CN106681389A - Machine learning apparatus and method, motor control apparatus and motor control system - Google Patents
Machine learning apparatus and method, motor control apparatus and motor control system Download PDFInfo
- Publication number
- CN106681389A CN106681389A CN201610602130.1A CN201610602130A CN106681389A CN 106681389 A CN106681389 A CN 106681389A CN 201610602130 A CN201610602130 A CN 201610602130A CN 106681389 A CN106681389 A CN 106681389A
- Authority
- CN
- China
- Prior art keywords
- mentioned
- motor
- chiller
- control device
- electric motor
- 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.)
- Granted
Links
Classifications
-
- 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
- G05D23/1927—Control of temperature characterised by the use of electric means using a plurality of sensors
- G05D23/193—Control of temperature characterised by the use of electric means using a plurality of sensors sensing the temperaure in different places in thermal relationship with one or more spaces
- G05D23/1931—Control of temperature characterised by the use of electric means using a plurality of sensors sensing the temperaure in different places in thermal relationship with one or more spaces to control the temperature of one space
Landscapes
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Motor Or Generator Cooling System (AREA)
- Feedback Control In General (AREA)
- Control Of Electric Motors In General (AREA)
- Devices That Are Associated With Refrigeration Equipment (AREA)
Abstract
A machine learning apparatus and method, a motor control apparatus and a motor control system are provided. The machine learning apparatus is a machine learning apparatus which learns the operation conditions of a cooling device for cooling a motor or a motor control apparatus. The machine learning apparatus is characterized by comprising a state observer for observing a state variable including at least one of temperature data of the motor and the motor control apparatus at a specific position during operation of the cooling device; a determination data acquisition unit for acquiring determination data that determines a margin of acceptable value of a loss in each of the motor, the motor control apparatus, and the cooling device and a margin of acceptable value of the temperature of each of the motor and the motor control apparatus at the specific position; and a learner for learning the operation conditions of the cooling device in accordance with a training data set constituted of a combination of the state variable and the determination data.
Description
Technical field
The present invention relates to a kind of rote learning device, control device of electric motor, motor control system and rote learning
Method, more particularly to it is a kind of study chiller service condition rote learning device, possess the electronic of rote learning device
Machine control device, motor control system and learning by rote.
Background technology
In motor (motor), along with driving, because of the heating that the iron loss of stator core or the copper loss of coil cause, electricity
The temperature of motivation rises, it is possible to the loss increase of motor or motor damage.Therefore, cool down what is generated to absorb
Heat, it is proposed that for cooling down method (for example, the Japanese Unexamined Patent Publication 5-236704 publication of motor.Hereinafter referred to as " patent is literary
Offer 1 ").
Additionally, the control device of drive motor, with the driving of motor, the power component heating inside control device
And temperature rises, power component is possible to impaired.Therefore, in order to maintain life-span of control device, it is proposed that cooling controller
Method.
Motor cooling described in patent documentation 1 is provided with：Cooling supplies road, its rotation in motor with cold-producing medium
It is axially disposed in rotating shaft；Cold-producing medium squit hole is used in cooling, and it is set to and the line formed by the coil for being wound in stator core
Circle ear is relative, and from cooling cold-producing medium supply road direction coil, ear sprays cooling cold-producing medium；Pump, it is to cooling cold-producing medium
Supply road provides cooling cold-producing medium；And pump control unit, it changes the cold of pump discharge according to the driving condition of motor
But refrigerant amount is used.By the motor cooling described in patent documentation 1, stator core can be efficiently cooled down, thus
Motor entirety can efficiently be cooled down.
However, existing control device of electric motor, with the temperature of motor accordingly, only change the operation of chiller
Rate.Accordingly, there exist following problem：It is difficult to while being defined by the temperature control of motor and control device of electric motor
Below temperature, while reducing the loss of motor, control device of electric motor and chiller.
