CN109612037A - Air-conditioner control system - Google Patents
Air-conditioner control system Download PDFInfo
- Publication number
- CN109612037A CN109612037A CN201811158832.0A CN201811158832A CN109612037A CN 109612037 A CN109612037 A CN 109612037A CN 201811158832 A CN201811158832 A CN 201811158832A CN 109612037 A CN109612037 A CN 109612037A
- Authority
- CN
- China
- Prior art keywords
- mentioned
- learning model
- environment
- air
- condition
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 claims abstract description 64
- 238000010801 machine learning Methods 0.000 claims abstract description 34
- 238000001514 detection method Methods 0.000 claims abstract description 12
- 238000004378 air conditioning Methods 0.000 claims description 66
- 238000012545 processing Methods 0.000 claims description 47
- 238000003860 storage Methods 0.000 claims description 46
- 238000004364 calculation method Methods 0.000 claims description 42
- 238000012360 testing method Methods 0.000 claims description 19
- 238000007726 management method Methods 0.000 description 21
- 230000007613 environmental effect Effects 0.000 description 15
- 238000010586 diagram Methods 0.000 description 14
- 238000004821 distillation Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000004387 environmental modeling Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000003754 machining Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Classifications
-
- 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
- F24F11/63—Electronic processing
- F24F11/64—Electronic processing using pre-stored data
-
- 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
-
- 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
- F24F11/30—Control or safety arrangements for purposes related to the operation of the system, e.g. for safety or monitoring
-
- 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
- F24F11/63—Electronic processing
- F24F11/65—Electronic processing for selecting an operating mode
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D23/00—Control of temperature
- G05D23/19—Control of temperature characterised by the use of electric means
- G05D23/1917—Control of temperature characterised by the use of electric means using digital means
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Fuzzy Systems (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Air Conditioning Control Device (AREA)
- General Factory Administration (AREA)
Abstract
The present invention provides a kind of air-conditioner control system for controlling air conditioner, condition in designated environment, detection indicates the quantity of state of environment, and according to the quantity of state detected come the control method of the air conditioner in the above-mentioned environment of reasoning, the control method gone out by inference controls air conditioner.Air-conditioner control system is also generated by using the machine learning of the quantity of state detected or renewal learning model, and the combination of the condition of these multiple learning models and environment is stored in association.
Description
Technical field
The present invention relates to a kind of air-conditioner control systems, more particularly to a kind of setting situation and fortune according to factory's inner machine
Turn situation to switch learning model and control the air-conditioner control system of air-conditioning.
Background technique
It is provided with multiple machines (such as lathe) in the factory, and the temperature in Precision Machining in factory influences processing
Precision, therefore it is provided with the air conditioner for controlling environmental condition in factory.Production in factory indicated by production plan,
Multiple the operation of a machine situations and processing conditions are controlled, so that producing the production indicated by production plan within specified delivery date
Product, and the operational situation of air conditioner is controlled to realize temperature condition (temperature and uniformity) in the required factory of processing.
Meet the technology of temperature condition in the required factory of processing as control air-conditioning, such as in Japanese Unexamined Patent Publication
Following technology is disclosed in 06-307703 bulletin, Japanese Unexamined Patent Publication 2015-206519 bulletin: controlling multiple air-conditionings in center
Temperature in factory is remained scheduled temperature by machine.
However, Temperature Distribution is according to the machine (lathe, robot etc.) as pyrotoxin being set in factory in factory
Setting situation and the difference of each the operation of a machine situation and it is different.Therefore, it is processed in required factory to realize
Temperature condition needs to consider setting situation, the operational situation of each machine to control air conditioner, is carried out in this way according to various situations
Airconditioning control appropriate is more difficult.
Also consider the case where importing machine learning device to carry out airconditioning control, but in order to generate cope with it is above-mentioned
The general-purpose machinery learner (general learning model) of various situations, many states letter for needing to detect under various conditions
Breath, additionally needs many media variables comprising data related with situation, therefore is generated known to overlearning etc. sometimes
Problem.
Summary of the invention
Therefore, the object of the present invention is to provide setting situation, fortune that one kind can broadly account for machine
Turn the air-conditioner control system of the airconditioning control appropriate of situation.
In air-conditioner control system of the invention, solved the above problems by the mechanism of setting switching learning model, on
State learning model for determine according to the setting situation of factory's inner machine and operational situation come using air-conditioning control action.This
The air-conditioner control system of invention has multiple learning models, is selected according to the setting situation of factory's inner machine and operational situation etc.
Learning model carries out the machine learning based on the quantity of state detected out of factory to the learning model selected, and by this
The learning model that sample generates correspondingly is used separately to carry out sky according to the setting situation of factory's inner machine and operational situation etc.
The control of tune machine.
The air-conditioner control system control of one embodiment of the present invention is provided with the air conditioner in the environment of at least one machine,
Air-conditioner control system control setting has: condition specifying part specifies the condition in above-mentioned environment；Quantity of state test section,
Detection indicates the quantity of state of the state of above-mentioned environment；Reasoning and calculation portion, according to above-mentioned quantity of state come in the above-mentioned environment of reasoning
The control method of above-mentioned air conditioner；Airconditioning control portion is controlled according to the control method inferred by above-mentioned reasoning and calculation portion
Above-mentioned air conditioner；Learning model generating unit, is generated or renewal learning mould by using the machine learning of above-mentioned quantity of state
Type；And learning model storage unit, by least one learning model generated by above-mentioned learning model generating unit with by above-mentioned
The combination of the specified condition of condition specifying part is stored in association.Moreover, above-mentioned reasoning and calculation portion is according to by above-mentioned condition
Condition in the specified above-mentioned environment of specifying part, selectively makes from the learning model for being stored in above-mentioned learning model storage unit
With at least one learning model, the control method of the above-mentioned air conditioner in the environment managed by above-mentioned air-conditioner control system 1 is calculated.
Above-mentioned air-conditioner control system can also have: characteristic quantity generating unit, detect according to by above-mentioned quantity of state test section
Quantity of state out come generate make above-mentioned environment have feature characteristic quantity, moreover, above-mentioned reasoning and calculation portion is according to features described above amount
Carry out the control method of the above-mentioned air conditioner in the above-mentioned environment of reasoning, in addition, above-mentioned learning model generating unit is by using above-mentioned
The machine learning of characteristic quantity generates or renewal learning model.
