CN111795488B - Intelligent temperature regulation and control system and method for distributed machine room - Google Patents
Intelligent temperature regulation and control system and method for distributed machine room Download PDFInfo
- Publication number
- CN111795488B CN111795488B CN202010662510.0A CN202010662510A CN111795488B CN 111795488 B CN111795488 B CN 111795488B CN 202010662510 A CN202010662510 A CN 202010662510A CN 111795488 B CN111795488 B CN 111795488B
- Authority
- CN
- China
- Prior art keywords
- machine room
- temperature
- module
- strategy
- temperature control
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
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
- F24F11/89—Arrangement or mounting of control or safety devices
-
- 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
- F24F11/46—Improving electric energy efficiency or saving
-
- 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
- F24F11/88—Electrical aspects, e.g. circuits
-
- H—ELECTRICITY
- H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
- H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
- H05K7/00—Constructional details common to different types of electric apparatus
- H05K7/20—Modifications to facilitate cooling, ventilating, or heating
- H05K7/20536—Modifications to facilitate cooling, ventilating, or heating for racks or cabinets of standardised dimensions, e.g. electronic racks for aircraft or telecommunication equipment
- H05K7/20554—Forced ventilation of a gaseous coolant
- H05K7/2059—Forced ventilation of a gaseous coolant within rooms for removing heat from cabinets, e.g. by air conditioning device
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
- F24F2110/00—Control inputs relating to air properties
- F24F2110/10—Temperature
Landscapes
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Signal Processing (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Aviation & Aerospace Engineering (AREA)
- Thermal Sciences (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Air Conditioning Control Device (AREA)
Abstract
The invention discloses an intelligent control system and method for temperature of a distributed machine room, relates to a system and method for controlling the temperature of the machine room by adjusting a cooling strategy, and aims to solve the problem of energy waste caused by rough temperature control management of the existing machine room, wherein the intelligent control system comprises an analysis device and at least one machine room control device; each machine room control device comprises an environment detection module, a temperature regulation module, a sub-strategy analysis module and a temperature control temporary instruction module; the environment detection module is simultaneously connected with the sub-strategy analysis module and the temperature control temporary instruction module; the sub-strategy analysis module is simultaneously connected with the temperature regulation module and the analysis device; the temperature control temporary instruction module is simultaneously connected with the temperature adjusting module and the sub-strategy analysis module; and the temperature adjusting module is connected with the machine room temperature control equipment.
Description
Technical Field
The invention relates to a temperature regulation and control system and method, in particular to a system and method for regulating and controlling the temperature of a machine room by adjusting a cooling strategy.
Background
Many rooms are limited to temperature management methods in which the temperature of the air conditioner is directly set or the flow rate of the pump is fixed. However, the machine room does not always need a set cooling measure. And the power consumption of the data centers all over the world already accounts for about 1 percent of the power consumption scale all over the world. A considerable amount of this is wasted due to the coarse granularity of the temperature control management of the machine room.
Disclosure of Invention
The invention aims to solve the problem of energy waste caused by rough temperature control management of the existing machine room, and provides an intelligent temperature control system and method for a distributed machine room.
The invention discloses an intelligent temperature control system for a distributed machine room, which comprises an analysis device and at least one machine room control device;
each machine room control device comprises an environment detection module, a temperature regulation module, a sub-strategy analysis module and a temperature control temporary instruction module;
the environment detection module is simultaneously connected with the sub-strategy analysis module and the temperature control temporary instruction module and is used for detecting and recording the machine room environment information, generating an environment information log, simultaneously sending the machine room environment information and the environment information log to the sub-strategy analysis module and sending the machine room environment information to the temperature control temporary instruction module;
the sub-strategy analysis module is simultaneously connected with the temperature regulation module and the analysis device and used for receiving the environmental information and the environmental information log of the machine room, predicting the load peak of the machine room according to the environmental information of the machine room, generating a temperature pre-regulation and control instruction according to the current temperature control strategy and the load peak of the machine room and sending the temperature pre-regulation and control instruction to the temperature regulation module; when a strategy optimization instruction sent by the temperature control temporary instruction module is received, the strategy optimization instruction and the environmental information log are sent to an analysis device; receiving the optimized temperature control strategy returned by the analysis device to replace the current temperature control strategy;
the temperature control temporary instruction module is simultaneously connected with the temperature adjusting module and the sub-strategy analysis module and is used for receiving the machine room environment information, generating a temperature temporary adjusting instruction to be sent to the temperature adjusting module when the machine room temperature in the machine room environment information is greater than or equal to a temperature threshold value, recording the times that the machine room temperature is greater than or equal to the temperature threshold value within a period of time, sending a strategy optimizing instruction to the sub-strategy analysis module when the times that the machine room temperature is greater than or equal to the temperature threshold value is greater than or equal to a set time, and returning the times to zero after the temperature control strategy is updated; or when detecting that the PUE or the energy consumption of the machine room in the environmental information log in the same period of time is more than or equal to the PUE threshold or the energy consumption threshold, sending a strategy optimization instruction to the sub-strategy analysis module; the machine room environment information comprises machine room temperature, machine room PUE and machine room energy consumption;
the temperature adjusting module is connected with the machine room temperature control equipment and used for receiving a temperature pre-adjusting instruction and controlling the machine room temperature control equipment to adjust the temperature of the machine room to the required temperature before the load peak of the machine room comes according to the temperature pre-adjusting instruction; or receiving a temporary temperature regulation and control instruction between load peaks of the two machine rooms, and regulating the temperature of the machine room to the required temperature by controlling the temperature control equipment of the machine room according to the temporary temperature regulation and control instruction;
the analysis device is used for receiving the environmental information log and the strategy optimization instruction; and the environmental information log is used as the input of the machine learning model to obtain an optimized temperature control strategy, and the optimized temperature control strategy is stored and returned to the sub-strategy analysis module respectively to replace the current temperature control strategy.