The content of the invention
It is an object of the invention to provide it is a kind of by rote learning, can be by motor and the temperature of control device of electric motor
Degree is controlled to below the temperature of defined, and the loss of motor, control device of electric motor and chiller can be reduced
Rote learning device, the control device of electric motor for possessing rote learning device, motor control system and learning by rote.
The rote learning device of one embodiment of the invention, learns for cooling down the cold of motor or control device of electric motor
But the service condition of device, in the rote learning device, possesses：State observation portion, it sees in the course of action of chiller
The state for surveying at least one of the temperature data of respective ad-hoc location comprising motor and control device of electric motor becomes
Amount；Judge data acquisition, it obtains the loss for judging motor, control device of electric motor and chiller and electronic
Judgement data of the temperature of the respective ad-hoc location of mechanical, electrical motivation control device relative to the surplus of feasible value；And study portion,
It learns the service condition of chiller according to the training dataset being made up of the combination of state variable and judgement data.
The control device of electric motor of one embodiment of the invention possesses above-mentioned rote learning device, the control device of electric motor
In, it is also equipped with：It is intended to determination section, its result according to obtained by study portion is learnt according to training dataset determines cooling dress
At least one of the group of refrigerant temperature and refrigerant flow of the rotating speed put and the group of run time and chiller group
Command value.
The motor control system of one embodiment of the invention, possesses above-mentioned control device of electric motor and output temperature data
Detector unit.
The learning by rote of one embodiment of the invention, learns for cooling down the cold of motor or control device of electric motor
But the service condition of device, in the learning by rote, comprises the steps：Observe in the course of action of chiller and including
The state variable of at least one of the temperature data of respective ad-hoc location of motor and control device of electric motor；Obtain
For judging, motor, control device of electric motor and the loss of chiller and the temperature of respective ad-hoc location relative to
The judgement data of the surplus of feasible value；And according to the training dataset being made up of the combination of state variable and judgement data, learn
Practise the service condition of chiller.
Description of the drawings
By the explanation of the following embodiment related to accompanying drawing, become the purpose of the present invention, feature and advantage
Definitely.
Fig. 1 is the structure chart of the rote learning device of embodiments of the invention.
Fig. 2 is the structure chart of the motor control system of embodiments of the invention.
Fig. 3 is the structure chart of the motor control system of the 1st variation of embodiments of the invention.
Fig. 4 is the structure chart of the motor control system of the 2nd variation of embodiments of the invention.
Fig. 5 is the structure chart of the motor control system of the 3rd variation of embodiments of the invention.
Fig. 6 is the structure chart of the control device of electric motor of embodiments of the invention.
Fig. 7 is the flow chart for illustrating the action step of the rote learning device of embodiments of the invention.
Fig. 8 is the flow chart for illustrating the action step of the motor control system of embodiments of the invention.
Specific embodiment
Hereinafter, referring to the drawings to the present invention rote learning device, control device of electric motor, motor control system and
Learning by rote is illustrated.
Fig. 1 is the structure chart of the rote learning device of embodiments of the invention.Fig. 2 is the motor of embodiments of the invention
The structure chart of control system.Fig. 6 is the structure chart of the control device of electric motor of embodiments of the invention.
The rote learning device 1 of embodiments of the invention is study for cooling down motor 5 or control device of electric motor 6
The rote learning device of the service condition of chiller 7, possesses state observation portion 2, judges data acquisition 3 and study portion 4.
In Fig. 2, rote learning device is contained in control device of electric motor 6.
State observation portion 2 is observed comprising motor 5 (with reference to Fig. 2) and motor control in the course of action of chiller 7
The state variable of the temperature data of the respective ad-hoc location of device processed 6.
Judge that data acquisition 3 is obtained and judge motor 5, control device of electric motor 6, the respective loss of chiller 7
And the temperature of the respective ad-hoc location of motor 5, control device of electric motor 6 is relative to the resulting number of the surplus of feasible value
According to.
Study portion 4 learns chiller 7 according to the training dataset being made up of the combination of state variable and judgement data
Service condition.