Above-mentioned learning model generating unit is by changing the existing learning model implementation stored by above-mentioned learning model storage unit
Become to generate new learning model.
The learning model generated by above-mentioned learning model generating unit is encrypted and is stored by above-mentioned learning model storage unit,
The learning model encrypted when reading learning model by above-mentioned reasoning and calculation portion is decoded.
The air-conditioner control system control of other way of the invention is provided with the air conditioner in the environment of at least one machine,
Air-conditioner control system control setting has: condition specifying part specifies the condition in above-mentioned environment；Quantity of state test section,
Detection indicates the quantity of state of above-mentioned environment；Reasoning and calculation portion, according to above-mentioned quantity of state come the above-mentioned sky in the above-mentioned environment of reasoning
The control method of tune machine；Airconditioning control portion controls above-mentioned sky according to the control method inferred by above-mentioned reasoning and calculation portion
Tune machine；And learning model storage unit, it stores and the combination of the condition in above-mentioned environment at least one associated in advance
Practise model.Moreover, above-mentioned reasoning and calculation portion is according to the condition in the above-mentioned environment specified by above-mentioned condition specifying part, from being stored in
At least one learning model is selectively used in the learning model of above-mentioned learning model storage unit, is calculated upper in above-mentioned environment
State the control method of air conditioner.
Above-mentioned air-conditioner control system can also have: characteristic quantity generating unit, generated according to above-mentioned quantity of state make it is above-mentioned
Environment have feature characteristic quantity, moreover, above-mentioned reasoning and calculation portion according to features described above amount come reasoning by above-mentioned airconditioning control system
The control method of above-mentioned air conditioner in the environment of 1 management of system.
The air conditioning control device of one embodiment of the present invention has above-mentioned condition specifying part and quantity of state test section.
The air conditioning control method of one embodiment of the present invention executes following steps: specified control is provided at least one machine
Environment in air conditioner condition the step of；Detection indicates the step of quantity of state of above-mentioned environment；According to above-mentioned quantity of state come
The step of control method of above-mentioned air conditioner in the above-mentioned environment of reasoning；Above-mentioned air conditioner is controlled according to above-mentioned control method
Step；And it is generated by using the machine learning of above-mentioned quantity of state or the step of renewal learning model.Moreover, above-mentioned push away
Selection basis in combination at least one above-mentioned learning model associated in advance of condition of the step from above-mentioned environment is managed to exist
Learning model used in the condition in above-mentioned environment specified in the step of specified above-mentioned condition, uses the study mould selected
Type calculates the control method of the above-mentioned air conditioner in above-mentioned environment.
Above-mentioned air conditioning control method also executes according to above-mentioned quantity of state the characteristic quantity for generating and above-mentioned environment being made to have feature
The step of,
Above-mentioned inference step can according to features described above amount come the control method of the above-mentioned air conditioner in the above-mentioned environment of reasoning,
Moreover, generating or generating or update by using the machine learning of features described above amount the step of updating above-mentioned learning model
Practise model.
The air conditioning control method of other way of the invention executes following steps: specified control is provided at least one machine
Environment in air conditioner condition the step of；Detection indicates the step of quantity of state of above-mentioned environment；It is pushed away according to above-mentioned quantity of state
The step of managing the control method of the above-mentioned air conditioner in above-mentioned environment；And above-mentioned air conditioner is controlled according to above-mentioned control method
The step of.Moreover, the combination of condition of the above-mentioned inference step from above-mentioned environment at least one study mould associated in advance
The learning model according to used in the condition in the above-mentioned environment specified in the step of specifying above-mentioned condition is selected in type, is used
The learning model selected calculates the control method of the above-mentioned air conditioner in above-mentioned environment.
Above-mentioned air conditioning control method can also execute according to above-mentioned quantity of state the spy for generating and above-mentioned environment being made to have feature
The step of sign amount, moreover, above-mentioned inference step can also be according to features described above amount come the above-mentioned air conditioner in the above-mentioned environment of reasoning
Control method.
The learning model collection of one embodiment of the present invention be make multiple learning models each with control be provided at least
The combination of the condition of air conditioner in the environment of one machine is associated to be formed, and each of above-mentioned multiple learning models is according to table
The quantity of state of the above-mentioned environment carried out under conditions of showing in above-mentioned environment is come the learning model that generates or update, from above-mentioned multiple
A learning model is selected in learning model according to condition set by environment, the learning model selected is above-mentioned for reasoning
The processing of the control method of above-mentioned air conditioner in environment.
In accordance with the invention it is possible to according to the quantity of state detected in each situation to the setting situation according to factory's inner machine
The learning model selected with operational situation carries out machine learning, therefore can prevent overlearning and to be able to carry out efficiency good
Good machine learning, in addition, using the learning model selected according to the setting situation of factory's inner machine and operational situation etc. into
The control of row air conditioner, therefore improve the control precision of air conditioner.
Detailed description of the invention
Fig. 1 is the summary functional block diagram of the air-conditioner control system of first embodiment.
Fig. 2 illustrates the model of the environment managed by air-conditioner control system 1.
Fig. 3 is the summary functional block diagram of the air-conditioner control system of second embodiment.
Fig. 4 is the summary functional block diagram of the air-conditioner control system of third embodiment.
Fig. 5 is the summary functional block diagram of the air-conditioner control system of the 4th embodiment.
Fig. 6 is the summary functional block diagram of the air-conditioner control system of the 5th embodiment.
Fig. 7 is the summary functional block diagram for indicating the variation of air-conditioner control system of the 5th embodiment.
Fig. 8 is the summary functional block diagram of the air-conditioner control system of sixth embodiment.
Fig. 9 is the outline flowchart of the processing executed on the air-conditioner control system that any of Fig. 6~Fig. 8 is shown.
Figure 10 is the stream summary of the processing executed on the air-conditioner control system that any of Fig. 1, Fig. 3~Fig. 5 is shown
Cheng Tu.
Specific embodiment
Fig. 1 is the summary functional block diagram of the air-conditioner control system 1 of first embodiment.
The processors such as CPU, GPU that the computers such as numerical control device, unit computer, master computer, Cloud Server have
The movement for coming each portion of control device according to each system and program is achieved in each functional block shown in fig. 1.