The invention discloses an intelligent temperature control method for a distributed machine room, which comprises the following steps:
s1, detecting and recording the environmental information of the machine room and generating an environmental information log;
s2, predicting the load peak of the machine room according to the environmental information of the machine room, generating a temperature pre-regulating instruction according to the current temperature control strategy and the load peak of the machine room of each machine room, and controlling the temperature control equipment of the machine room to regulate the temperature of the machine room to the required temperature before the load peak of the machine room comes according to the temperature pre-regulating instruction;
s3, detecting the temperature of the machine room in a first set period between the load peak of the previous machine room and the load peak of the next machine room, and judging whether the temperature of the machine room is greater than or equal to a temperature threshold value;
if the temperature of the machine room is less than the temperature threshold, executing step S5;
if the temperature of the machine room is greater than or equal to the temperature threshold, executing step S4;
s4, generating a temporary temperature adjusting instruction, controlling the machine room temperature control equipment to adjust the temperature of the machine room to the required temperature according to the temporary temperature adjusting instruction, and executing the step S5;
s5, detecting whether the PUE and/or energy consumption of the machine room exceeds the PUE threshold and/or energy consumption threshold of the machine room, if so, executing a step S7, otherwise, executing a step S6;
s6, accumulating the times that the temperature of the machine room is more than or equal to the temperature threshold value in the second set period, judging whether the times is more than or equal to the set times, executing the step S7, and otherwise, turning to the step S3; the second setting period comprises a plurality of first setting periods;
and S7, sending a strategy optimization instruction, obtaining an optimized temperature control strategy according to the environmental information log, replacing the current temperature control strategy with the optimized temperature control strategy, and returning to the step S1.
The invention has the beneficial effects that: the invention can manage the temperature regulation and control of a plurality of machine rooms, optimize the temperature control strategy through machine learning when the temperature control is unfavorable, and then utilize the optimized temperature control strategy to manage the temperature regulation and control of the machine rooms, can carry out the temperature regulation and control according to the actual equipment running condition, and effectively reduces the energy consumption of the machine rooms.
Drawings
FIG. 1 is a schematic diagram of a module structure of an intelligent temperature control system for a distributed machine room, in which a sub-strategy analysis module and a learning balance combiner module are connected according to the present invention;
FIG. 2 is a schematic diagram of a module structure of the intelligent temperature control system for a distributed machine room, in which a sub-strategy analysis module is connected with a learning balance combiner module and a knowledge base module at the same time;
fig. 3 is a schematic diagram of a work flow of a machine room control end of the intelligent temperature control method for a distributed machine room according to the present invention;
fig. 4 is a schematic diagram of an analysis end workflow of the intelligent temperature control method for a distributed machine room according to the present invention.
Detailed Description
In a first specific embodiment, the system for intelligently regulating and controlling the temperature of a distributed machine room in this embodiment includes an analysis device 2 and at least one machine room control device 1;
each machine room control device 1 comprises an environment detection module 1-1, a temperature regulation module 1-2, a sub-strategy analysis module 1-3 and a temperature control temporary instruction module 1-4;
the environment detection module 1-1 is connected with the sub-strategy analysis module 1-3 and the temperature control temporary instruction module 1-4 simultaneously and is used for detecting and recording the machine room environment information, generating an environment information log, sending the machine room environment information and the environment information log to the sub-strategy analysis module 1-3 simultaneously and sending the machine room environment information to the temperature control temporary instruction module 1-4;
the sub-strategy analysis module 1-3 is connected with the temperature regulation module 1-2 and the analysis device 2, and is used for receiving the environmental information and the environmental information log of the machine room, predicting the load peak of the machine room according to the environmental information of the machine room, generating a temperature pre-regulation and control instruction according to the current temperature control strategy and the load peak of the machine room, and sending the temperature pre-regulation and control instruction to the temperature regulation module 1-2; when a strategy optimization instruction sent by the temperature control temporary instruction module 1-4 is received, the strategy optimization instruction and the environmental information log are sent to the analysis device 2; receiving the optimized temperature control strategy returned by the analysis device 2 to replace the current temperature control strategy;
the temperature control temporary instruction module 1-4 is connected with the temperature adjusting module 1-2 and the sub-strategy analysis module 1-3, and is used for receiving the machine room environment information, generating a temperature temporary adjusting instruction when the machine room temperature in the machine room environment information is greater than or equal to a temperature threshold value, sending the temperature temporary adjusting instruction to the temperature adjusting module 1-2, recording the times that the machine room temperature is greater than or equal to the temperature threshold value within a period of time, sending a strategy optimizing instruction to the sub-strategy analysis module 1-3 when the times that the machine room temperature is greater than or equal to the temperature threshold value is greater than or equal to a set time, and returning the times to zero after the temperature control strategy is updated; or when detecting that the PUE or the energy consumption of the machine room in the environmental information log in the same period of time is more than or equal to the PUE threshold or the energy consumption threshold, sending a strategy optimization instruction to the sub-strategy analysis module 1-3; the machine room environment information comprises machine room temperature, machine room PUE and machine room energy consumption;
the temperature adjusting module 1-2 is connected with the machine room temperature control equipment and used for receiving a temperature pre-adjusting instruction and controlling the machine room temperature control equipment to adjust the temperature of the machine room to a required temperature before a load peak of the machine room comes according to the temperature pre-adjusting instruction; or receiving a temporary temperature regulation and control instruction between load peaks of the two machine rooms, and regulating the temperature of the machine room to the required temperature by controlling the temperature control equipment of the machine room according to the temporary temperature regulation and control instruction;
the analysis device 2 is used for receiving the environmental information log and the strategy optimization instruction; and the environmental information log is used as the input of the machine learning model to obtain an optimized temperature control strategy, and the optimized temperature control strategy is stored and returned to the sub-strategy analysis module 1-3 respectively to replace the current temperature control strategy.