Detector unit (the 8,9) inspection that temperature data is set to be respectively provided on motor 5 and control device of electric motor 6
The coil temperature of the motor 5 measured and the power component temperature of control device of electric motor 6.
Rote learning device 1 is preferably also equipped with：Loss presumption unit 11, it estimates motor 5 and electricity according to temperature data
The loss of motivation control device 6；And loss calculation portion 12, it calculates chiller 7 according to the service condition of chiller 7
Loss.Loss presumption unit 11 and loss calculation portion 12 can also be contained in judgement data acquisition 3.
The loss of chiller 7 is calculated using the rotating speed and run time of chiller 7.Additionally, being based on
In the case of the cooling of liquid, it is also possible to the refrigerant temperature detected using flow and detector unit (10,10').
In addition, detector unit 10 detects the refrigerant temperature in the direction flow through to chiller 7 from motor 5, temperature detection
Element 10' detects the refrigerant temperature in the direction flow through to motor 5 from chiller 7.
The structure chart of the motor control system shown in Fig. 2 shows only possess the example of 1 chiller, but does not limit
In such example, it is also possible to possess multiple stage chiller (Fig. 3, Fig. 4, Fig. 5).
Fig. 3 is the structure chart of the motor control system 100' of the 1st variation of embodiments of the invention.As shown in figure 3,
2 chillers i.e. the 1st chiller 71 and the 2nd chiller 72 can also be set, each chiller cooling He of motor 5
Control device of electric motor 6.
Fig. 4 is the motor control system 100 of the 2nd variation of embodiments of the invention " structure chart.Shown in Fig. 4,
2 chillers i.e. the 1st chiller 71' and the 2nd chiller 72', the 1st chiller 71' cooling motor can be set
5, the 2nd chiller 72' cools down control device of electric motor 6.
Fig. 5 is the motor control system 100 of the 3rd variation of embodiments of the invention " ' structure chart.Shown in Fig. 5,
4 chillers i.e. the 1st chiller 71 ", the 2nd chiller 72 ", the 3rd chiller 73 can also be set " and the 4th is cold
But device 74 ", the 1st chiller 71 " and the 2nd chiller 72 " cooling motor 5, the 3rd chiller 73 " and the 4th cooling dress
Put 74 and " cool down control device of electric motor 6.In addition, the number of units of chiller is not limited to the example shown in Fig. 3~5, it is also possible to
The chiller of 3 or more than 5 is set.
Study portion 4 can also learn service condition according to the training dataset obtained for multiple chillers.
Control device of electric motor 6 possesses according to the return calculating part 14 for judging data calculating return.Study portion 4 possesses function
Update section 15, it updates from current state variable and determines service condition (rotating speed, the fortune of appropriate chiller 7 according to return
Row time etc.) function.
Return calculating part 14 according to judge data, that is, judge motor 5, control device of electric motor 6, chiller 7 it is each
From the total and motor 5 of loss, the temperature of the ad-hoc location of control device of electric motor 6 relative to feasible value surplus and
The result for obtaining, calculates return.
Specifically, returning calculating part 14 can be according to chiller 7, motor 5, the respective damage of control device of electric motor 6
Consumption and chiller 7, motor 5, the temperature of the respective ad-hoc location of control device of electric motor 6 relative to feasible value whether
There is surplus to determine return.For example, chiller 7, motor 5, the respective loss of control device of electric motor 6 aggregate value
Than last time, the value of (before 1 time is attempted) was reduced, also, the respective spy of chiller 7, motor 5 and control device of electric motor 6
In the case that the temperature put of positioning is less than feasible value, accordingly increase return with surplus, chiller 7, motor 5 and
The aggregate value of the respective loss of control device of electric motor 6 increases than the value of last time (1 time attempt before), or chiller 7, electronic
The temperature of machine 5 and the respective ad-hoc location of control device of electric motor 6 be feasible value more than in the case of, it is also possible to subtract return
It is few.In addition, in the present embodiment, with addition to the loss of chiller 7, also from motor 5 and control device of electric motor 6
Loss total calculating return as a example by, but a less side is lost in motor 5 and control device of electric motor 6 can also be without
In the calculating of return.