The air-conditioner control system 1 of present embodiment at least has: environmental management portion 100, manages the observation as state
With the environment (room etc. for being provided with air conditioner 130) of reasoning object；Reasoning processing unit 200 pushes away the state of environment
Reason；And learning model storage unit 300, it stores and manages multiple learning models.The air-conditioner control system 1 is also equipped with: air-conditioning
Control unit 400, result obtained from processing unit 200 makes inferences the state of environment by inference control air-conditioning；And
Learning model generating unit 500 generates and updates storage the learning model in learning model storage unit 300.
The specified ring for being provided with the air conditioner 130 controlled by air-conditioner control system 1 in the environmental management portion 100 of present embodiment
The control condition of air conditioner 130 in border (environment managed by air-conditioner control system 1) and acquirement indicate the state of environment
Quantity of state.The environmental management portion 100 can for example be mounted on the central management device of air conditioner, be provided with by air-conditioner control system
On numerical control device in the environment of the environment of 1 management etc..Environmental management portion 100 is connect with the network 150 of wire/wireless, and is set
The machine 120 being placed in the environment managed by air-conditioner control system 1, the air conditioner 130 controlled by air-conditioner control system 1 are counted
According to exchange.
Item in the specified environment managed by air-conditioner control system 1 of the condition specifying part 110 that environmental management portion 100 has
Part (operational situation etc. of the setting situation and machine 120 of machine 120 and air conditioner 130).
Such as shown in Fig. 2, the setting situation of machine 120 and air conditioner 130 is shown as, and will be managed by air-conditioner control system 1
Environment be divided into multiple regions, the machine 120 of heat source how is provided as in each region.Such as Fig. 2 shows example
In, the setting situation of machine 120 and air conditioner 130 is the machine 120 that heat source is provided as in region 2,7,11,12, in area
Air conditioner 130 is set in domain 5.Environmental management portion 100 for example can also according to by operating personnel through not shown input and output
Setting that device carries out etc. obtains the setting situation of this machine 120 and air conditioner 130.
Fortune of the operational situation of machine 120 as each machine 120 operated in the environment managed by air-conditioner control system 1
Turn situation and is obtained.The operational situation of machine 120 for example can as stopping, low-speed running (low temperature), middling speed operating (medium temperature),
Run at high speed (high temperature) etc. be classified as calorific value corresponding with the movement of machine 120 in this wise.It environmental management portion 100 can also root
According to running-active status (internal temperature, main shaft, feed shaft of machine etc. of the machine 120 obtained via network 150 from each machine 120
Movement, load state etc.) determine the operational situation of machine 120.
Condition specifying part 110 specifies (output) to obtain in this way learning model storage unit 300, learning model generating unit 500
The environment managed by air-conditioner control system 1 in condition.Condition specifying part 110 plays the role of following: by environmental management portion 100
Condition in current air conditioner control is notified as selecting the condition of learning model to each of air-conditioner control system 1
Portion.
The quantity of state test section 140 that environmental management portion 100 has is by the state of the environment managed by air-conditioner control system 1
It is detected as quantity of state.For example comprising each machine 120 in the quantity of state of the environment managed by air-conditioner control system 1
Peripheral temperature or internal temperature, operational situation, the temperature in each region of environment managed by air-conditioner control system 1 etc..Quantity of state
Test section 140 by by be set in machine 120 or air conditioner 130 temperature sensor, be arranged and manage by air-conditioner control system 1
Environment each region in the detected value that detects of temperature sensor detected as quantity of state.Quantity of state test section 140
The quantity of state detected is output to reasoning processing unit 200, learning model generating unit 500.
What the observation of reasoning processing unit 200 of present embodiment was obtained from environmental management portion 100 is managed by air-conditioner control system 1
Environment state, and the environment managed come reasoning by air-conditioner control system 1 according to the result observed.Reasoning processing unit
200 such as can be installed on air conditioning control device, unit computer, master computer, Cloud Server or machine learning device on.
The characteristic quantity generating unit 210 that reasoning processing unit 200 has is according to the state detected by quantity of state test section 140
It measures to generate the characteristic quantity for the feature for indicating the environment managed by air-conditioner control system 1.Expression is generated by characteristic quantity generating unit 210
The environment managed by air-conditioner control system 1 feature characteristic quantity as airconditioning control portion 400 airconditioning control judgement material
Material is useful information.In addition, indicating the spy of the environment managed by air-conditioner control system 1 generated by characteristic quantity generating unit 210
The characteristic quantity of sign becomes input data when aftermentioned reasoning and calculation portion 220 is made inferences using learning model.It indicates by feature
The characteristic quantity of the feature for the environment managed by air-conditioner control system 1 that amount generating unit 210 generates for example can be by scheduled
Sampling period adopts the peripheral temperature of each machine 120 detected by quantity of state test section 140 in the past scheduled period
The characteristic quantity that sample obtains is also possible to the past pre- of the operational situation of each machine 120 detected by quantity of state test section 140
Peak value during fixed, alternatively, it is also possible to be will be from each regional temperature that quantity of state test section 140 detects to time series
Frequency field carries out integral transformation or is standardized or complies with transmission function or Xiang Teding for amplitude or power density
Time or Frequency and Amplitude carry out the combination of the signal processing of dimension reduction etc..Characteristic quantity generating unit 210 is by reasoning and calculation portion 220
The mode handled pre-process to the quantity of state detected by quantity of state test section 140 and normalization.
The reasoning and calculation portion 220 that reasoning processing unit 200 has according to from learning model storage unit 300 according to by current
The condition learning model selected in the environment of the management of air-conditioner control system 1 and the spy generated by characteristic quantity generating unit 210
Sign amount makes inferences the control method of the air conditioner 130 in the environment managed by air-conditioner control system 1.By that will be stored in
Learning model in learning model storage unit 300 is applied to execute the platform of the reasoning processing of machine learning, is achieved in
Reasoning and calculation portion 220.Reasoning and calculation portion 220 can be used for example multilayer neural network and make inferences processing, alternatively, it is also possible to make
Use Bayesian network, support vector machines, mixed Gauss model etc. as machine learning and well known learning algorithm makes inferences place
Reason.The learning algorithm such as also can be used supervised learning, unsupervised learning, intensified learning of reasoning and calculation portion 220 makes inferences
Processing.In addition, reasoning and calculation portion 220 can also execute the processing of the reasoning based on a variety of learning algorithms respectively.Reasoning and calculation portion 220
The machine learning device based on the learning model selected from the learning model storage unit 300 of machine learning is constituted, executes to be used as and be somebody's turn to do
The input data of machine learning device and used the reasoning of the characteristic quantity generated by characteristic quantity generating unit 210 to handle, thus to by
The control method for the air conditioner 130 in environment that air-conditioner control system 1 manages makes inferences.It is pushed away as by reasoning and calculation portion 220
The control method of the air conditioner 130 for the result managed out is for example also possible to be set in the environment managed by air-conditioner control system 1
The set temperature of each air conditioner 130, wind direction (including curling wobbling action etc.) etc..