Specifically, as shown in fig. 1, the intelligent regulation and control system of the present invention includes: the system comprises a machine room control device 1 and an analysis device 2, wherein the machine room control device 1 is further detailed into an environment detection module 1-1, a temperature regulation module 1-2, a sub-strategy analysis module 1-3 and a temperature control temporary instruction module 1-4;
the environment monitoring module 1-1 is configured to uniformly manage all machine room environment information of the machine room, such as temperature, humidity and monitoring devices (for monitoring energy consumption, load, and the like), and share the machine room environment information in real time (or at a certain period) to the sub-policy analysis module 1-3 and the temperature control temporary instruction module 1-4. And the environmental information of the machine room between the last temperature control strategy update and the current temperature control strategy update is stored and processed to generate an environmental information log.
The environment monitoring module 1-1 may specifically include an infrared temperature measuring device, a humidity detecting device, and a network connection device to obtain local environmental information such as air temperature and air speed (for recording local climate environment differences and adjusting the temperature of the machine room according to the climate differences). The devices are arranged beside the servers in the machine room and are arranged according to the division of the machine room server groups. The devices are directly connected with a processor of the environment monitoring module 1-1 through a data line or through a wireless network signal connection, and are used for transmitting monitoring data in real time.
The temperature adjusting module 1-2 is used for controlling all temperature control equipment of the machine room in real time, cooling the equipment before a load peak comes, and controlling the temperature based on the instructions sent by the sub-strategy analysis module 1-3 and the temperature control temporary instruction module 1-4.
The temperature adjusting module 1-2 specifically comprises a cooler (an air conditioner and the like), a cooling water pump is uniformly installed indoors according to requirements, and is directly connected with a processor of the temperature adjusting module 1-2 through a data line or a wireless network. Every computer lab temperature control equipment all has 3 ~ 8 control by temperature change gears for realize the temperature control (cooling) intensity of different degrees.
The temperature adjusting module 1-2 usually performs a temperature reduction operation in advance on a predicted load peak according to a current temperature control strategy and machine room environment information, and if the temperature of a few servers in the machine room exceeds a temperature threshold value for several times within a certain period of time, the temperature adjusting module 1-2 may perform a temporary temperature reduction by temporarily increasing the operation of a cooling water pump or reducing the temperature of a cooler. However, if the cooling power consumption temporarily increases due to an increase in the operation of the cooling water pump or a decrease in the temperature of the cooling machine, and the PUE (total power consumption/energy consumption for IT) in the machine room fluctuates several times (exceeds the PUE threshold), the environmental information log is transmitted to the analysis device 2 for analysis.
And the sub-strategy analysis module 1-3 is a module for specifically bearing a load peak prediction task of the machine room. And can be primarily analyzed and adjusted when the temperature of the machine room is not adjusted conveniently. If the optimum level is not reached, it reports the existing strategy and allows the analysis device 2 to perform an in-depth analysis process.
The sub-policy analysis module 1-3 may specifically include a computing device with a relatively high performance, and the computing device is connected to the main processor of the machine room control apparatus 1 through a data line or a wireless network.
And the temperature control temporary instruction module 1-4 is used for sending a temperature control temporary instruction to the temperature adjusting module 1-2 or sending a strategy optimizing instruction to the sub-strategy analysis module 1-3 when the conditions are met.
The working process of the intelligent regulation and control system is that the machine room control device 1 monitors and records the environmental information of the machine room in real time through the environmental monitoring module 1-1; when the environmental temperature reaches a temperature threshold, the energy consumption of the machine room or the PUE of the machine room is abnormal (exceeds the threshold or other set values, such as the maximum value of the energy consumption of the machine room or the PUE of the machine room when the previous temperature control strategy works), or after strategy information is updated, the sub-strategy analysis module is called (a strategy optimization instruction is sent or the temperature control strategy is updated and other modules are called to work under the current temperature control strategy).
The analysis device 2 may be a server with strong computing capability, large memory space and good network bandwidth, software for analyzing the temperature control strategy is deployed on the server, and after receiving the environmental information log and the strategy optimization instruction uploaded by the machine room control device 1 and other relevant information of the machine room, the current temperature control strategy is analyzed by using a machine learning algorithm to obtain an optimized temperature control strategy, so that the machine room control device 1 performs temperature regulation and control based on the optimized temperature control strategy again.
The temperature control strategy is added manually (at the beginning or at any other stage) by an operator or automatically by a system, if the PUE of the machine room exceeds the past (exceeds the PUE threshold value or the maximum value of the PUE of the machine room when the previous temperature control strategy works), the energy consumption of the machine room is larger than that before the unmodified temperature control strategy (exceeds the energy consumption threshold value or the maximum value of the energy consumption of the machine room when the previous temperature control strategy works) or the temperature of the machine room exceeds the threshold value for a plurality of times in a certain time period, the temperature adjusting module 1-2 starts to call the sub-strategy analyzing module 1-3, the sub-strategy analyzing module 1-3 starts to enter an analyzing mode, and the specific action is to request the environment detecting module 1-1 for the environment information of the machine room around a server more frequently and perform temporary temperature adjustment on a preset machine room area or the whole machine room.