Function update section 15 preferably uses so-called Q study and carries out intensified learning.Q study is study under certain environment " s "
Value (behavior memory) Q (s, method a) of housing choice behavior " a ".Under certain environment " s ", by Q (s, a) highest behavior " a " choosing
It is selected as optimal behavior.Function update section 15 uses following formula (1) renewal function (behavior memory function Q (st, at))。
Here, Q (st, at) be behavior memory function, stFor the state (environment) of moment t, atFor the behavior of moment t, α is
Learning coefficient, rt+1For return, γ is discount rate.The expected value of behavior memory function representation return.Item with max is in ring
Border st+1Under Q-value when have selected Q-value highest behavior a be multiplied by γ and obtain.
, it is known that learning coefficient and discount rate are set by 0 ＜ α, γ≤1, but here is for simplicity, if by learning coefficient and
Discount rate is set to 1, then represented by following formula (2).
If the newer represents the value Q (s of the optimal behavior under the next ambient condition of Behavior-based control " a "t+1, max
at+1) more than the value Q (s of the behavior " a " under environment " s "t, at), then by Q (st, at) be set to it is larger, if conversely, be less than environment
Value Q (the s of the behavior " a " under " s "t, at), then by Q (st, at) be set to it is less.That is, the value of certain behavior under certain state is made
The value of the optimal behavior being close under the next state being induced by it.For the newer state corresponding to can by train number
According to the state variable that collection is obtained.Additionally, return can be obtained from return calculating part 14.So-called behavior, is the fortune for changing chiller 7
Row condition is rotating speed of chiller 7 etc..Consider behavior memory Q (s for example for each environment " s ", behavior " a "t、at) deposit
Chu Weibiao (hereinafter referred to as behavior memory table).
As shown in fig. 6, causative juristic act and the state that changes indirectly and straight because of behavior in the state of control device of electric motor 6
Connect the state of change.The state changed indirectly because of behavior, including the temperature (coil of the ad-hoc location of control device of electric motor 6
Temperature, temperature of power component etc.), chiller 7, the loss of motor 5, power component.The shape directly changed because of behavior
State, including rotating speed, the run time of chiller 7 of chiller 7.In addition, in the case where the cooling based on liquid is carried out,
The state changed indirectly because of behavior is also including (direction flow through to chiller 7 from motor 5) refrigeration of chiller 7
Agent temperature.Additionally, refrigerant flow and chiller 7 of the state directly changed because of behavior also including chiller 7
(direction flow through to motor 5 from chiller 7) refrigerant temperature.
Study portion 4 updates in subordinate act value table and current state variable and can obtain according to newer and return
The corresponding behavior memory of behavior.
Study portion 4 can also according to motor 5 or mutually isostructural other motor of control device of electric motor 6, electronic
The state variable of machine control device (not shown) and return, regeneration behavior value table.
Then, the control device of electric motor 6 of embodiments of the invention is illustrated.In addition, the present embodiment is described being based on
The situation of the cooling of liquid.The control device of electric motor 6 of embodiments of the invention is the electricity of the rote learning device 1 for possessing above-mentioned
Motivation control device 6, is also equipped with：It is intended to determination section 16, it is tied according to obtained by study portion 4 is learnt according to training dataset
Really, the instruction of the service conditions such as the rotating speed of change chiller 7 is sent.
Study portion 4 according to by current state variable and judging the additional training dataset that the combination of data is constituted,
Relearn the service condition of chiller 7 and update.
Can also be configured to, rote learning device 1 is connected via network with control device of electric motor 6, the Jing of state observation portion 2
Current state variable is obtained by network.
Rote learning device 1 is preferably on Cloud Server.
Rote learning device 1 can also be built in the control device of electric motor 6 of control motor 5.
Motor control system 100 possesses above-mentioned control device of electric motor 6, for cooling down motor 5 and motor control
The chiller 7 of device processed 6 and output temperature data detector unit (8,9,10,10').Control device of electric motor 6
Receive alternating electromotive force, drive motor 5 from alternating current power supply 20.