The learning model storage unit 300 of present embodiment can store with by condition specifying part 110 specify by air-conditioning control
The multiple learning models obtained after the combination for the condition in environment that system 1 processed manages is associated.Learning model storage unit 300
Can such as be installed on environmental management portion 100, unit computer, master computer, Cloud Server, on database server.
It is stored in learning model storage unit 300 and is managed with what is specified by condition specifying part 110 by air-conditioner control system 1
Environment in condition (setting situation and operational situation of factory's inner machine etc.) combination it is associated after obtained multiple study
Model 1,2 ..., N.Condition (the setting situation of factory's inner machine in this signified environment managed by air-conditioner control system 1
With operational situation etc.) combination indicate to enumerate related combination, such as Fig. 2 with value acquired by each condition, the range of value, value
It is shown, make to be managed by air-conditioner control system 1 it is environmental modeling in the case where, such as can by using the situation in each region as
Matrix [nothing, machine (high temperature), nothing, nothing, air conditioner, nothing, machine (medium temperature), nothing, nothing, nothing, machine (medium temperature), machine of element
(stopping)] it is used as one of the combination of condition in the environment managed by air-conditioner control system 1.
The learning model conduct being stored in learning model storage unit 300, which can be constituted, to be suitable in reasoning and calculation portion 220
The information of a learning model of reasoning processing stored.When the learning model example being stored in learning model storage unit 300
In the case where using the learning model of the learning algorithm of multilayer neural network in this way, each layer neuron (perceptron) can be stored as
Weight parameter etc. between quantity, each layer neuron (perceptron), in addition, being using the learning algorithm of Bayesian network
In the case where practising model, it can be stored as constituting the transition probabilities etc. between the node and node of Bayesian network.It is stored in
Practising each learning model in model storage unit 300 can be the learning model for using identical learning algorithm, alternatively, it is also possible to be to make
With the learning model of different learning algorithms, if it is possible to be used in reasoning and calculation portion 220 reasoning processing be then also possible to using
The learning model of any learning algorithm.
Learning model storage unit 300 can close the combination of the condition in the environment managed by an air-conditioner control system 1
Join a learning model and stored, alternatively, it is also possible to the condition in the environment managed by an air-conditioner control system 1
Combination association has used the learning model of more than two different learning algorithms and has been stored.Learning model storage unit 300 can also
It is used with each combination association of the condition in the environment managed by multiple air-conditioner control systems 1 that is overlapped to the range of the combination
The learning models of different learning algorithms is simultaneously stored.At this point, learning model storage unit 300 is directed to by air-conditioner control system 1
The corresponding learning model of the combination of the condition in environment managed, the kind of processing capacity and learning algorithm needed for further determining
Thus the use conditions such as class for example can select can be performed to the combination of the condition in the environment managed by air-conditioner control system 1
Reasoning processing or the different corresponding learning model in reasoning and calculation portion 220 of processing capacity.
When being received externally the combined learning model comprising the condition in the environment that is managed by air-conditioner control system 1
When read/write is requested, group of the learning model storage unit 300 to the condition in the environment managed with the air-conditioner control system 1
The learning model stored in association is closed to be read out/be written.At this point, can also be in the read/write request of learning model
Comprising can be by the information for the reasoning processing, processing capacity that reasoning and calculation portion 220 executes.In this case, learning model is deposited
Storage portion 300 is to the combination of the condition in the environment managed with air-conditioner control system 1 and can be executed by reasoning and calculation portion 220
Reasoning processing or processing capacity it is associated after obtained learning model be read out/be written.Learning model storage unit 300 can also
With with the following functions: requesting the read/write from external learning model, according to what is specified by condition specifying part 110
Condition is read out/is written to learning model associated with condition (combination).By the way that this function is arranged, to reasoning
Calculation part 220, learning model generating unit 500 do not need study of the setting requirements based on the condition specified by condition specifying part 110
The function of model.
In addition, learning model storage unit 300 can also be added the learning model generated by learning model generating unit 500
It is close and store, the learning model encrypted when reading learning model by reasoning and calculation portion 220 is decrypted.
In the environment that airconditioning control portion 400 is managed according to the air-conditioner control system 1 inferred by reasoning processing unit 200
The control method of air conditioner 130 controls the movement for the air conditioner 130 being set in the environment managed by air-conditioner control system 1.It should
Airconditioning control portion 400 for example sends each air conditioner 130 for control instruction via network 150, thus controls each air conditioner 130.
In addition, the airconditioning control portion 400 can also be via the communication path (infrared ray, other wireless means etc.) different from network 150
To control each air conditioner 130.
Learning model generating unit 500 is according in the environment managed by air-conditioner control system 1 specified by condition specifying part 110
Condition and indicate the environment managed by air-conditioner control system 1 generated by characteristic quantity generating unit 210 feature characteristic quantity,
It generates or updates (machine learning) and be stored in the learning model in learning model storage unit 300.The learning model generating unit 500
It selects to generate according to the condition in the environment managed by air-conditioner control system 1 specified by condition specifying part 110 or be modernized into
The learning model of object, according to the feature for the environment managed by air-conditioner control system 1 for indicating to be generated by characteristic quantity generating unit 210
Characteristic quantity, machine learning is carried out to the learning model selected.The opportunity learnt by learning model generating unit 500 is for example
Inferior progress the case where changing the setting of each air conditioner 130 in a manual manner by operating personnel.In this case, learning model
The learning model that generating unit 500 will be selected according to the condition in the environment managed by air-conditioner control system 1 will be indicated by spy
The characteristic quantity of the feature for the environment managed by air-conditioner control system 1 that sign amount generating unit 210 generates is set as state variable, thus raw
At or update the learning model that the set temperature of each air conditioner 130 and wind direction etc. be set as label data by (machine learning).