If after the temporary adjustment, it is still detected for many times or once that the PUE of the machine room exceeds the threshold value or the average energy consumption of the machine room is higher than past within a period of time, the machine room control device 1 sends the environmental information log and other information obtained from the environmental monitoring module 1-1 to the analysis device 2. And after receiving the data, the analysis device 2 optimizes or corrects the current temperature control strategy, returns the optimized temperature control strategy to the machine room control device 1, and repeats the process until the energy consumption of the machine room returns to the point before the temperature control strategy is modified or is better than the prior art.
The temperature control strategy is a control protocol for managing the temperature of the machine room, and the temperature control strategy can be composed of a plurality of instructions, wherein the instructions can be composed of 4 tuples:
for example, < 'location _ id', 'average _ temperature', 'average _ load', 'action' >, the element location _ id represents an identifier that uniquely identifies a certain machine room, the average _ temperature represents the average temperature of the machine room reached, the average _ load represents the average load of the machine room, and the action represents the operation performed on the temperature adjustment module 1-2 when the average temperature of the machine room and the average load of the machine room reach corresponding values. The temperature control strategy is used for predicting the load peak of the machine room which is possibly reached in the future in advance and cooling in time.
And the information sent to the analysis means 2 may be constituted by the following elements:
for example, machine information in the environmental information log: < 'device _ position', 'log _ time', 'load', 'temperature', where 'device _ position' represents the location of the machine room where the machine is located, 'log _ time' represents the time of information recording, 'load' represents the load, 'temperature' represents the temperature of the machine at the time of recording.
The machine refers to energy consumption equipment such as a server and machine room temperature control equipment which work in a machine room.
And other information, including the machine room information: "location _ id '," local _ weather', "local _ hub '," server _ intensity', "temperature _ control _ intensity '," city', "availability _ info '," location _ id to uniquely identify a room, "server _ intensity' to represent room server density," "temperature _ control _ intensity 'to represent average density of room temperature control devices (number of room temperature control devices/number of servers)," city' to represent city of the room, "" availability 'to represent altitude information, "" wind _ info' to represent local weather information such as local wind speed, etc.
Further, the environment detection module 1-1 comprises a server temperature acquisition sub-module, a server load and energy consumption monitoring sub-module, a machine room area temperature monitoring sub-module, a machine room area load and energy consumption information monitoring sub-module, a machine room temperature control equipment energy consumption monitoring module, a machine room PUE monitoring module and an environment information log generation module;
the server temperature acquisition submodule is used for acquiring the temperature information of a single server in real time;
the server load and energy consumption monitoring submodule is used for acquiring load and energy consumption information of a single server in real time;
the machine room region temperature information monitoring submodule is used for receiving the temperature information of a single server in the machine room region according to the pre-divided machine room region and calculating to obtain the average temperature information in the machine room region;
the machine room region load and energy consumption information monitoring submodule is used for receiving load information of a single server in a machine room region according to the machine room region which is divided in advance and calculating to obtain average energy consumption information in the machine room region;
the energy consumption monitoring module of the machine room temperature control equipment is used for acquiring energy consumption information of the machine room temperature control equipment in real time;
the machine room PUE monitoring module is used for calculating according to the energy consumption information of the single server and the energy consumption information of the machine room temperature control equipment to obtain a machine room PUE;
the environmental information log generating module is used for collecting the temperature information of the single server, the average temperature information in the machine room area, the load information of the single server, the energy consumption information of the single server, the average energy consumption information in the machine room area, the energy consumption information of the machine room temperature control equipment and the machine room PUE to generate logs.
Specifically, the temperature of the machine room in the first embodiment may be subdivided into temperature information of a single server, average temperature information in the machine room area, or average temperature information of the whole machine room according to actual needs (i.e., the machine room area is set as the whole machine room).
The energy consumption of the machine room can be subdivided into energy consumption information of a single server, average energy consumption information in a machine room area, energy consumption information of temperature control equipment of the machine room and average energy consumption (total energy consumption/energy consumption equipment number) of the machine room according to actual needs.
The machine room PUE can be calculated according to a formula, and corresponding values are substituted.
The environmental information of the machine room is transmitted in real time or in a certain period, and the environmental information log can be regarded as a data table obtained by storing and processing the environmental information of the machine room for a period of time (between the last time of updating the temperature control strategy and the current time of updating the temperature control strategy).
The server temperature acquisition submodule is used for monitoring the temperature information of a single server;
the server load and energy consumption monitoring submodule is used for detecting the load information of a single server and calculating the energy consumption information;
the machine room region temperature monitoring submodule is used for collecting single-server temperature information in a region and calculating the average temperature of the machine rooms in the region in real time;
the machine room area load and energy consumption monitoring submodule is used for collecting load information of a single server in the area and calculating the average energy consumption of the machine room area;
the energy consumption monitoring module of the machine room temperature control equipment (cooling equipment such as a pump, an air conditioner and the like) is used for collecting energy consumption information of the machine room temperature control equipment;
the environment detection module 1-1 records information generated by each sub-module in real time, and records a generated log to be called by the temperature regulation module 1-2 and the sub-strategy analysis module 1-3.
The servers in each computer room may be divided, for example, into 8 groups of servers. When performing peak prediction, this group of servers is treated as one server. The data for this set of servers is made with the average temperature and average load of these servers.
And the energy consumption of the machine room also comprises the energy consumption of power consumption equipment such as machine room lamplight and an alarm, but the quantity of the devices is generally fixed and the energy consumption is fixed, so that the device only needs to count in advance and is added as a part of total energy consumption when average energy consumption is carried out every time, and the device can be ignored if the numerical value is small.