Chiller 7 is also equipped with controlling the chiller control of the refrigerant flow of chiller 7 and refrigerant temperature
Portion 13, the refrigerant flow and refrigerant temperature of the observation of state observation portion 2 chiller 7.It is cold that chiller 7 is also equipped with detection
But the tachometer 21 of the rotating speed of device 7, detection chiller 7 refrigerant flow effusion meter 22.
Then, the learning by rote of embodiments of the invention is illustrated.The rote learning of embodiments of the invention
Method is study for cooling down the rote learning side of the service condition of the chiller 7 of motor 5 or control device of electric motor 6
Method, wherein, respective certain bits of the observation comprising motor 5 and control device of electric motor 6 in the course of action of chiller 7
The state variable of at least one of the temperature data put, obtains and judges motor 5, control device of electric motor 6, chiller 7
Respective loss and motor 5, the respective ad-hoc location of control device of electric motor 6 temperature relative to feasible value surplus
Judgement data, according to by state variable and judging the training dataset that constitutes of combination of data, the operation of study chiller 7
Condition.
Fig. 7 represents the flow chart for illustrating the action step of the rote learning device of embodiments of the invention.First, exist
In step S101, respective spy of the observation comprising motor 5 and control device of electric motor 6 in the course of action of chiller 7
The state variable of at least one of the temperature data that positioning is put.
Then, in step s 102, obtain and judge motor 5, control device of electric motor 6, the respective damage of chiller 7
Consumption and motor 5, the respective ad-hoc location of control device of electric motor 6 temperature relative to the surplus of feasible value resulting number
According to.
Then, in step s 103, according to the training dataset being made up of the combination of state variable and judgement data, study
The service condition of chiller 7.
Then, the learning by rote of the use motor control system of embodiments of the invention is illustrated.Fig. 8 tables
Show the flow chart for illustrating the action step of the motor control system of embodiments of the invention.First, in step s 201,
Proceed by study.
Then, in step S202, the service condition (rotating speed, run time etc.) of chiller 7 is set.
Then, in step S203, with certain hour drive motor 5.
Then, in step S204, the temperature (coil temperature etc.) of motor 5 is determined, damage is estimated by loss presumption unit 11
Consumption.Also, the temperature of the power component of control device of electric motor 6 is determined, loss is estimated by loss presumption unit 11.Also, determine
The situation (rotating speed, run time etc.) of chiller 7, by the computed losses of loss calculation portion 12.Herein it is assumed that loss presumption unit 11
There is in advance lossy data at each temperature.
Then, in step S205, according to the total and each of each loss (motor 5, chiller 7, power component)
The temperature in portion, carries out return calculating.
Increase total with last time phase specific loss, or in the case that the temperature in each portion becomes more than feasible value, in step
Reduce the fraction of the value of the behavior after change in S206.Afterwards, in step S208, regeneration behavior value table.
On the other hand, reduce total with last time phase specific loss, and in the case that the temperature in each portion is less than feasible value, in step
Increase the fraction of the value of the behavior after change in rapid S207.Afterwards, in step S208, regeneration behavior value table.
However, only for the first time not to value increase and decrease, in step S208, regeneration behavior value table.
Then, it is preferential with the big project of the fraction of behavior memory in step S209, determines to become in subordinate act value table
The more project of the service condition of chiller 7.
According to the service condition of the chiller 7 determined in step S209, returning to step S202 transports chiller 7
Go to make behavior memory become optimal.
As described above, by the rote learning device of embodiments of the invention, possess rote learning device
Control device of electric motor, motor control system and learning by rote, can be by rote learning while will be electronic mechanical, electrical
The temperature control of motivation control device is into below the temperature of regulation, while reduction motor, control device of electric motor and cooling
The loss of device.
By means of the invention it is possible to provide it is a kind of by rote learning can reduce motor, control device of electric motor and
Total rote learning device of the loss of chiller, control device of electric motor, the motor control for possessing rote learning device
System processed and learning by rote.