Learning model generating unit 500 in the environment that is managed specified by condition specifying part 110 by air-conditioner control system 1
Condition (combination) associated learning model be not stored in learning model storage unit 300 in the case where, give birth to again
At learning model associated with above-mentioned condition (combination), on the other hand, with by condition specifying part 110 specify by air-conditioning
Condition (combination) associated learning model in environment that control system 1 manages is stored in learning model storage unit 300
In in the case where, update the learning model by carrying out machine learning to the learning model.Learning model generating unit 500 is being learned
Practise model storage unit 300 in be stored with it is multiple with by condition specifying part 110 specify the environment managed by air-conditioner control system 1 in
Condition (combination) associated learning model in the case where, can to each learning model carry out machine learning, in addition,
Can according to can by study processing and processing capacity that learning model generating unit 500 executes, only to a part learning model
Carry out machine learning.
The learning model being stored in learning model storage unit 300 can also be changed by learning model generating unit 500,
Generate new learning model.As the example that learning model generating unit 500 is changed learning model, such as illustrate distillation model
Generation.Distillation model is to export using for obtained from the input for being mounted with to learn the machine learning device for finishing model at it
Study obtained from being learnt since 1 in its machine learning device finishes model.Learning model generating unit 500 can will be via
The distillation model that this process (referred to as distillation process) obtains is stored in learning model storage unit 300 as new learning model
And it uses.In general, the size of distillation model, which is less than original study, finishes model, but mould is finished with original study due to having
The equal accuracy of type, therefore be suitable for being distributed to other computers via network etc..It learns as 500 Duis of learning model generating unit
Other examples that model is changed are practised, there is the integration of learning model.Learning model generating unit 500 can also with by airconditioning control
The similar feelings of the construction for more than two learning models that the condition (combination) in environment that system 1 manages stores in association
Under condition, such as in the case where the value of each weight parameter is in pre-determined predetermined threshold, will be with these learning model phases
After condition (combination) in the associated environment managed by air-conditioner control system 1 is integrated, structure correspondingly is stored
Make any one in similar more than two learning models.
Fig. 3 is the summary functional block diagram of the air-conditioner control system 1 of second embodiment.
In the air-conditioner control system 1 of present embodiment, each functional block is installed on an air conditioning control device 2 (in sky
Central management device, numerical control device of tune machine etc. are upper to be constructed) in.Pass through this structure, the air-conditioner control system 1 of present embodiment
According to setting situation, the operational situation of the machine 120 with the environment managed by air-conditioner control system 1, different study moulds is used
Type makes inferences the control method of the air conditioner 130 in the environment managed by air-conditioner control system 1, to by air-conditioner control system
Air conditioner 130 in the environment of 1 management is controlled.In addition, be able to use 2 generations of an air conditioning control device/update with by
The corresponding each learning model of condition in environment that air-conditioner control system 1 manages.
Fig. 4 is the summary functional block diagram of the air-conditioner control system 1 of third embodiment.
In the air-conditioner control system 1 of present embodiment, by environmental management portion 100, reasoning processing unit 200 and air-conditioning control
Portion 400 processed is installed on air conditioning control device 2, in addition, learning model storage unit 300 and learning model generating unit 500 are installed
In on the machine learning device 3 being connected via standard interface and network with air conditioning control device 2.Machine learning device 3 can also
Be installed on unit computer, master computer, Cloud Server, on database server.It, can be in air-conditioning control by this structure
The reasoning processing that the model for having used study to finish as lighter processing is executed on device 2 processed, can fill in machine learning
Generation/update the processing for executing the learning model of the processing as specific gravity on 3 is set, therefore air conditioning control device 2 can not be interfered
Original movement and the utilization for realizing air-conditioner control system 1.
Fig. 5 is the summary functional block diagram of the air-conditioner control system 1 of the 4th embodiment.
In the air-conditioner control system 1 of present embodiment, environmental management portion 100 is installed on air conditioning control device 2, it will
Reasoning and calculation portion 220, learning model storage unit 300, learning model generating unit 500 are installed on via standard interface and network and sky
On the machine learning device 3 for adjusting control device 2 to be connected.In addition, separately preparing airconditioning control portion 400.In addition, in this embodiment party
In the air-conditioner control system 1 of formula, the quantity of state detected by quantity of state test section 140, which is assumed to be, can be directly used for reasoning and calculation
The data that the reasoning processing in portion 220, generation/update of the learning model of learning model generating unit 500 are handled, to omit feature
Measure the structure of generating unit 210.By this structure, the model for having used study to finish can be executed on machine learning device 3
Reasoning processing and learning model generation/update processing, therefore movement that air conditioning control device 2 can not be interfered original and reality
The utilization of existing air-conditioner control system 1.
Fig. 6 is the summary functional block diagram of the air-conditioner control system 1 of the 5th embodiment.
In the air-conditioner control system 1 of present embodiment, each functional block is installed on an air conditioning control device 2.This
Outside, in the air-conditioner control system of present embodiment 1, in learning model storage unit 300 it is stored with by airconditioning control system
The learning model that the multiple study obtained after the combination of condition in the environment of 1 management of system is associated finish, it is assumed that for do not generate/
Renewal learning model, and omit the structure of learning model generating unit 500.Pass through this structure, the airconditioning control of present embodiment
System 1 is for example according to setting situation, the operational situation of the machine 120 of the environment managed by air-conditioner control system 1, using different
Learning model makes inferences the control method of the air conditioner 130 in the environment managed by air-conditioner control system 1, to by air-conditioning
The air conditioner 130 in environment that control system 1 manages is controlled.In addition, will not optionally renewal learning model, therefore for example
It can be used as the structure for the air conditioning control device 2 for being delivered to customer.
Fig. 7 is the summary functional block diagram for indicating the variation of air-conditioner control system 1 of the 5th embodiment.
In the air-conditioner control system 1 of present embodiment, shows and store learning model in the 5th embodiment (Fig. 6)
Portion 300 is mounted on the example on the external memory 4 being connected with air conditioning control device 2.In this variation, by large capacity
Learning model is stored in external register 4, thus, it is possible to use many learning models, and can not be come via network etc.
It is effective in the case where reading learning model, therefore need real-time in reasoning processing.
Fig. 8 is the summary functional block diagram of the air-conditioner control system 1 of sixth embodiment.
In the air-conditioner control system 1 of present embodiment, environmental management portion 100 is installed on air conditioning control device 2, it will
Reasoning and calculation portion 220, learning model storage unit 300, which are installed on, to be connected via standard interface and network with air conditioning control device 2
Machine learning device 3 on.Machine learning device 3 can also be installed on unit computer, master computer, Cloud Server, data
On the server of library.In addition, in the air-conditioner control system 1 of present embodiment, in learning model storage unit 300 it is stored with
The multiple study obtained after the combination of the condition in environment managed by air-conditioner control system 1 is associated finish learning model, false
It is set as not generating/renewal learning model, to omit the structure of learning model generating unit 500.In addition, in the sky of present embodiment
It adjusts in control system 1, the quantity of state detected by quantity of state test section 140, which is assumed to be, can be directly used for reasoning and calculation portion 220
Reasoning processing data, omit characteristic quantity generating unit 210 structure.Pass through this structure, the airconditioning control of present embodiment
System 1 is for example according to setting situation, the operational situation of the machine 120 of the environment managed by air-conditioner control system 1, using different
Learning model makes inferences the control method of the air conditioner 130 in the environment managed by air-conditioner control system 1, to by air-conditioning control
The air conditioner 130 in environment that system 1 processed manages is controlled.In addition, will not optionally renewal learning model, therefore for example can
Enough structures as the air conditioning control device 2 for being delivered to customer use.
Fig. 9 is the outline flowchart of the processing executed on air-conditioner control system 1 of the invention.
Do not have (the five, the 6th in the case where renewal learning model in flowchart illustration air-conditioner control system 1 shown in Fig. 9
Embodiment) process flow.
Condition in the specified environment managed by air-conditioner control system 1 of [step SA01] condition specifying part 110.
The state for the environment that the detection of [step SA02] quantity of state test section 140 is managed by air-conditioner control system 1 is as shape
State amount.
[step SA03] characteristic quantity generating unit 210 according to the quantity of state detected in step SA02, generate indicate by
The characteristic quantity of the feature for the environment that air-conditioner control system 1 manages.
What the reasoning and calculation portion 220 [step SA04] was specified from the selection of learning model storage unit 300 and in step SA01
The corresponding learning model of condition in the environment managed by air-conditioner control system 1 is as the learning model for reasoning and reading.
The reasoning and calculation portion 220 [step SA05] is according to the learning model read in step SA04 and in step SA03
The characteristic quantity of generation makes inferences the control method of the air conditioner 130 in the environment managed by air-conditioner control system 1.
The airconditioning control portion 400 [step SA06] is controlled according to the control method of the air-conditioning inferred in step SA05
Air-conditioning.
Figure 10 is the outline flowchart of the processing executed in air-conditioner control system 1 of the invention.
Flowchart illustration shown in Figure 10 generates and in the case where renewal learning model (first in air-conditioner control system 1
~the four embodiment) process flow.
Condition in the specified environment managed by air-conditioner control system 1 of [step SB01] condition specifying part 110.
The state for the environment that the detection of [step SB02] quantity of state test section 140 is managed by air-conditioner control system 1 is as shape
State amount.
[step SB03] characteristic quantity generating unit 210 according to the quantity of state detected in step SB02, generate indicate by
The characteristic quantity of the feature for the environment that air-conditioner control system 1 manages.
What the reasoning and calculation portion 220 [step SB04] was specified from the selection of learning model storage unit 300 and in step SB01
The corresponding learning model of condition in the environment managed by air-conditioner control system 1 is as the learning model for reasoning and reading.
Whether [step SB05] learning model generating unit 500 judges to generate in learning model storage unit 300 and in step
The corresponding study of the condition in the environment managed by air-conditioner control system 1 specified in rapid SB01 finishes learning model.It is generating
In the case that study finishes learning model, processing is transferred to step SB07, the case where no generation study finishes learning model
Under, processing is transferred to step SB06.
[step SB06] learning model generating unit 500 is generated and is updated according to the characteristic quantity generated in step SB03
Learning model corresponding with the condition in the environment managed by air-conditioner control system 1 specified in step SB01, processing transfer
To step SB01.
The reasoning and calculation portion 220 [step SB07] is according to the learning model read in step SB04 and in step SB03
The characteristic quantity of generation makes inferences the control method of the air conditioner 130 in the environment managed by air-conditioner control system 1.
The airconditioning control portion 400 [step SB08] is controlled according to the control method of the air-conditioning inferred in step SB05
Air-conditioning.
It this concludes the description of embodiments of the present invention, but the example that the present invention is not limited to the above embodiments, lead to
It crosses and applies change appropriate and can implement in various ways.
Claims (12)
1. a kind of air-conditioner control system controls the air conditioner being provided in the environment of at least one machine, which is characterized in that
The air-conditioner control system has:
Condition specifying part specifies the condition in above-mentioned environment；
Quantity of state test section, detection indicate the quantity of state of the state of above-mentioned environment；
Reasoning and calculation portion, according to above-mentioned quantity of state come the control method of the above-mentioned air conditioner in the above-mentioned environment of reasoning；
Airconditioning control portion controls above-mentioned air conditioner according to the control method inferred by above-mentioned reasoning and calculation portion；
Learning model generating unit, is generated or renewal learning model by using the machine learning of above-mentioned quantity of state；And
Learning model storage unit, by least one learning model generated by above-mentioned learning model generating unit and by above-mentioned condition
The combination of the specified condition of specifying part is stored in association,
Above-mentioned reasoning and calculation portion is according to the condition in the above-mentioned environment specified by above-mentioned condition specifying part, from being stored in above-mentioned study
At least one learning model is selectively used in learning model in model storage unit, is calculated by above-mentioned air-conditioner control system 1
The control method of above-mentioned air conditioner in the environment of management.
2. air-conditioner control system according to claim 1, which is characterized in that
The air-conditioner control system is also equipped with: characteristic quantity generating unit, according to the quantity of state detected by above-mentioned quantity of state test section
Come generate make above-mentioned environment have feature characteristic quantity,
Above-mentioned reasoning and calculation portion according to features described above amount come the control method of the above-mentioned air conditioner in the above-mentioned environment of reasoning,
Above-mentioned learning model generating unit is generated by using the machine learning of features described above amount or renewal learning model.
3. air-conditioner control system according to claim 1 or 2, which is characterized in that
Above-mentioned learning model generating unit passes through changing for the existing learning model for implementing to be stored for above-mentioned learning model storage unit
Become, generates new learning model.
4. air-conditioner control system according to any one of claims 1 to 3, which is characterized in that
Above-mentioned learning model generating unit learning model generated is encrypted and is stored by above-mentioned learning model storage unit, will be led to
The learning model encrypted when crossing above-mentioned reasoning and calculation portion reading learning model is decoded.
5. a kind of air-conditioner control system controls the air conditioner being provided in the environment of at least one machine, which is characterized in that
The air-conditioner control system has:
Condition specifying part specifies the condition in above-mentioned environment；
Quantity of state test section, detection indicate the quantity of state of above-mentioned environment；
Reasoning and calculation portion, according to above-mentioned quantity of state come the control method of the above-mentioned air conditioner in the above-mentioned environment of reasoning；
Airconditioning control portion controls above-mentioned air conditioner according to the control method inferred by above-mentioned reasoning and calculation portion；And
Learning model storage unit stores and the combination of the condition in above-mentioned environment at least one study mould associated in advance
Type,
Above-mentioned reasoning and calculation portion is according to the condition in the above-mentioned environment specified by above-mentioned condition specifying part, from being stored in above-mentioned study
At least one learning model is selectively used in learning model in model storage unit, calculates the above-mentioned air-conditioning in above-mentioned environment
The control method of machine.
6. air-conditioner control system according to claim 5, which is characterized in that
The air-conditioner control system is also equipped with: characteristic quantity generating unit, and being generated according to above-mentioned quantity of state makes above-mentioned environment have spy
The characteristic quantity of sign,
The above-mentioned sky in environment that above-mentioned reasoning and calculation portion is managed by above-mentioned air-conditioner control system 1 come reasoning according to features described above amount
The control method of tune machine.
7. a kind of air conditioning control device, which is characterized in that
The air conditioning control device has condition specifying part described in any one in claim 1~6 and quantity of state detection
Portion.
8. a kind of air conditioning control method, which is characterized in that
The air conditioning control method executes following steps:
Specified control is provided with the step of condition of the air conditioner in the environment of at least one machine；
Detection indicates the step of quantity of state of above-mentioned environment；
According to above-mentioned quantity of state come the above-mentioned air conditioner in the above-mentioned environment of reasoning control method the step of；
According to above-mentioned control method come the step of controlling above-mentioned air conditioner；And
It is generated by using the machine learning of above-mentioned quantity of state or the step of renewal learning model,
In the combination of condition of the above-mentioned inference step from above-mentioned environment at least one above-mentioned learning model associated in advance
The learning model that selection is used according to the condition in specified above-mentioned environment in the step of specifying above-mentioned condition, uses choosing
The learning model at the place of selecting calculates the control method of the above-mentioned air conditioner in above-mentioned environment.
9. air conditioning control method according to claim 8, which is characterized in that
The air conditioning control method also executes according to above-mentioned quantity of state the step of generating the characteristic quantity for making above-mentioned environment have feature,
Above-mentioned inference step according to features described above amount come the control method of the above-mentioned air conditioner in the above-mentioned environment of reasoning,
The step of generating or updating above-mentioned learning model generates by using the machine learning of features described above amount or updates
Practise model.
10. a kind of air conditioning control method, which is characterized in that
The air conditioning control method executes following steps:
Specified control is provided with the step of condition of the air conditioner in the environment of at least one machine；
Detection indicates the step of quantity of state of above-mentioned environment；
According to above-mentioned quantity of state come the above-mentioned air conditioner in the above-mentioned environment of reasoning control method the step of；And according to above-mentioned control
Method processed come the step of controlling above-mentioned air conditioner,
It is selected in combination at least one learning model associated in advance of condition of the above-mentioned inference step from above-mentioned environment
According to the condition in specified above-mentioned environment in the specified above-mentioned condition the step of come using learning model, using selecting
Learning model calculate the control method of the above-mentioned air conditioner in above-mentioned environment.
11. air conditioning control method according to claim 10, which is characterized in that
The air conditioning control method also executes according to above-mentioned quantity of state the step of generating the characteristic quantity for making above-mentioned environment have feature,
Above-mentioned inference step is according to features described above amount come the control method of the above-mentioned air conditioner in the above-mentioned environment of reasoning.
12. a kind of learning model collection, make each and the sky in the environment for being provided at least one machine of multiple learning models
The combination for the condition that tune machine is controlled is associated to be formed,
The shape of the above-mentioned environment of expression carried out under conditions of according to each of above-mentioned multiple learning models in above-mentioned environment
State amount come the learning model that generates or update,
A learning model, the study selected are selected according to the condition being set in environment from above-mentioned multiple learning models
Model is used for the processing made inferences to the control method of the above-mentioned air conditioner in above-mentioned environment.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2017194043A JP2019066135A (en) | 2017-10-04 | 2017-10-04 | Air-conditioning control system |
| JP2017-194043 | 2017-10-04 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109612037A true CN109612037A (en) | 2019-04-12 |
Family
ID=65728135
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811158832.0A Pending CN109612037A (en) | 2017-10-04 | 2018-09-30 | Air-conditioner control system |
Country Status (4)
| Country | Link |
|---|---|
| US (1) | US20190101305A1 (en) |
| JP (1) | JP2019066135A (en) |
| CN (1) | CN109612037A (en) |
| DE (1) | DE102018007640A1 (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114341564A (en) * | 2019-08-23 | 2022-04-12 | 大金工业株式会社 | Air conditioner control system, air conditioner, and machine learning device |
| CN114761732A (en) * | 2019-12-13 | 2022-07-15 | 三菱电机株式会社 | Model sharing system, model management device, and control device for air conditioning device |
| CN115280077A (en) * | 2020-03-27 | 2022-11-01 | 三菱电机株式会社 | Learning device and inference device for air conditioner control |
| CN116097046A (en) * | 2020-09-04 | 2023-05-09 | 大金工业株式会社 | Generating method, program, information processing device, information processing method, and learning-completed model |
| CN116209964A (en) * | 2020-09-23 | 2023-06-02 | 大金工业株式会社 | Information processing device, information processing method, and program |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP7389314B2 (en) * | 2019-03-05 | 2023-11-30 | ダイキン工業株式会社 | Air conditioner control system |
| US11448412B2 (en) * | 2019-04-02 | 2022-09-20 | Lg Electronics Inc. | Air conditioner with an artificial intelligence |
| JP7392394B2 (en) * | 2019-10-31 | 2023-12-06 | 株式会社富士通ゼネラル | Air conditioning system and air conditioner |
| EP4130595A4 (en) * | 2020-03-25 | 2024-04-03 | Daikin Industries, Ltd. | Air-conditioning control system |
| JP7212654B2 (en) * | 2020-03-25 | 2023-01-25 | ダイキン工業株式会社 | air conditioning control system |
| JP6997401B1 (en) * | 2020-09-04 | 2022-01-17 | ダイキン工業株式会社 | Generation method, program, information processing device, information processing method, and trained model |
| KR102537407B1 (en) * | 2021-06-03 | 2023-05-26 | (주)더블유에스에이 | Vacuum Pump Smart AI Module |
| WO2024034110A1 (en) * | 2022-08-12 | 2024-02-15 | 三菱電機株式会社 | Control device, chiller system, simulation unit, actual machine unit, and control method |
| US20240191898A1 (en) * | 2022-12-12 | 2024-06-13 | Computime Ltd. | Transfer Learning Model for Newly Setup Environment |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS637703A (en) | 1986-06-27 | 1988-01-13 | 井関農機株式会社 | Posture control apparatus of tractor working machine |
| JP2899484B2 (en) * | 1991-08-09 | 1999-06-02 | 松下電器産業株式会社 | Pattern classification device, environment recognition device and air conditioner |
| JPH05240974A (en) * | 1991-10-04 | 1993-09-21 | Sanyo Electric Co Ltd | Controlling device for machinery and apparatus |
| JP6420565B2 (en) | 2014-04-18 | 2018-11-07 | 株式会社竹中工務店 | Air conditioning system |
| TWI546506B (en) * | 2014-12-04 | 2016-08-21 | 台達電子工業股份有限公司 | Controlling system for environmental comfort value and controlling method of the controlling system |
| KR20170079109A (en) * | 2015-12-30 | 2017-07-10 | 에스케이하이닉스 주식회사 | Circuit for generating periodic signal and memory device including same |
| JP6926429B2 (en) * | 2016-09-27 | 2021-08-25 | 日本電気株式会社 | Data processing equipment, data processing methods, and programs |
| US11137161B2 (en) * | 2017-03-30 | 2021-10-05 | Samsung Electronics Co., Ltd. | Data learning server and method for generating and using learning model thereof |
-
2017
- 2017-10-04 JP JP2017194043A patent/JP2019066135A/en active Pending
-
2018
- 2018-09-27 US US16/144,509 patent/US20190101305A1/en not_active Abandoned
- 2018-09-27 DE DE102018007640.8A patent/DE102018007640A1/en not_active Withdrawn
- 2018-09-30 CN CN201811158832.0A patent/CN109612037A/en active Pending
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114341564A (en) * | 2019-08-23 | 2022-04-12 | 大金工业株式会社 | Air conditioner control system, air conditioner, and machine learning device |
| CN114761732A (en) * | 2019-12-13 | 2022-07-15 | 三菱电机株式会社 | Model sharing system, model management device, and control device for air conditioning device |
| CN114761732B (en) * | 2019-12-13 | 2024-03-19 | 三菱电机株式会社 | Model sharing system, model management device, and control device for air conditioner |
| CN115280077A (en) * | 2020-03-27 | 2022-11-01 | 三菱电机株式会社 | Learning device and inference device for air conditioner control |
| CN115280077B (en) * | 2020-03-27 | 2024-03-08 | 三菱电机株式会社 | Learning device and reasoning device for air conditioner control |
| CN116097046A (en) * | 2020-09-04 | 2023-05-09 | 大金工业株式会社 | Generating method, program, information processing device, information processing method, and learning-completed model |
| CN116097046B (en) * | 2020-09-04 | 2023-12-08 | 大金工业株式会社 | Generating method, information processing apparatus, information processing method, and learning-completed model |
| CN116209964A (en) * | 2020-09-23 | 2023-06-02 | 大金工业株式会社 | Information processing device, information processing method, and program |
| CN116209964B (en) * | 2020-09-23 | 2024-04-02 | 大金工业株式会社 | Information processing device, information processing method, and program product |
Also Published As
| Publication number | Publication date |
|---|---|
| DE102018007640A1 (en) | 2019-04-04 |
| US20190101305A1 (en) | 2019-04-04 |
| JP2019066135A (en) | 2019-04-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109612037A (en) | Air-conditioner control system | |
| Huang et al. | Deep reinforcement learning based preventive maintenance policy for serial production lines | |
| Thrift | Fuzzy Logic Synthesis with Genetic Algorithms. | |
| US5485390A (en) | Inductive-deductive process design for machined parts | |
| Mahadevan et al. | Optimizing production manufacturing using reinforcement learning. | |
| CN103440368B (en) | A kind of multi-model dynamic soft measuring modeling method | |
| Moore et al. | Indirect adaptive fuzzy control | |
| CN109613886A (en) | Thermal displacement correction system | |
| Zhu et al. | A digital twin–driven method for online quality control in process industry | |
| CN109613887A (en) | Thermal displacement correction system | |
| CN108693822A (en) | Control device, storage medium, control system and control method | |
| Saldivar et al. | Identifying smart design attributes for Industry 4.0 customization using a clustering Genetic Algorithm | |
| WO1998037465A1 (en) | Adaptive objet-oriented optimization software system | |
| JP2022065773A (en) | Control device, controller, control system, control method, and control program | |
| CN115544775A (en) | Digital twin workshop multi-dimensional multi-level model construction and dynamic configuration method | |
| CN113821903B (en) | Temperature control method and equipment, modularized data center and storage medium | |
| Ignatyev et al. | System for automatic adjustment of intelligent controller parameters | |
| CN110837857A (en) | Industrial electricity load prediction method, system and storage medium thereof | |
| CN115097737B (en) | Multi-level regulation and control method capable of being re-entered into manufacturing system | |
| CN112184007A (en) | Workshop equipment remote diagnosis method based on digital twins | |
| Shen et al. | Digital twins in additive manufacturing: a state-of-the-art review | |
| CN114265363B (en) | Intelligent optimization method and system for machining path of numerical control machine tool | |
| Würtz et al. | Controlled emergence and self-organization | |
| Mirazimzadeh et al. | Unsupervised clustering approach for recognizing residual stress and distortion patterns for different parts for directed energy deposition additive manufacturing | |
| Dohmen et al. | LongiControl: A Reinforcement Learning Environment for Longitudinal Vehicle Control. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190412 |
|
| WD01 | Invention patent application deemed withdrawn after publication |