And the system is also provided with a local meteorological environment monitoring submodule for collecting local real-time information such as temperature, humidity, wind power, wind direction and the like. And storing specific information such as the altitude of the place where the machine room is located.
Further, the analysis device 2 comprises a learning balance combiner module, a strategy optimizer module 2-2 and a knowledge base module 2-1;
the learning balance combiner module comprises at least one learner module 2-3, is connected with the sub-strategy analysis module 1-3, the strategy optimizer module 2-2 and the knowledge base module 2-1, and is used for receiving the environmental information log and the strategy optimization instruction sent by the sub-strategy analysis module 1-3 and then inputting the environmental information log into a machine learning model in the learner module 2-3 to obtain an optimized temperature control strategy; sending the optimized temperature control strategy to a strategy optimizer module 2-2, and sending the environment information log to a knowledge base module 2-1;
the strategy optimizer module 2-2 is simultaneously connected with the sub-strategy analysis module 1-3 and the knowledge base module 2-1 and is used for receiving and processing the optimized temperature control strategy, wherein the processing comprises duplication removal processing and conflict removal processing; respectively sending the optimized temperature control strategy to the sub-strategy analysis module 1-3 to replace the current temperature control strategy, sending the optimized temperature control strategy to the knowledge base module 2-1, updating the knowledge base module 2-1, and marking the optimized temperature control strategy as the current temperature control strategy; the knowledge base module 2-1 is used for receiving all environment information logs sent by the learning balance combiner module and all temperature control strategies sent by the strategy optimizer module 2-2, storing the environment information logs and the temperature control strategies according to the corresponding relation, completing updating and sending an updating signal to the learning balance combiner module;
and the learning balance combiner module is also used for training the machine learning model in the learner module 2-3 by taking all the environmental information logs and all the temperature control strategies stored in the knowledge base module 2-1 as a training set according to the updating signal to obtain an optimized machine learning model.
Specifically, as shown in fig. 1, when the environmental information log and the policy optimization command sent by the sub-policy analysis module 1-3 are sent to the learning balance combiner module, the environmental information log is transmitted to different learner modules 2-3 to obtain at least one optimized temperature control policy, the newly learned optimized temperature control policies are combined, the combined optimized temperature control policies are transmitted to the policy optimizer module 2-2 to perform operations such as deduplication and conflict elimination, and finally the optimized temperature control policy is returned to the knowledge base module 2-1 to update the knowledge base module 2-1, and the optimized temperature control policy is returned to the machine room control device 1 to replace the current temperature control policy.
The knowledge base module 2-1 fully utilizes the difference data generated by each machine room to integrate. And the knowledge base module 2-1 is dynamically updated in real time because the strategy is continuously optimized. The learner modules 2-3 adopt different machine learning methods and apply machine learning algorithms to optimize strategies which possibly have problems, and energy wasted by a machine room is reduced as much as possible. After the temperature control strategy is updated, the learning balance combiner module takes out the environmental information logs collected from all the machine rooms and the corresponding temperature control strategies from the knowledge base module 2-1, and trains the machine learning model in each learner module 2-3 by calling the learner module 2-3.
The training process of the middle machine learning model of the learner module 2-3 is as follows: when the analyzer 2 receives the analysis request from the room control apparatus 1, the analyzer 2 first normalizes the environment log information into a pattern (pattern) having a format similar to a rule in a policy but lacking an action field. And then, the analysis end starts to call different learner modules 2-3, and inputs the converted environmental information log and the temperature control strategy into the learner modules.
The training process for the three classes of learner modules 2-3 is briefly set forth below:
1. neighbor learning device
The policies in the schema collection and the knowledge base are mapped into the same high-dimensional space. For each mode, several (set threshold) instructions similar to it are fetched (each rule in the policy is called an instruction). These similar vectors are voted and the final mode is marked as the most numerous actions.
2. Pattern rule distance learning device
And (4) establishing a tree for the strategies in the knowledge base, wherein each layer corresponds to one field in the strategy. And for each mode, comparing layer by layer according to the fields. When a field is matched and the matching depth exceeds a threshold value, the field cannot be matched. And counting decision fields of all subtrees under the field. Labeling is done in most cases.
3. Commonly-appearing learning device
Within a certain time period, certain two patterns occur in succession, referred to as co-occurrence.
The co-occurrence probability is represented by cij, and the calculation formula is as follows: number of co-occurrences of Pattern i and Pattern j/Total number of occurrences of Pattern i
The probability of the common occurrence of all n different access patterns occurring in the log can be represented by an n x n matrix
After the learner computes the matrix, the learner ranks the rows that first represent the new pattern from large to small. Thereafter, truncation is performed according to a threshold and voting is performed according to patterns known to occur in the strategy. The final majority of actions are the final actions of the current mode.
The strategy optimizer module 2-2 receives the updated strategy of the combination sent by the learning balance combiner module, optimizes the strategy (including the operation of removing the duplicate and conflict), and finally updates the knowledge base module 2-1 and returns the updated knowledge base module to the machine room control device 1.
The temperature control method based on the machine room temperature regulation and analysis system utilizes a machine learning technology to improve the use efficiency of machine room energy.
Further, each of the learning balance combiner modules 2-3 is implemented using a different machine learning model.