Claims (17)
1. a kind of rote learning device, learns the operation bar for cooling down the chiller of motor or control device of electric motor
Part, the rote learning device is characterised by possessing：
State observation portion, it is observed comprising above-mentioned motor and above-mentioned motor control in the course of action of above-mentioned chiller
The state variable of at least one of the temperature data of respective ad-hoc location of device processed；
Judge data acquisition, it is obtained for judging above-mentioned motor, above-mentioned control device of electric motor and above-mentioned chiller
Respective loss and above-mentioned motor, the temperature of the respective ad-hoc location of above-mentioned control device of electric motor relative to allowing
The judgement data of the surplus of value；And
Study portion, it learns above-mentioned according to the training dataset being made up of the combination of above-mentioned state variable and above-mentioned judgement data
The service condition of chiller.
2. rote learning device according to claim 1, it is characterised in that
Said temperature data include being set in the detector unit of above-mentioned motor and above-mentioned control device of electric motor respectively
At least one of the coil temperature of the above-mentioned motor for detecting and the power component temperature of above-mentioned control device of electric motor.
3. rote learning device according to claim 1 and 2, it is characterised in that
The rote learning device is also equipped with：
Loss presumption unit, it estimates the loss of above-mentioned motor and above-mentioned control device of electric motor according to said temperature data；
And
Loss calculation portion, it calculates the loss of above-mentioned chiller according to the service condition of above-mentioned chiller.
4. rote learning device according to any one of claim 1 to 3, it is characterised in that
Using the rotating speed of above-mentioned chiller and the refrigerant temperature and cold-producing medium of the group of run time and above-mentioned chiller
At least one of group of flow group is calculating the loss of above-mentioned chiller.
5. rote learning device according to any one of claim 1 to 4, it is characterised in that
Above-mentioned study portion is configured to learn above-mentioned service condition according to the training dataset obtained for multiple chillers.
6. rote learning device according to any one of claim 1 to 5, it is characterised in that
Above-mentioned study portion possesses：
Return calculating part, it calculates according to above-mentioned judgement data and returns；And
Function update section, it is used for according to above-mentioned return, updates from current state variable and speculates and reduce above-mentioned motor, above-mentioned
Total, the appropriate above-mentioned chiller of the loss of at least one of control device of electric motor and above-mentioned chiller
At least one of the refrigerant temperature of the group and above-mentioned chiller of rotating speed and run time and the group of refrigerant flow group
Function.
7. rote learning device according to claim 6, it is characterised in that
Above-mentioned study portion according to the state variable of at least one of above-mentioned motor and above-mentioned control device of electric motor and on
Return is stated, is updated and the rotating speed of above-mentioned chiller and the refrigerant temperature and system of the group of run time and above-mentioned chiller
At least one of group of cryogen flow organizes corresponding behavior memory table.
8. rote learning device according to claim 6, it is characterised in that
Above-mentioned study portion according to above-mentioned motor or mutually isostructural other motor or electronic of above-mentioned control device of electric motor
The state variable of machine control device and above-mentioned return, update and are used to cool down the cold of other motor or control device of electric motor
But in the group of the refrigerant temperature and refrigerant flow of the rotating speed of device and the group of run time and above-mentioned chiller extremely
The corresponding behavior memory table of a few group.
9. the rote learning device according to claim 7 or 8, it is characterised in that
Above-mentioned return calculating part is according to the copper loss and iron loss of above-mentioned motor, the damage of the power component of above-mentioned control device of electric motor
Total, or above-mentioned motor and the above-mentioned control device of electric motor of the loss of at least one of consumption and above-mentioned chiller
The temperature of ad-hoc location calculates return relative to the surplus of feasible value.
10. a kind of control device of electric motor, possesses the rote learning device any one of claim 1 to 9, the motor
Control device is characterised by,
It is also equipped with：Intention determination section, its result according to obtained by above-mentioned study portion is learnt according to above-mentioned training dataset, certainly
The rotating speed and the group of run time of fixed above-mentioned chiller and the refrigerant temperature of above-mentioned chiller and refrigerant flow
The command value of at least one of group group.
11. control device of electric motor according to claim 10, it is characterised in that
Above-mentioned study portion is configured to according to the additional instruction being made up of the combination of current state variable and above-mentioned judgement data
Practice data set, relearn above-mentioned service condition and update above-mentioned service condition.
12. control device of electric motor according to claim 10 or 11, it is characterised in that
Above-mentioned rote learning device via network connection to above-mentioned control device of electric motor,
Above-mentioned state observation portion is configured to the above-mentioned networks of Jing and obtains current state variable.
13. control device of electric motor according to claim 12, it is characterised in that
Above-mentioned rote learning device is present in Cloud Server.
14. control device of electric motor according to any one of claim 10 to 12, it is characterised in that above-mentioned rote learning
Device is built in the above-mentioned control device of electric motor for controlling above-mentioned motor.
15. a kind of motor control systems, it is characterised in that possess：
Control device of electric motor any one of claim 10 to 14；
Chiller, it is used to cool down above-mentioned motor or above-mentioned control device of electric motor；And
Detector unit, its output said temperature data.
16. motor control systems according to claim 15, it is characterised in that
Above-mentioned chiller is also equipped with：Chiller control unit, the refrigerant flow and cold-producing medium temperature of its control chiller
Degree,
The refrigerant flow and refrigerant temperature of above-mentioned chiller is observed in above-mentioned state observation portion.
A kind of 17. learning by rote, it learns the operation for cooling down the chiller of motor or control device of electric motor
Condition, the learning by rote is characterised by, comprises the steps：
Observation is respective comprising above-mentioned motor and above-mentioned control device of electric motor in the course of action of above-mentioned chiller
At least one of the temperature data of ad-hoc location state variable；
Obtain respective loss for judging above-mentioned motor, above-mentioned control device of electric motor and above-mentioned chiller and each
From ad-hoc location temperature relative to the surplus of feasible value judgement data；And
According to the training dataset being made up of the combination of above-mentioned state variable and above-mentioned judgement data, learn above-mentioned chiller
Service condition.
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2015152415 | 2015-07-31 | ||
| JP2015-152415 | 2015-07-31 | ||
| JP2016129081A JP6093076B2 (en) | 2015-07-31 | 2016-06-29 | Machine learning device for learning operating condition of cooling device, motor control device, machine control system and machine learning method provided with machine learning device |
| JP2016-129081 | 2016-06-29 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106681389A true CN106681389A (en) | 2017-05-17 |
| CN106681389B CN106681389B (en) | 2018-03-20 |
Family
ID=57986262
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610602130.1A Active CN106681389B (en) | 2015-07-31 | 2016-07-27 | Rote learning device and method, control device of electric motor and motor control system |
Country Status (2)
| Country | Link |
|---|---|
| JP (1) | JP6093076B2 (en) |
| CN (1) | CN106681389B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109816624A (en) * | 2017-11-17 | 2019-05-28 | 发那科株式会社 | Appearance inspection device |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP7452511B2 (en) * | 2021-10-01 | 2024-03-19 | 株式会社豊田中央研究所 | motor control device |
| JP7671261B2 (en) * | 2022-02-24 | 2025-05-01 | 株式会社豊田中央研究所 | Temperature control device for rotating electrical machines |
| JP2024107543A (en) * | 2023-01-30 | 2024-08-09 | 株式会社豊田中央研究所 | Fluid prediction device and rotating electrical machine system using same |
| US12500543B2 (en) | 2023-06-13 | 2025-12-16 | Ge Infrastructure Technology, Llc | Active temperature control for generator in power train |
| WO2025203612A1 (en) * | 2024-03-29 | 2025-10-02 | ソフトバンク株式会社 | Information processing device, information processing method, and information processing program |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2000350413A (en) * | 1999-06-01 | 2000-12-15 | Mitsubishi Electric Corp | Cooling gas temperature control device for rotating electric machine |
| JP2007100967A (en) * | 2005-09-30 | 2007-04-19 | Daikin Ind Ltd | Refrigeration equipment |
| CN103430433A (en) * | 2011-03-29 | 2013-12-04 | 希尔莱特有限责任公司 | Method and apparatus for operating a motor with optimized efficiency |
| JP2014236622A (en) * | 2013-06-04 | 2014-12-15 | 三菱電機エンジニアリング株式会社 | Power supply controller and power supply control method |
-
2016
- 2016-06-29 JP JP2016129081A patent/JP6093076B2/en active Active
- 2016-07-27 CN CN201610602130.1A patent/CN106681389B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2000350413A (en) * | 1999-06-01 | 2000-12-15 | Mitsubishi Electric Corp | Cooling gas temperature control device for rotating electric machine |
| JP2007100967A (en) * | 2005-09-30 | 2007-04-19 | Daikin Ind Ltd | Refrigeration equipment |
| CN103430433A (en) * | 2011-03-29 | 2013-12-04 | 希尔莱特有限责任公司 | Method and apparatus for operating a motor with optimized efficiency |
| JP2014236622A (en) * | 2013-06-04 | 2014-12-15 | 三菱電機エンジニアリング株式会社 | Power supply controller and power supply control method |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109816624A (en) * | 2017-11-17 | 2019-05-28 | 发那科株式会社 | Appearance inspection device |
| CN109816624B (en) * | 2017-11-17 | 2021-07-09 | 发那科株式会社 | Appearance inspection device |
Also Published As
| Publication number | Publication date |
|---|---|
| JP6093076B2 (en) | 2017-03-08 |
| JP2017033546A (en) | 2017-02-09 |
| CN106681389B (en) | 2018-03-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106681389A (en) | Machine learning apparatus and method, motor control apparatus and motor control system | |
| CN103967871B (en) | For the method and apparatus controlling to be used for the electro-motor for fluid pump energy supply | |
| EP2850724B1 (en) | Electrical device with emergency cooling system | |
| US9973135B2 (en) | Rotary electric machine control system and control method for rotary electric machine | |
| CN105144014B (en) | The method that desired value for the air adjustment to IT environment is adjusted | |
| WO2016130152A1 (en) | Coolant distribution unit | |
| CN107487176A (en) | Electromotor cooling system, control method, device and electric automobile | |
| EP3769014B1 (en) | Machine-learning method for conditioning individual or shared areas | |
| JP2007060848A (en) | Apparatus and method for controlling electric energy and its program | |
| JP6841747B2 (en) | Rotating machine cooling device and rotating machine cooling method | |
| CN106802060A (en) | Refrigeration for refrigerator control method and control device and refrigerator | |
| US9819300B2 (en) | Machine learning apparatus for learning operation conditions of cooling device, motor control apparatus and motor control system having the machine learning apparatus, and machine learning method | |
| EP3596411B1 (en) | Fault detecting system and method for refrigeration system and refrigeration system | |
| JP2013223284A (en) | Preventive maintenance device for motor | |
| JP6295622B2 (en) | Equipment and cooling method | |
| WO2021239681A1 (en) | Controlling of a district thermal energy distribution system | |
| JP2015117918A (en) | Thermal storage system and control method thereof | |
| CN109163429A (en) | Multi-line system and its mode control method, device and storage medium | |
| EP3380929B1 (en) | Method and system for controlling cooling system of power equipment | |
| CN119717945B (en) | Temperature control method and device for hollow cooling screw | |
| JP2009290109A (en) | Control method for cooling transformer and equipment | |
| JP2014056322A (en) | Electric motor preventive maintenance device, electric motor preventive maintenance method, and electric motor preventive maintenance system | |
| JP5916436B2 (en) | Estimation device, degradation determination device, estimation method, and degradation determination method | |
| JP4857051B2 (en) | Refrigerator equipment operation method and equipment comprising a refrigerator | |
| JP5955206B2 (en) | Demand control apparatus and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |