established based on a water chilling unit with fixed performance, and as the service time of the water chilling unit increases, the inherent performance of the water chilling unit may change, thereby causing the accuracy of the model to be reduced. The water chiller adjusting model of the machine learning scheme is obtained by training according to historical operating data, so that the water chiller adjusting model is not influenced by inherent performance change, and the accuracy of the water chiller adjusting model can be maintained only by frequently updating the water chiller adjusting model through the historical operating data. And each water cooler corresponds to the water cooler adjusting model, and the water cooler adjusting model is more flexible and convenient when being newly added.
However, when the water chiller is just started to be used, the accuracy of the water chiller adjustment model trained in this case is not high because sufficient training data is not available for the water chiller adjustment model of the water chiller. Therefore, the training method of the water chiller adjusting model is used for ensuring that the trained water chiller adjusting model has good accuracy.
Fig. 1 is a schematic flowchart of a training method for a water chiller adjustment model according to an embodiment of the present application, and as shown in fig. 1, the method includes the following steps:
and S110, acquiring a first sample set and a second sample set.
Assuming that the hvac system is first used in a new building, sufficient training data is not generated at this time due to the short usage time of the target chiller. In order to ensure that the water chiller regulation model has good accuracy, a first sample set can be generated according to historical operating data of a target water chiller, and a second sample set can be generated according to historical operating data of the same type of water chiller of the target water chiller. It should be noted that the target water chiller is typically a new water chiller and therefore has a shorter usage time, while the same type of water chiller is typically an old water chiller and therefore has a longer operation time, and therefore the data size of the second sample set is greater than the data size of the first sample set. The historical operating data may include: the system comprises a chilled water supply temperature, a chilled water return temperature, a chilled water flow rate, a water cooler load, water cooler energy consumption, a water cooler working life, an outdoor temperature, a performance efficiency coefficient and the like.
For example, in the first scenario, building a is a new building that is used for the first time, in which a plurality of new water coolers (i.e., target water coolers) are provided, and building B, building C, and building D are old buildings that have been put into service for a long time, in each of which a plurality of old water coolers of the same type are also provided. For each target water chiller, the target water chiller may be first controlled to operate for a short period of time, such as one or three days, and then a first sample set may be generated based on the collected historical operating data. Since the old water coolers in buildings B, C and D have been operating for a longer period of time, for example, half a year or a year, the second sample set may be generated directly from all or part of the historical operating data collected for the old water coolers.
In the second scenario, the building E is an old building which has been put into use for a long time, a plurality of old water coolers are arranged in the building E, and a new water cooler (i.e. a target water cooler) needs to be added to the building E. For the target water chiller, the target water chiller may be first controlled to operate for a short period of time, such as one or three days, and then a first sample set may be generated based on the collected historical operating data. A second sample set is then generated based on historical operating data of other old water coolers in the building E.
And S120, performing data cleaning on the first sample set and the second sample set.
In order to improve the quality of the sample and facilitate subsequent data processing, a data cleaning may be performed on the first sample set and the second sample set. The data cleansing specifically may include: normalization processing and/or missing value processing.
To unify the data dimension between different samples, the electronic device may perform normalization processing on the first sample set and the second sample set, for example, performing max-min normalization processing on the first sample set and the second sample set. The specific normalization method can be selected according to actual requirements, and the application is not limited.
In practical application of the water chiller, due to the lack of effective management on historical operating data, a phenomenon that part of the historical operating data is lost exists. For example, a complete sample should include 8 sub-items of chilled water supply temperature, chilled water return temperature, chilled water flow rate, water chiller load, water chiller energy consumption, water chiller working age, outdoor temperature and performance efficiency coefficient, but due to management problems, data of chilled water return temperature and chilled water flow rate are not collected in some samples, and a sample lacking at least one item of data is referred to as a missing sample in the present application.
The electronic device may perform missing value processing on each missing type sample in the first sample set and the second sample set. Through missing value processing, missing type samples can be perfected, sample quality is improved, and accuracy of a water chiller adjusting model is improved.
Specifically, for each missing sample, the electronic device may determine, according to non-missing items in the missing sample, a similarity between the missing sample and each intact sample, where the intact sample is a sample in which all items of data in a sample set in which the missing sample is located exist. The similarity between the samples can be obtained by calculating the euclidean distance between the samples and also can be obtained by calculating the cosine distance between the samples, and the specific selection of which way to obtain the similarity can be determined according to actual requirements, which is not limited in the present application.
Further, the electronic device may determine each non-missing item in the missing sample, then determine a sub-item corresponding to the non-missing item in each complete sample, and then determine the similarity between the complete sample and the missing sample according to each non-missing item in the missing sample and the corresponding sub-item in the complete sample.
For example, there are one missing type sample and a plurality of complete type samples in the second sample set, the complete type samples include a, b, c, d and e, 5 sub-entries are included, and the c entry in the missing type sample has no specific data. The electronic device may determine non-missing items a, b, d, and e in the missing type sample, then determine corresponding sub-items a, b, d, and e in the full type sample, and then determine the similarity between the full type sample and the missing type sample according to the sub-items a, b, d, and e in the missing type sample and the sub-items a, b, d, and e in the full type sample.
After obtaining the similarity between the missing type sample and each of the intact type samples, the electronic device may perform missing value processing on the missing item of the missing type sample according to a plurality of intact type samples whose similarities meet a second preset condition. Specifically, a second similarity threshold may be set, and if the similarity of the complete sample exceeds the second similarity threshold, the complete sample may be considered to satisfy a second preset condition; or setting a preset number N, sequencing from large to small according to the similarity of all the complete samples, and determining the first N complete samples as samples meeting a second preset condition. The second similarity threshold and the preset number N can be determined by a cross-validation method.
Specifically, when the missing value processing is performed on the missing item of the missing type sample, for each missing item in the missing type sample, an average value of the data of the corresponding item of the missing item in the complete type sample meeting the second preset condition may be determined, and the average value may be used as the value of the missing item.
For example, if c-term in the missing-type sample is missing term, and there are 50 complete-type samples meeting the second preset condition, the electronic device may calculate an average value of c-term data of the 50 complete-type samples, and use the average value as the value of the missing term.
And S130, determining the similarity between the samples in the first sample set and the samples in the second sample set.
The electronic device may determine a similarity between the samples in the first sample set and the samples in the second sample set according to a euclidean distance or a cosine distance between the samples in the first sample set and the samples in the second sample set.
And S140, training a water chiller regulation model of the target water chiller according to a plurality of samples with the similarity meeting the first preset condition in the second sample set.
After obtaining the similarity between the samples in the first sample set and the samples in the second sample set, the electronic device may determine, for each sample in the first sample set, a plurality of samples meeting a first preset condition in the second sample set, and then train a water chiller adjustment model of the target water chiller according to the determined samples; or summing the euclidean distances corresponding to the samples in the first sample set for each sample in the second sample set to serve as the similarity between the sample and the first sample set, then determining a plurality of samples meeting a first preset condition in the second sample set, and then training the water chiller adjustment model of the target water chiller according to the determined samples. In addition, the electronic equipment can train the water chiller regulation model together with the first sample set according to a plurality of samples of which the similarity in the second sample set meets the first preset condition, so that the number of training samples is further increased.
The number of samples for training the adjusting model of the water chiller can be increased by selecting a proper number of samples from the second sample set as training samples, and the trained adjusting model of the water chiller can have higher accuracy under the condition of enough training samples.
For example, the first sample set has 5 samples, the second sample set has 90 samples, and for each sample in the first sample set, the electronic device may determine 10 samples in the second sample set that meet the first preset condition, and then may train the water chiller adjustment model of the target water chiller according to the determined 50 samples.
For another example, the first sample set has 5 samples, the second sample set has 90 samples, for each sample in the second sample set, the electronic device may sum the similarities corresponding to the 5 samples in the first sample set to serve as the similarity between the sample and the first sample set, then, 50 samples meeting the first preset condition are determined in the second sample set, and then, the water chiller adjustment model of the target water chiller is trained according to the determined 50 samples.
Specifically, in the embodiment of the present application, the water chiller regulation model may be a supervised machine learning model.
Further, the electronic device may set a first similarity threshold, and if the similarity of the samples in the second sample set exceeds the first similarity threshold, the complete sample may be considered to satisfy a first preset condition; or setting a preset number N, sorting the samples from large to small according to the similarity of the samples in the second sample set, and determining the samples in the first N second sample sets as the samples meeting the first preset condition. The first similarity threshold and the preset number N can be determined by a cross-validation method.
In addition, in a specific application, if in a first scenario, namely, the building a is a new building which is used for the first time, a plurality of new water coolers (namely target water coolers) are arranged in the new building, and the buildings B, C and D are old buildings which are already put into use for a long time, a plurality of old water coolers of the same type are also arranged in each building. When the electronic equipment judges the first preset condition, the electronic equipment can compare the similarity by taking a building as a unit, namely the total similarity of the samples of which building is high, and then the corresponding samples are used for training. Because all the water coolers have the same working environment in the same building, certain correlation exists among historical operating data, and the accuracy of the water cooler adjusting model can be improved by selecting the correlated data as a training sample.
Fig. 2 is a schematic diagram of a system used by a water chiller adjustment model according to an embodiment of the present application, and as shown in fig. 2, a heating, ventilation, air conditioning system and a building automatic control system are installed in a certain building, and after a water chiller adjustment model of a target water chiller in the building is trained, a trained water chiller adjustment model is obtained. During formal work, the heating ventilation air conditioning system can acquire various operating data of the water chiller in real time through the sensors and send the operating data to the building automatic control system through the controller, the building automatic control system can input the operating data into the water chiller adjusting model, and then the water chiller adjusting model outputs performance efficiency coefficients under different loads. And selecting the power which can meet the cold quantity requirement and has the lowest energy consumption by the building automatic control system according to the cold quantity requirement of the building, and finally determining the load sequence of the water chilling unit according to the selected power. The controller in the heating ventilation air-conditioning system can control the water chilling unit to work under the conditions of ensuring the cold quantity demand and ensuring the minimum energy consumption according to the determined water chilling unit load sequence.
In this embodiment, the electronic device may obtain a first sample set and a second sample set, determine similarity between samples in the first sample set and samples in the second sample set, and then train a water chiller adjustment model of a target water chiller according to a plurality of samples in the second sample set, where the similarity meets a first preset condition, where the samples in the first sample set are generated according to historical operating data of the target water chiller, the samples in the second sample set are generated according to historical operating data of a similar water chiller to the target water chiller, and a data volume of the first sample set is smaller than a data volume of the second sample set. The method selects the sample with high similarity as the training sample in the second sample set similar to the first sample set, can effectively increase the training data of the target water cooler, and ensures that the water cooler adjusting model has good accuracy.
Based on the same inventive concept, as an implementation of the foregoing method, an embodiment of the present application provides a training apparatus for a water chiller adjustment model, where the apparatus embodiment corresponds to the foregoing method embodiment, and for convenience of reading, details in the foregoing method embodiment are not repeated one by one in the apparatus embodiment, but it should be clear that the apparatus in the embodiment can correspondingly implement all the contents in the foregoing method embodiment.
Fig. 3 is a schematic structural diagram of a training device of a water chiller adjustment model according to an embodiment of the present application, and as shown in fig. 3, the device according to the embodiment includes:
an obtaining module 110, configured to obtain a first sample set and a second sample set, where samples in the first sample set are generated according to historical operating data of a target water chiller, samples in the second sample set are generated according to historical operating data of a similar water chiller of the target water chiller, and a data amount of the first sample set is smaller than a data amount of the second sample set;
the training module 120 is configured to determine similarity between samples in the first sample set and samples in the second sample set, and train a water chiller adjustment model of the target water chiller according to a plurality of samples in the second sample set, where the similarity meets a first preset condition.
Optionally, the apparatus further comprises:
and a data washing module 130 for performing data washing on the first sample set and the second sample set.
Optionally, the data cleansing module 130 is specifically configured to:
and carrying out normalization processing on the first sample set and the second sample set.
Optionally, the data cleansing module 130 is specifically configured to:
and performing missing value processing on each missing sample in the first sample set and the second sample set, wherein the missing sample is a sample missing at least one item of data.
Optionally, the data cleansing module 130 is specifically configured to:
for each deletion type sample, determining the similarity between the deletion type sample and each integral type sample according to non-deletion items in the deletion type sample, wherein the integral type sample is a sample in which all items of data in a sample set of the deletion type sample exist;
and according to the plurality of intact samples with the similarity meeting the second preset condition, carrying out missing value processing on the missing items of the missing samples.
Optionally, the data cleansing module 130 is specifically configured to:
for each missing item in the missing type sample, determining an average value of corresponding items of the missing item in a plurality of complete type samples;
the mean value is taken as the value of the missing term.
Optionally, the water chiller adjustment model is a supervised machine learning model.
The training device of the water chiller adjustment model provided by the embodiment can execute the method embodiment, the implementation principle and the technical effect are similar, and the details are not repeated here.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions. Each functional unit and module in the embodiments may be integrated in one processing unit, or each unit may exist alone physically, or two or more units are integrated in one unit, and the integrated unit may be implemented in a form of hardware, or in a form of software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working processes of the units and modules in the system may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
Based on the same inventive concept, the embodiment of the application also provides the electronic equipment. Fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application, and as shown in fig. 4, the electronic device according to the embodiment includes: a memory 21 and a processor 20, the memory 21 being for storing a computer program; the processor 20 is arranged to perform the method according to the above-described method embodiment when the computer program 22 is invoked.
The electronic device provided by this embodiment may perform the above method embodiments, and the implementation principle and the technical effect are similar, which are not described herein again.
Embodiments of the present application further provide a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the computer program implements the method described in the above method embodiments.
The integrated unit may be stored in a computer-readable storage medium if it is implemented in the form of a software functional unit and sold or used as a separate product. Based on such understanding, all or part of the processes in the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium and can implement the steps of the embodiments of the methods described above when the computer program is executed by a processor. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable storage medium may include at least: any entity or device capable of carrying computer program code to a photographing apparatus/terminal apparatus, a recording medium, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signal, telecommunication signal, and software distribution medium. Such as a usb-disk, a removable hard disk, a magnetic or optical disk, etc. In certain jurisdictions, computer-readable media may not be an electrical carrier signal or a telecommunications signal in accordance with legislative and patent practice.
In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus/device and method may be implemented in other ways. For example, the above-described apparatus/device embodiments are merely illustrative, and for example, the division of the modules or units is only one logical division, and there may be other divisions when actually implemented, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
It will be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It should also be understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.
As used in this specification and the appended claims, the term "if" may be interpreted contextually as "when", "upon" or "in response to" determining "or" in response to detecting ". Similarly, the phrase "if it is determined" or "if a [ described condition or event ] is detected" may be interpreted contextually to mean "upon determining" or "in response to determining" or "upon detecting [ described condition or event ]" or "in response to detecting [ described condition or event ]".
Furthermore, in the description of the present application and the appended claims, the terms "first," "second," "third," and the like are used for distinguishing between descriptions and not necessarily for describing or implying relative importance.
Reference throughout this specification to "one embodiment" or "some embodiments," or the like, means that a particular feature, structure, or characteristic described in connection with the embodiment is included in one or more embodiments of the present application. Thus, appearances of the phrases "in one embodiment," "in some embodiments," "in other embodiments," or the like, in various places throughout this specification are not necessarily all referring to the same embodiment, but rather "one or more but not all embodiments" unless specifically stated otherwise. The terms "comprising," "including," "having," and variations thereof mean "including, but not limited to," unless expressly specified otherwise.
Finally, it should be noted that: the above embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application 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 or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present application.
Claims (7)
1. A training method of a water chiller regulation model is characterized by comprising the following steps:
obtaining a first sample set and a second sample set, wherein samples in the first sample set are generated according to historical operating data of a target water chiller, samples in the second sample set are generated according to historical operating data of the same type of water chiller of the target water chiller, and the data volume of the first sample set is smaller than that of the second sample set; the target water chiller is a new water chiller, and the similar water chiller is an old water chiller; the historical operating data may include: the system comprises a chilled water supply temperature, a chilled water return temperature, a chilled water flow rate, a water chiller load, water chiller energy consumption, a water chiller service life, an outdoor temperature and a performance efficiency coefficient;
performing data washing on the first sample set and the second sample set, including: for each deletion type sample, determining the similarity between the deletion type sample and each intact type sample according to non-deletion items in the deletion type sample, wherein the intact type sample is a sample in which all items of data in a sample set of the deletion type sample exist; according to a plurality of intact samples with similarity meeting a second preset condition, carrying out missing value processing on missing items of the missing samples; the missing type sample is a sample missing at least one item of data;
determining a similarity between samples in the first set of samples and samples in the second set of samples;
and training a water chiller adjusting model of the target water chiller according to a plurality of samples with the similarity meeting a first preset condition in the second sample set.
2. The method of claim 1, wherein the data washing the first set of samples and the second set of samples comprises:
and normalizing the first sample set and the second sample set.
3. The method according to claim 1, wherein the performing missing value processing on the missing item of the missing type sample according to a plurality of intact type samples with similarity meeting a second preset condition comprises:
for each missing term in the missing type sample, determining an average of the corresponding terms of the missing term in the plurality of intact type samples;
and taking the average value as the value of the missing item.
4. The method of any of claims 1-3, wherein the water chiller conditioning model is a supervised machine learning model.
5. The utility model provides a trainer of model is adjusted to cold water machine which characterized in that includes:
the system comprises an acquisition module, a storage module and a processing module, wherein the acquisition module is used for acquiring a first sample set and a second sample set, samples in the first sample set are generated according to historical operating data of a target water chiller, samples in the second sample set are generated according to historical operating data of the same type of water chiller of the target water chiller, and the data volume of the first sample set is smaller than that of the second sample set; the target water chiller is a new water chiller, and the similar water chiller is an old water chiller; the historical operating data may include: the system comprises a chilled water supply temperature, a chilled water return temperature, a chilled water flow rate, a water chiller load, water chiller energy consumption, a water chiller service life, an outdoor temperature and a performance efficiency coefficient;
the data cleaning module is used for cleaning data of the first sample set and the second sample set; the data cleaning module is specifically configured to: for each deletion type sample, determining the similarity between the deletion type sample and each integral type sample according to non-deletion items in the deletion type sample, wherein the integral type sample is a sample in which all items of data in a sample set of the deletion type sample exist; according to a plurality of intact samples with similarity meeting a second preset condition, carrying out missing value processing on missing items of the missing samples, wherein the missing samples are samples missing at least one item of data;
the training module is used for determining the similarity between the samples in the first sample set and the samples in the second sample set, and training a water chiller regulation model of the target water chiller according to a plurality of samples of which the similarity in the second sample set meets a first preset condition.
6. An electronic device, comprising: a memory for storing a computer program and a processor; the processor is adapted to perform the method of any of claims 1-4 when the computer program is invoked.
7. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the method according to any one of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011119619.6A CN112364999B (en) | 2020-10-19 | 2020-10-19 | Training method and device for water chiller adjustment model and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011119619.6A CN112364999B (en) | 2020-10-19 | 2020-10-19 | Training method and device for water chiller adjustment model and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112364999A CN112364999A (en) | 2021-02-12 |
| CN112364999B true CN112364999B (en) | 2021-11-19 |
Family
ID=74508230
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011119619.6A Active CN112364999B (en) | 2020-10-19 | 2020-10-19 | Training method and device for water chiller adjustment model and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112364999B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113623924B (en) * | 2021-06-24 | 2023-01-10 | 杭州大和热磁电子有限公司 | Semiconductor refrigeration cold water machine with standby control system |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104966097A (en) * | 2015-06-12 | 2015-10-07 | 成都数联铭品科技有限公司 | Complex character recognition method based on deep learning |
| CN108805048A (en) * | 2018-05-25 | 2018-11-13 | 腾讯科技（深圳）有限公司 | A kind of method of adjustment of human face recognition model, device and storage medium |
| CN109840588A (en) * | 2019-01-04 | 2019-06-04 | 平安科技（深圳）有限公司 | Neural network model training method, device, computer equipment and storage medium |
| CN110008338A (en) * | 2019-03-04 | 2019-07-12 | 华南理工大学 | A kind of electric business evaluation sentiment analysis method of fusion GAN and transfer learning |
| WO2019214365A1 (en) * | 2018-05-10 | 2019-11-14 | 腾讯科技（深圳）有限公司 | Translation model training method, sentence translation method and apparatus, and storage medium |
| CN111047048A (en) * | 2019-11-22 | 2020-04-21 | 支付宝(杭州)信息技术有限公司 | Energized model training and merchant energizing method and device, and electronic equipment |
| CN111126700A (en) * | 2019-12-25 | 2020-05-08 | 远景智能国际私人投资有限公司 | Energy consumption prediction method, device, equipment and storage medium |
| CN111191791A (en) * | 2019-12-02 | 2020-05-22 | 腾讯云计算（北京）有限责任公司 | Application method, training method, device, equipment and medium of machine learning model |
| CN111256294A (en) * | 2020-01-17 | 2020-06-09 | 深圳市得益节能科技股份有限公司 | Model prediction-based optimization control method for combined operation of water chilling unit |
| CN111768000A (en) * | 2020-06-23 | 2020-10-13 | 中南大学 | Industrial process data modeling method for online adaptive fine-tuning deep learning |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106874581B (en) * | 2016-12-30 | 2021-03-30 | 浙江大学 | Building air conditioner energy consumption prediction method based on BP neural network model |
| CN108399414B (en) * | 2017-02-08 | 2021-06-01 | 南京航空航天大学 | Sample selection method and device applied to cross-modal data retrieval field |
| CN108898120A (en) * | 2018-07-04 | 2018-11-27 | 四川泰立智汇科技有限公司 | Water cooler method for diagnosing faults based on hybrid neural networks |
| CN109540562A (en) * | 2018-12-12 | 2019-03-29 | 上海理工大学 | A kind of water cooler method for diagnosing faults |
| CN110162014A (en) * | 2019-05-29 | 2019-08-23 | 上海理工大学 | A kind of breakdown of refrigeration system diagnostic method of integrated multi-intelligence algorithm |
-
2020
- 2020-10-19 CN CN202011119619.6A patent/CN112364999B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104966097A (en) * | 2015-06-12 | 2015-10-07 | 成都数联铭品科技有限公司 | Complex character recognition method based on deep learning |
| WO2019214365A1 (en) * | 2018-05-10 | 2019-11-14 | 腾讯科技（深圳）有限公司 | Translation model training method, sentence translation method and apparatus, and storage medium |
| CN108805048A (en) * | 2018-05-25 | 2018-11-13 | 腾讯科技（深圳）有限公司 | A kind of method of adjustment of human face recognition model, device and storage medium |
| CN109840588A (en) * | 2019-01-04 | 2019-06-04 | 平安科技（深圳）有限公司 | Neural network model training method, device, computer equipment and storage medium |
| CN110008338A (en) * | 2019-03-04 | 2019-07-12 | 华南理工大学 | A kind of electric business evaluation sentiment analysis method of fusion GAN and transfer learning |
| CN111047048A (en) * | 2019-11-22 | 2020-04-21 | 支付宝(杭州)信息技术有限公司 | Energized model training and merchant energizing method and device, and electronic equipment |
| CN111191791A (en) * | 2019-12-02 | 2020-05-22 | 腾讯云计算（北京）有限责任公司 | Application method, training method, device, equipment and medium of machine learning model |
| CN111126700A (en) * | 2019-12-25 | 2020-05-08 | 远景智能国际私人投资有限公司 | Energy consumption prediction method, device, equipment and storage medium |
| CN111256294A (en) * | 2020-01-17 | 2020-06-09 | 深圳市得益节能科技股份有限公司 | Model prediction-based optimization control method for combined operation of water chilling unit |
| CN111768000A (en) * | 2020-06-23 | 2020-10-13 | 中南大学 | Industrial process data modeling method for online adaptive fine-tuning deep learning |
Non-Patent Citations (2)
| Title |
|---|
| "基于聚类的连续型数据缺失值充填方法";李国和等;《计算机工程》;20190930;第45卷(第9期);第33页第1节 * |
| "零样本学习研究进展";张鲁宁等;《自动化学报》;20200131;第46卷(第1期);第2页左栏第4段和第10-11页第4节 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112364999A (en) | 2021-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108302719B (en) | Control method, device and system of multi-split air conditioning system and storage medium | |
| CN111442489B (en) | Control method and system of air conditioner, air conditioner and readable storage medium | |
| CN101231018A (en) | System and method for controlling demand of multi-air-conditioner | |
| WO2020199682A1 (en) | Air conditioner control method, air conditioner control apparatus, storage medium, memory and air conditioner | |
| CN107036238B (en) | Intelligent energy-saving control method for dynamically predicting external air and load | |
| CN108613326B (en) | Air conditioning system, intelligent regulation control method and device thereof and computer equipment | |
| CN112364999B (en) | Training method and device for water chiller adjustment model and electronic equipment | |
| CN110454941B (en) | Intelligent temperature control method and device and air conditioning equipment | |
| CN115017006A (en) | Auxiliary energy-saving control method and system for data center | |
| CN115388520A (en) | Air conditioner control method, air conditioner control device, air conditioner and storage medium | |
| CN115264768A (en) | Building air conditioner energy-saving control method and system based on Internet of things | |
| CN114608128A (en) | Method and device for controlling temperature of air conditioner chip, air conditioner and storage medium | |
| CN113137701B (en) | Method and device for controlling air conditioner and air conditioner | |
| CN110762739B (en) | Data center air conditioner control method, device, equipment and storage medium | |
| CN110285536A (en) | A kind of control method and device of temperature control equipment | |
| CN113446713A (en) | Digital twin-based intelligent data optimal control and energy saving method and system | |
| CN115327930B (en) | Visual energy-saving control method and system | |
| CN113803862B (en) | Valve control method of air conditioning system, electronic device, and readable storage medium | |
| CN115013938A (en) | Air conditioner control method, device, equipment and storage medium | |
| CN111578453A (en) | Cooling water optimization control system and method based on big data analysis | |
| CN114383293B (en) | Method and device for controlling air conditioner, air conditioner and storage medium | |
| CN214406424U (en) | Cooling water optimization control system based on big data analysis | |
| CN117537450B (en) | Method, device, equipment and medium for controlling operation feedback of central air-conditioning refrigeration system | |
| CN113465109B (en) | Low-energy-consumption standby control method and device for multi-split air conditioner and medium | |
| CN114002971B (en) | Building energy-saving control method and system |
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 |