In a second specific embodiment, as shown in fig. 3, the method for intelligently regulating and controlling the temperature of the distributed machine room in the second embodiment includes:
s1, detecting and recording the environmental information of the machine room and generating an environmental information log;
s2, predicting the load peak of the machine room according to the environmental information of the machine room, generating a temperature pre-regulating instruction according to the current temperature control strategy and the load peak of the machine room of each machine room, and controlling the temperature control equipment of the machine room to regulate the temperature of the machine room to the required temperature before the load peak of the machine room comes according to the temperature pre-regulating instruction;
s3, detecting the temperature of the machine room in a first set period between the load peak of the previous machine room and the load peak of the next machine room, and judging whether the temperature of the machine room is greater than or equal to a temperature threshold value;
if the temperature of the machine room is less than the temperature threshold, executing step S5;
if the temperature of the machine room is greater than or equal to the temperature threshold, executing step S4;
s4, generating a temporary temperature adjusting instruction, controlling the machine room temperature control equipment to adjust the temperature of the machine room to the required temperature according to the temporary temperature adjusting instruction, and executing the step S5;
s5, detecting whether the PUE and/or energy consumption of the machine room exceeds the PUE threshold and/or energy consumption threshold of the machine room, if so, executing a step S7, otherwise, executing a step S6;
s6, accumulating the times that the temperature of the machine room is more than or equal to the temperature threshold value in the second set period, judging whether the times is more than or equal to the set times, executing the step S7, and otherwise, turning to the step S3; the second setting period comprises a plurality of first setting periods;
and S7, sending a strategy optimization instruction, obtaining an optimized temperature control strategy according to the environmental information log, replacing the current temperature control strategy with the optimized temperature control strategy, and returning to the step S1.
Specifically, the method of the present invention generally implements three functions:
1. when the temperature of the machine room exceeds a machine room threshold value, carrying out temporary temperature regulation and control;
2. predicting the load peak of the machine room according to the current temperature control strategy, and performing temperature control regulation (cooling) in advance;
3. and regularly checking indexes such as average energy consumption of the machine room PUE and the machine room, the temporary temperature regulation and control frequency and the like, and optimizing a temperature control strategy when the condition is met, wherein the aim is to reduce the total energy consumption.
The overall procedure can be divided into two major timing tasks:
the first timing task is short in time interval (the first set period), namely temperature information is monitored in real time, temporary temperature regulation and control are immediately carried out when the temperature of the machine room is high, and the number of the temporary temperature regulation and control can be recorded.
The second timing task, which is a longer time interval (second set period), includes several first timing tasks. If the times that the temperature of the machine room exceeds the temperature threshold is greater than the set times, the temperature is uploaded to the analysis device 2 for analysis, namely, the function 2 is executed without force;
the second timing task also includes that when other indexes, such as the PUE of the machine room exceeds the PUE threshold of the machine room, are also uploaded to the analysis device 2 for analysis, that is, the function 3 is executed without force.
Claims (4)
1. The intelligent temperature control system for the distributed machine room is characterized by comprising an analysis device (2) and at least one machine room control device (1);
each machine room control device (1) comprises an environment detection module (1-1), a temperature regulation module (1-2), a sub-strategy analysis module (1-3) and a temperature control temporary instruction module (1-4);
the environment detection module (1-1) is connected with the sub-strategy analysis module (1-3) and the temperature control temporary instruction module (1-4) at the same time, and is used for detecting and recording the machine room environment information, generating an environment information log, sending the machine room environment information and the environment information log to the sub-strategy analysis module (1-3) and sending the machine room environment information to the temperature control temporary instruction module (1-4);
the sub-strategy analysis module (1-3) is connected with the temperature regulation module (1-2) and the analysis device (2) at the same time, and is used for receiving the machine room environment information and the environment information log, predicting a machine room load peak according to the machine room environment information, generating a temperature pre-regulation instruction according to a current temperature control strategy and the machine room load peak, and sending the temperature pre-regulation instruction to the temperature regulation module (1-2); when a strategy optimization instruction sent by the temperature control temporary instruction module (1-4) is received, the strategy optimization instruction and the environmental information log are sent to the analysis device (2); receiving the optimized temperature control strategy returned by the analysis device (2) to replace the current temperature control strategy;
the temperature control temporary instruction module (1-4) is connected with the temperature adjusting module (1-2) and the sub-strategy analysis module (1-3) and is used for receiving the machine room environment information, generating a temperature temporary adjusting instruction to be sent to the temperature adjusting module (1-2) when the machine room temperature in the machine room environment information is greater than or equal to a temperature threshold value, recording the times that the machine room temperature is greater than or equal to the temperature threshold value within a period of time, sending a strategy optimizing instruction to the sub-strategy analysis module (1-3) when the times that the machine room temperature is greater than or equal to the temperature threshold value is greater than or equal to a set time, and resetting the times to zero after the temperature control strategy is updated; or when detecting that the PUE of the machine room in the environmental information log in the same period of time is greater than or equal to the PUE threshold or the energy consumption of the machine room is greater than or equal to the energy consumption threshold, sending a strategy optimization instruction to the sub-strategy analysis module (1-3); the machine room environment information comprises machine room temperature, machine room PUE and machine room energy consumption;
the temperature adjusting module (1-2) is connected with the machine room temperature control equipment and used for receiving the temperature pre-adjusting and controlling instruction and controlling the machine room temperature control equipment to adjust the temperature of the machine room to the required temperature before the load peak of the machine room comes according to the temperature pre-adjusting and controlling instruction; or receiving the temporary temperature regulating and controlling instruction between two load peaks of the machine room, and regulating the temperature of the machine room to the required temperature by controlling the temperature control equipment of the machine room according to the temporary temperature regulating and controlling instruction;
the analysis device (2) is used for receiving the environmental information log and the strategy optimization instruction; the environmental information log is used as the input of the machine learning model to obtain an optimized temperature control strategy, and the optimized temperature control strategy is stored and returned to the sub-strategy analysis module (1-3) respectively to replace the current temperature control strategy;
the environment detection module (1-1) comprises a server temperature acquisition sub-module, a server load and energy consumption monitoring sub-module, a machine room area temperature monitoring sub-module, a machine room area load and energy consumption information monitoring sub-module, a machine room temperature control equipment energy consumption monitoring module, a machine room PUE monitoring module and an environment information log generation module;
the server temperature acquisition submodule is used for acquiring the temperature information of a single server in real time;
the server load and energy consumption monitoring submodule is used for acquiring load and energy consumption information of a single server in real time;
the machine room region temperature monitoring submodule is used for receiving the temperature information of a single server in the machine room region according to the pre-divided machine room region and calculating to obtain the average temperature information in the machine room region;
the machine room region load and energy consumption information monitoring submodule is used for receiving load information of a single server in a machine room region according to a pre-divided machine room region and calculating to obtain average energy consumption information in the machine room region;
the energy consumption monitoring module of the machine room temperature control equipment is used for acquiring energy consumption information of the machine room temperature control equipment in real time;
the machine room PUE monitoring module is used for calculating according to the energy consumption information of the single server and the energy consumption information of the machine room temperature control equipment to obtain a machine room PUE;
the environmental information log generating module is used for collecting the temperature information of the single server, the average temperature information in the machine room area, the load information of the single server, the energy consumption information of the single server, the average energy consumption information in the machine room area, the energy consumption information of the machine room temperature control equipment and the machine room PUE to generate logs.
2. The intelligent control system for the temperature of the distributed machine room according to claim 1, wherein the analysis device (2) comprises a learning balance combiner module, a strategy optimizer module (2-2) and a knowledge base module (2-1);
the learning balance combiner module comprises at least one learner module (2-3), is connected with the sub-strategy analysis module (1-3), the strategy optimizer module (2-2) and the knowledge base module (2-1) and is used for receiving the environment information log and the strategy optimization instruction sent by the sub-strategy analysis module (1-3) and then inputting the environment information log into a machine learning model in the learner module (2-3) to obtain an optimized temperature control strategy; the optimized temperature control strategy is sent to a strategy optimizer module (2-2), and the environment information log is sent to a knowledge base module (2-1);
the strategy optimizer module (2-2) is simultaneously connected with the sub-strategy analysis module (1-3) and the knowledge base module (2-1) and is used for receiving and processing the optimized temperature control strategy, wherein the processing comprises duplicate removal processing and conflict removal processing; respectively sending the optimized temperature control strategy to a sub-strategy analysis module (1-3) to replace the current temperature control strategy, sending the optimized temperature control strategy to a knowledge base module (2-1), updating the knowledge base module (2-1), and marking the optimized temperature control strategy as the current temperature control strategy; the knowledge base module (2-1) is used for receiving all environment information logs sent by the learning balance combiner module and all temperature control strategies sent by the strategy optimizer module (2-2), storing the environment information logs and the temperature control strategies according to the corresponding relation, completing updating and sending an updating signal to the learning balance combiner module;
and the learning balance combiner module is also used for training the machine learning model in the learner module (2-3) by taking all the environmental information logs and all the temperature control strategies stored in the knowledge base module (2-1) as a training set according to the updating signal to obtain an optimized machine learning model.
3. The intelligent temperature control system for a distributed machine room according to claim 2, wherein each learner module (2-3) in the learning balance combiner module is implemented by using a different machine learning model.
4. An intelligent control method for temperature of a distributed machine room, which is based on the intelligent control system for temperature of a distributed machine room of claim 3, and comprises:
s1, detecting and recording the environmental information of the machine room and generating an environmental information log;
s2, predicting a load peak of the machine room according to the environmental information of the machine room, generating a temperature pre-regulation and control instruction according to the current temperature control strategy of each machine room and the load peak of the machine room, and controlling a temperature control device of the machine room to regulate the temperature of the machine room to a required temperature before the load peak of the machine room comes according to the temperature pre-regulation and control instruction;
s3, detecting the temperature of the machine room in a first set period between the load peak of the previous machine room and the load peak of the next machine room, and judging whether the temperature of the machine room is greater than or equal to a temperature threshold value;
if the temperature of the machine room is less than the temperature threshold, executing step S5;
if the temperature of the machine room is greater than or equal to the temperature threshold, executing step S4;
s4, generating a temporary temperature adjusting instruction, controlling the machine room temperature control equipment to adjust the temperature of the machine room to the required temperature according to the temporary temperature adjusting instruction, and executing the step S5;
s5, detecting whether the PUE of the machine room exceeds the PUE threshold of the machine room and/or whether the energy consumption of the machine room exceeds the energy consumption threshold of the machine room, if so, executing a step S7, otherwise, executing a step S6;
s6, accumulating the times that the temperature of the machine room is more than or equal to the temperature threshold value in the second set period, judging whether the times is more than or equal to the set times, executing the step S7, and otherwise, turning to the step S3; the second setting period comprises a plurality of first setting periods;
and S7, sending a strategy optimization instruction, obtaining an optimized temperature control strategy according to the environmental information log, replacing the current temperature control strategy with the optimized temperature control strategy, and returning to the step S1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010662510.0A CN111795488B (en) | 2020-07-10 | 2020-07-10 | Intelligent temperature regulation and control system and method for distributed machine room |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010662510.0A CN111795488B (en) | 2020-07-10 | 2020-07-10 | Intelligent temperature regulation and control system and method for distributed machine room |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111795488A CN111795488A (en) | 2020-10-20 |
| CN111795488B true CN111795488B (en) | 2021-11-02 |
Family
ID=72806773
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010662510.0A Active CN111795488B (en) | 2020-07-10 | 2020-07-10 | Intelligent temperature regulation and control system and method for distributed machine room |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111795488B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112229045B (en) * | 2020-10-23 | 2021-12-31 | 深圳市华创智能工程技术有限公司 | Intelligent temperature control method, device and system of matrix switching system |
| CN112987895B (en) * | 2021-05-20 | 2021-08-10 | 凌渡(南京)科技有限公司 | Intelligent server chip refrigeration method, device and system based on cooling liquid |
| CN113726732B (en) * | 2021-07-14 | 2023-05-19 | 华迪计算机集团有限公司 | Monitoring and auxiliary decision making system and method for data machine room |
| CN113934241A (en) * | 2021-10-15 | 2022-01-14 | 浪潮思科网络科技有限公司 | SDN-based machine room temperature adjusting method and device |
| CN114007400B (en) * | 2022-01-05 | 2022-04-01 | 浙江德塔森特数据技术有限公司 | Machine room energy-saving control method of distributed architecture |
| CN114838485B (en) * | 2022-02-28 | 2024-04-26 | 青岛海尔空调器有限总公司 | Method and device for controlling air conditioner, air conditioner and storage medium |
| CN114901057B (en) * | 2022-07-12 | 2022-09-27 | 联通(广东)产业互联网有限公司 | Multi-point energy consumption detection and dynamic regulation system in data center machine room |
| CN116027829A (en) * | 2022-07-18 | 2023-04-28 | 中国电信股份有限公司 | Machine room temperature control method, device, equipment and storage medium |
| CN115686108B (en) * | 2023-01-03 | 2023-04-07 | 浙江德塔森特数据技术有限公司 | Multi-level cloud management and control method and system for edge data machine room |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2014236605A (en) * | 2013-06-03 | 2014-12-15 | 三菱電機株式会社 | Management system of air-conditioner |
| CN105183032A (en) * | 2015-07-21 | 2015-12-23 | 浪潮(北京)电子信息产业有限公司 | Method and system for controlling and managing temperature of computer room in cloud platform |
| CN109556236A (en) * | 2018-10-22 | 2019-04-02 | 珠海格力电器股份有限公司 | energy efficiency control method and device |
| JP6534038B2 (en) * | 2015-07-16 | 2019-06-26 | 清水建設株式会社 | Demand power control apparatus and demand power control method |
| CN110691491A (en) * | 2018-07-05 | 2020-01-14 | 百度（美国）有限责任公司 | Controlling and optimizing indirect evaporative cooling units for data center cooling |
-
2020
- 2020-07-10 CN CN202010662510.0A patent/CN111795488B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2014236605A (en) * | 2013-06-03 | 2014-12-15 | 三菱電機株式会社 | Management system of air-conditioner |
| JP6534038B2 (en) * | 2015-07-16 | 2019-06-26 | 清水建設株式会社 | Demand power control apparatus and demand power control method |
| CN105183032A (en) * | 2015-07-21 | 2015-12-23 | 浪潮(北京)电子信息产业有限公司 | Method and system for controlling and managing temperature of computer room in cloud platform |
| CN110691491A (en) * | 2018-07-05 | 2020-01-14 | 百度（美国）有限责任公司 | Controlling and optimizing indirect evaporative cooling units for data center cooling |
| CN109556236A (en) * | 2018-10-22 | 2019-04-02 | 珠海格力电器股份有限公司 | energy efficiency control method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111795488A (en) | 2020-10-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111795488B (en) | Intelligent temperature regulation and control system and method for distributed machine room | |
| CN109800066B (en) | Energy-saving scheduling method and system for data center | |
| US7356548B1 (en) | System and method for remote monitoring and controlling of facility energy consumption | |
| CN112650580B (en) | Industrial big data monitoring system based on edge calculation | |
| CN116320831B (en) | Intelligent park security system based on edge calculation | |
| CN117171590B (en) | Intelligent driving optimization method and system for motor | |
| CN116070795B (en) | Intelligent energy management and control method and system based on Internet of things | |
| CN111507520A (en) | Dynamic prediction method and system for load of heat exchange unit | |
| CN116365717B (en) | High-voltage circuit safety monitoring system and method | |
| CN116365713A (en) | Photovoltaic power station cluster monitoring method and system | |
| CN111045363A (en) | Intelligent operation and maintenance management and control cloud platform of information communication network | |
| CN117391591B (en) | Intelligent storage optimizing system based on digital twinning | |
| CN117650628B (en) | Energy efficiency management system based on self-adaptive preconditioning scene | |
| CN113625639A (en) | Agricultural intelligent monitoring system and monitoring method thereof | |
| CN111914000B (en) | Server power capping method and system based on power consumption prediction model | |
| CN111290354B (en) | Modularized data center waste heat efficient utilization control management system based on cloud computing | |
| CN117346417B (en) | Water chiller temperature control system and method based on Internet of things technology | |
| CN116320833A (en) | Heat supply pipe network monitoring method based on Internet of things technology | |
| CN118012711B (en) | Internet-based integrated chip running state supervision system and method | |
| CN111488319A (en) | Log association processing method, device and equipment | |
| CN117478017B (en) | Control method of servo motor | |
| CN116452367B (en) | Intelligent power supply management system | |
| CN117096955B (en) | Distributed photovoltaic cluster operation control system | |
| CN103853107A (en) | Data center control system based on power consumption monitoring | |
| CN118536687A (en) | Big data machine room inspection method and system based on using action frequency |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |