WO2017000738A1 - 分布式计算网络系统及用于该系统的计算节点 - Google Patents

分布式计算网络系统及用于该系统的计算节点 Download PDF

Info

Publication number
WO2017000738A1
WO2017000738A1 PCT/CN2016/084332 CN2016084332W WO2017000738A1 WO 2017000738 A1 WO2017000738 A1 WO 2017000738A1 CN 2016084332 W CN2016084332 W CN 2016084332W WO 2017000738 A1 WO2017000738 A1 WO 2017000738A1
Authority
WO
WIPO (PCT)
Prior art keywords
cpn
computing
computing node
algorithm
node
Prior art date
Application number
PCT/CN2016/084332
Other languages
English (en)
French (fr)
Inventor
姜子炎
Original Assignee
邻元科技(北京)有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Priority claimed from CN201510378076.2A external-priority patent/CN106331037B/zh
Priority claimed from CN201510377980.1A external-priority patent/CN106325229B/zh
Application filed by 邻元科技(北京)有限公司 filed Critical 邻元科技(北京)有限公司
Priority to JP2017568390A priority Critical patent/JP6742353B2/ja
Priority to EP16817102.3A priority patent/EP3318938A4/en
Priority to US15/740,146 priority patent/US10732588B2/en
Publication of WO2017000738A1 publication Critical patent/WO2017000738A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05BCONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B15/00Systems controlled by a computer
    • G05B15/02Systems controlled by a computer electric
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/50Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5061Partitioning or combining of resources
    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05BCONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00Programme-control systems
    • G05B19/02Programme-control systems electric
    • G05B19/04Programme control other than numerical control, i.e. in sequence controllers or logic controllers
    • G05B19/05Programmable logic controllers, e.g. simulating logic interconnections of signals according to ladder diagrams or function charts
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/54Interprogram communication
    • G06F9/545Interprogram communication where tasks reside in different layers, e.g. user- and kernel-space
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/12Discovery or management of network topologies
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W84/00Network topologies
    • H04W84/18Self-organising networks, e.g. ad-hoc networks or sensor networks
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/12Discovery or management of network topologies
    • H04L41/122Discovery or management of network topologies of virtualised topologies, e.g. software-defined networks [SDN] or network function virtualisation [NFV]

Definitions

  • the present invention relates to the field of computing network technologies, and in particular, to a distributed computing network system and a computing node for the system.
  • the automatic control system in the prior art adopts a centralized organizational structure, and all terminal measurement and control points (sensors, actuators, field controllers) All are connected through the bus communication network, and the information measurement and control points at the end of each subsystem (lighting system, air conditioning system, fire protection system, security system) are largely distributed in the same building subspace, but according to different sub-categories The system performs vertical and vertical integration.
  • This centralized automatic control system has the following main disadvantages:
  • the terminal's measurement and control points need to be named for global communication, define physical attributes and define the relationship between each other.
  • this site configuration and configuration work becomes extremely laborious and difficult. This work needs to be carried out after the construction of the building is completed and the electromechanical equipment is in place.
  • the construction period that can be utilized is very short, so the time is hasty; when the building layout or function division changes in the later stage, the automatic control system is difficult to change flexibly;
  • the self-control platform is closed and the versatility is poor.
  • the control software is often designed separately for individual buildings. Therefore, in the process of transforming and expanding the system, the new control strategy is difficult to implement flexibly and simply on the existing automatic control platform.
  • the present invention aims to solve at least one of the technical problems existing in the prior art.
  • the computing network system and the computing node (CPN) provided by the present invention are generated and applicable to the field of building automation, this does not constitute a limitation of the scope of the present invention.
  • the computing network system of the present invention serves as a The basic computing network can also be applied to other areas.
  • a computing node for a distributed computing network
  • the computing node (CPN) being a computer having information receiving, processing, and transmitting functions
  • the computing node (CPN) has a central processing unit, memory, communication interface; several A computing node (CPN) constitutes a distributed computing network; each computing node (CPN) performs data interaction with its topologically adjacent computing node (CPN); the computing node (CPN) has spatial attributes, the spatial attributes Reflected as an absolute spatial location of the computing node (CPN) and/or a relative spatial location of the computing node (CPN) in the topology network in which it is located; the computing node has an operating system built in, the operating system provides an API An interface by which a user can convert various management/control requirements and/or policies into a standard computing sequence; the computing nodes (CPNs) in the distributed computing network are collectively and self-organized Calculate the sequence.
  • a distributed computing network system which is composed of a plurality of computing nodes (CPNs) having information receiving, processing, and transmitting. a functional computer; each compute node (CPN) is in data interaction with its topologically adjacent compute node (CPN); the compute node (CPN) has spatial attributes, which are embodied as the compute node (CPN) An absolute spatial location and/or a relative spatial location of the computing node (CPN) in the topology network in which it is located; the distributed computing network system embedding an operating system, the operating system or the operating system At least a portion is distributed within each of the computing nodes (CPNs); the operating system provides an API interface through which a user can translate various management/control requirements and/or policies into a standard computing sequence; The computing nodes (CPNs) in the distributed computing network system collectively and self-organized to complete the computing sequence.
  • each compute node (CPN) is in data interaction with its topologically adjacent compute node (CPN); the compute node (CPN) has spatial attributes
  • Each computing node is associated with a basic spatial unit or a regional control system of an electromechanical device.
  • a computing node is associated with the regional control system, the basic spatial unit or electromechanical
  • the location space information, relative positional relationship or topological relationship of the device is naturally embodied on the computing node (CPN), thus having the advantage of rapid deployment, and eliminating the massive and repeated field wiring of the original control system.
  • Adaptation, debugging, and definition work save a lot of manpower;
  • Standardization The relevant information of the basic space unit or electromechanical device is described in the form of a standard data table. After the computing node is associated with the regional control system, the computing node can automatically recognize that it is associated with a basic space unit or some Electromechanical devices, thus enabling the calculation of node (CPN) plug-and-play, automatic identification;
  • CPN node
  • no central computing the entire computing network system is flat, non-centralized, each node The status is completely equal, and the global calculation is distributedly distributed through data interaction between the nodes, and various management control strategies running on the system are embodied and completed by the distributed calculation;
  • This system provides an open, user-friendly programming platform, users can easily complete the definition of events / tasks using the operator / algorithm library provided by the system, the system automatically compiles the underlying program code, thus Realize the rapid management of control management strategy software, with the advantages of agile development; on the programming platform can develop a large number of applications, with great compatibility and flexibility.
  • 1 is an organizational diagram of a master control system in the prior art.
  • FIG. 2 is a schematic structural diagram of a distributed computing network system according to an embodiment of the present invention.
  • FIG. 3 is a schematic diagram of a first implementation manner of a communication manner between a computing node and a DCS system according to an embodiment of the present invention.
  • FIG. 4 is a schematic diagram of another implementation manner of a communication manner between a computing node and a DCS system according to an embodiment of the present invention.
  • FIG. 5 is a schematic structural diagram of a computing node (CPN) according to an embodiment of the present invention.
  • FIG. 6 is a schematic diagram of a library module of a user module of an operating system according to an embodiment of the present invention.
  • FIG. 7 is a schematic diagram of a user interaction module of a user module of an operating system according to an embodiment of the present invention.
  • FIG. 8 is a schematic diagram of a kernel module of an operating system of an embodiment of the present invention.
  • FIG. 9 is a schematic diagram of a topology and naming rules of a distributed computing network system according to an embodiment of the present invention.
  • FIG. 10 is a schematic structural diagram of a communication frame between a user module and a kernel module according to an embodiment of the present invention.
  • Figure 11 is a schematic diagram showing the principle of representing information from three dimensions.
  • Figure 12 is a schematic diagram showing the composition of an event of an embodiment of the present invention.
  • Fig. 13 is a schematic diagram of a first embodiment of the embodiment of the present invention.
  • Fig. 14 is another schematic view of the first embodiment of the embodiment of the present invention.
  • the computing network system is composed of a plurality of computing nodes (CPNs, computing nodes), and all computing nodes together form a flat, non-centered
  • the computing network, the status of each computing node is equal.
  • the computing node is a computer (such as a small computer) having information receiving, processing, and transmitting functions.
  • the structure and composition of the computing node are as shown in FIG. 5, and each computing node has a processor 1, a memory 2, and a communication interface 3.
  • Each computing node performs data interaction with its neighboring computing nodes, and the data interaction is one-hop communication. After the processing of the information, data interaction is performed with the computing nodes adjacent to the topology, and the data interaction is also a hop. Communication, all computing nodes share computing tasks collectively; that is to say, the distributed computing network provided by the present invention disassembles tasks into typical, replicable basic calculations, and the basic calculation is through each calculation The node obtains the input information from the neighbor node, completes the local calculation, and transmits the calculation result to the neighbor node. The entire network has no concept of center or summit, and each node cooperates with each other to complete the system. Calculation.
  • a compute node has a spatial attribute that is embodied by the absolute spatial location of the compute node (CPN) and/or the relative spatial location of the compute node (CPN) in the topology network in which it is located.
  • the distributed computing network system has an embedded operating system.
  • the operating system includes a kernel module and a user module.
  • the kernel modules of the operating system are distributed in each computing node, and the kernel modules in each computing node are identical.
  • the network formed by the interconnection of computing nodes forms a large computer, and the calculation process is distributed in each computing node (CPN); a single computing node can also be a comprehensive information processing and running calculation.
  • the ability of the computer, but the ability to process information and computing is weak, the system composed of multiple computing nodes interconnected has the ability to store, calculate and communicate more information.
  • cache modules in the memory of the computing node (CPN), which are used to store the status of tasks and/or events.
  • the cache module can be 1024 or any number of multiple modules.
  • the system can process multiple tasks or events in parallel at the same time. Complex and varied control task requirements.
  • each computing node (CPN) adopts a localized naming manner when accessing the network, that is, each computing node (CPN) naming only needs to distinguish adjacent nodes, and does not need to be globally named. Or through the routing configuration, as shown in Figure 9, the node named 1, 4 or 5 appears multiple times in the network, and this does not affect the normal operation of the computing network, each computing node (CPN) can be directly and Data interaction only with compute nodes directly adjacent to the topology, and thus adjacent node names are different That is, it can effectively avoid the work of massive site naming and system configuration.
  • Each computing node has several communication interfaces (ie, API interfaces), and the communication interface is divided into two types, A and B.
  • the computing nodes are interconnected with their neighboring computing nodes through the class A communication interface, and pass through B.
  • the class communication interface is interconnected with the District Control System.
  • the regional control system is associated with a basic space unit (such as a subspace of a building: an office, a corridor, etc.) or an electromechanical device (such as a cold machine, a water pump, etc.), and the regional control system is used for the collection center.
  • All the measurement and control information of the subspace or electromechanical equipment or the controller or actuator for controlling the subspace or electromechanical equipment, the computing node (CPN) and the corresponding DCS realize the information interaction according to the standard information set, thereby being able to
  • the relevant information required for the calculation is obtained at the DCS, or the calculated result is sent to the corresponding DCS so that it performs the relevant control.
  • the communication between the computing node (CPN) and the regional control system (DCS) includes but is not limited to the following two types:
  • the first communication method is shown in Figure 3:
  • the computing node interacts directly with the regional controller (DCU) in the DCS, and the DCU performs data interaction with each controller, sensor or actuator.
  • DCU regional controller
  • the DCU and each controller, sensor or actuator form a master-slave relationship.
  • the second communication method is shown in Figure 4:
  • the regional controller (DCU) in the DCS and each controller, sensor, actuator or lamp, louver, FCU intelligent device are co-located on a local network system (which may be a local area network), and the computing node is interconnected with the local network system.
  • Information exchange can be performed between the CPN, the DCU, and each controller, sensor, actuator, or lamp, louver, and FCU smart device, and there is no master-slave relationship between the above components.
  • the functions of the DCU can be integrated into the CPN, that is, a module equivalent to the DCU is added to the CPN, which corresponds to the subspace or electromechanical device of the building, collects relevant information or performs related control.
  • the network formed by each CPN interconnection belongs to a physical tangible network, on which multiple virtual functional subnets can be defined, such as indoor transportation network, air flow network, heat transfer network, power distribution system, Air conditioning water system, water supply system, domestic hot water system, air conditioning air system, Exhaust ventilation system, fresh air system, gas supply system, cold station, thermal station, etc., can be calculated within a single functional subnet, or multiple functional subnets can be associated, so that information can be realized in multiple subnets. The true sense of sharing.
  • the functional subnet is determined by the content of the specific application, and is a virtual network above the distributed computing network provided by the present invention, and its topology is consistent with the physical network.
  • the same computing node may belong to multiple functional subnets; multiple functional subnets do not affect each other, so the computing nodes can simultaneously Handle computing tasks for multiple functional subnets in parallel and in parallel.
  • each computing node plays a different role in the functional subnet.
  • the CPNs of all spatial units in the air flow subnet are equal to each other; and the air conditioning in the air conditioning box air system
  • the CPN of the box and the CPN of the subordinate space constitute a master-slave relationship.
  • Standard Computational Sequence The user translates all control management requirements and tasks into a standard computational sequence through an API interface in the operating system provided by the distributed computing network.
  • the standard calculation sequence includes several calculation units, which are specifically a task or event, so the definition of the standard calculation sequence includes the following:
  • Task A control management strategy that is coordinated by multiple computing nodes (CPNs). It consists of several events in a certain sequence.
  • Event is a centerless calculation done by the Computational Node (CPN), which is determined by defining operators/algorithms, input variables, output variables, and intermediate variables.
  • CPN Computational Node
  • Variable individual physical information related to the basic space unit and/or individual physical information related to the electromechanical device and/or user-defined temporary variables; for example: area of the area in the basic information, indoor temperature in the environmental and system measurement parameters, intelligence Feedback of the lighting switch status in the equipment operating parameters, zone The set value of the fan coil operating state in the domain, the indoor ambient temperature setting value, the fire alarm of the fire equipment, the total average power consumption of the area in the past 10 minutes, and so on.
  • Algorithm The operator provided by the system (belonging to the lower concept of the algorithm), the basic algorithm, the advanced algorithm or the user-defined algorithm; the user is allowed to download the algorithm to the computing node (CPN), and the support system adopts a mechanism similar to the transmission calculation, and the algorithm is used. Passed to each relevant compute node (CPN); user-defined algorithms can be deleted by defining the effective range and/or effective time of the algorithm.
  • Each information point involved in running the management control strategy generally has three dimensions of information, as shown in FIG. 11, the first dimension solves where/what is, that is, where the information point is located (including the space in which it is located) , location, close to the trunk or branch, etc. with spatial attributes), what is it (is a room or a water pump or a cold machine), the second dimension solves the data represented by the information points The specific meaning (indicating temperature or humidity or specific content information such as energy consumption), the third dimension is the change of data as time goes by.
  • the DCS itself has a spatial attribute (the DCS is either a basic space unit, that is, a room or a corridor, at the moment when the computing node (CPN) is combined or communicated with the DCS.
  • the DCS is either a basic space unit, that is, a room or a corridor, at the moment when the computing node (CPN) is combined or communicated with the DCS.
  • the network formed after the CPN network forms an Internet of Things with spatial attributes;
  • the standard information set is composed of several standard data tables, and each basic space unit or electromechanical device corresponds to A standard data table (as shown in Table 1), which defines which variables need to be collected, variable names, standard formats of variables, etc., all basic space units or electromechanical devices follow this unified standard information set.
  • Table 1 A standard data table (as shown in Table 1), which defines which variables need to be collected, variable names, standard formats of variables, etc., all basic space units or electromechanical devices follow this unified standard information set.
  • the CPN can obtain the information of the corresponding area or device after data interaction with the DCS.
  • the No. 5 data in the standard information set is “room air temperature”. If the device connected to the DCS has a temperature measuring point, Then the No.
  • CPN can obtain information from the DCS regardless of the device and communication protocol used at the bottom; in this way, When the CPN is connected to the DCS in the field, it can be identified through the information exchange CPN whether the DCS corresponds to the building space or some kind of electromechanical equipment, such as a cold machine or a water pump, so that the CPN plug-and-play and automatic identification can be realized. .
  • Table 1 is an example of the standard data table.
  • the characterization of the third dimension is realized by the record storage function of the DCS or CPN, and the information points can be recorded according to the changing development of time, and will not be described again.
  • the computing node in the distributed computing network system provided by the present invention can be plugged and inserted. With intelligent identification, it is extremely robust and flexible.
  • the user module is specifically an API based on a communication protocol, and the API includes a library module and a user interaction module, that is, the operating system provides various levels of algorithm libraries from simple mathematical calculations to professional application algorithms, and the user can call when writing a standard calculation sequence.
  • the algorithm in the algorithm library the operating system automatically forms the underlying program code to achieve agile programming; combined with Figure 6, the library module has a three-level operator/algorithm library, which is an operator library, a basic algorithm library, and an advanced algorithm library.
  • the operator library includes: addition, subtraction, multiplication, division, weighted summation, product, logic operation (sum, OR, non-etc.), seeking the maximum value, the minimum value, the set operation (intersection, union, etc.), generating Tree, Jacobi/Gaussell iteration, and other common basic mathematical operations.
  • the basic algorithm library includes: matrix calculation algorithm, steepest descent method, Newton method, genetic algorithm, neuron algorithm, and other common basic mathematical algorithms.
  • the advanced algorithm library includes: sensor fault diagnosis algorithm, number distribution check algorithm, fire inversion algorithm, region-based CFD algorithm, and other advanced algorithms applied to various professional fields.
  • the user interaction module provides the following interfaces to the user: a task definition interface, an event definition interface, an algorithm definition interface, and a variable definition interface;
  • the user can define one or more of the following items: task name, task execution condition, total number of task steps, events included in the task, execution steps of the event, and calculation structure included in the task return result. .
  • event definition interface the user can define one or more of the following: event name, which task the event belongs to, event execution condition, input variable of the event, output variable, name of the intermediate variable, event call Operator/algorithm.
  • the user can define one or more of the following items: which task/event, algorithm name, and specific content of the algorithm;
  • variable definition interface the user can define one or more of the following: the variable name, which task and/or event the variable belongs to, the variable byte length, and the initial value of the variable.
  • the user module can use the computing node as the carrier, or the regional control system as the carrier, and can also adopt another independent software and hardware as the carrier, and the setting mode is flexible and changeable.
  • the kernel module further includes:
  • Calculation processing module used to perform a basic calculation
  • Parallel computing collaboration service module for supporting and servicing parallel computing; enabling the system to perform parallel computation of multiple standard computational sequences.
  • Communication module used to complete communication tasks.
  • the parallel computing collaboration service module further includes the following sub-modules: a task management module, an event management module, an operator/algorithm management module, and a variable management module.
  • the communication module further includes the following sub-modules: an underlying communication interface driver module, a communication protocol parsing module, and a communication frame editing module.
  • the modules have the following synergy:
  • the underlying communication interface driver module receives information from the communication interface and stores the received information in a receive buffer; the receive buffer is located in a memory of a compute node (CPN).
  • CPN compute node
  • the communication protocol parsing module retrieves the information from the receiving buffer and parses the information into four subclasses of task related information, event related information, operator/algorithm related information, and variable related information, and presses the processed information.
  • the categories are respectively stored in the task management space, the event management space, the operator/algorithm management space, and the variable management space, and the task management space, the event management space, the operator/algorithm management space, and the variable management space are all located at the computing node (CPN). In memory.
  • the task management module manages and maintains task related information in the task management space.
  • the event management module manages and maintains event related information in the event management space.
  • the operator/algorithm management module manages and maintains operator/algorithm related information in the operator/algorithm management space.
  • the variable management module manages and maintains variable related information in the variable management space.
  • the calculation processing module calls relevant information from the task management space, the event management space, the operator/algorithm management space, and the variable management space, and performs calculation, and then stores the calculated results in the task management space and the event management space according to the categories. , operator / algorithm management space, variable management space.
  • the communication frame editing module calls relevant information from the task management space, the event management space, the operator/algorithm management space, and the variable management space and edits it into a communication frame, which is stored in the sending buffer, and the sending buffer is located in the memory of the computing node.
  • the underlying communication interface driver module sends the communication frame in the transmission buffer through the communication interface.
  • the kernel module and the user module exchange information under a certain communication protocol.
  • the communication protocol may be various communication protocols existing in the prior art, or may be a customized communication protocol, such as a knot.
  • the communication frame sent by the user module to the kernel module may adopt a three-layer structure, and the bottom layer is a physical link layer protocol, including a synchronization code, a length, a check portion, and a data portion, and the physical link layer may be adopted.
  • the second layer is the underlying layer of the application, including the frame header and data part, for handling communication problems such as communication response, verification, and fragment transmission;
  • the third layer is the application upper layer including The message type and the message content part are used to process task definitions, event definitions, algorithm downloads, variable assignments, and system upgrades.
  • the structure of the application layer can also be in other ways.
  • the user module sends one or several forms of user-defined tasks, events, algorithms, and variables to the kernel module in the form of the communication frame shown in FIG.
  • the various operational management tasks and strategies running on the distributed computing network system provided by the present invention are embodied and completed by a standard computing sequence. After the network completes the standard computing sequence, the tasks and strategies for running management are basically completed.
  • the following example is used to introduce how the computing nodes of the system cooperate to complete distributed computing.
  • the standard calculation sequence is composed of several calculation units, which refer to tasks or events, and the task is composed of several events according to a certain sequence.
  • the event is a basic calculation completed by a computing node, which passes Input variables, output variables, intermediate variables (which may not be available in some cases), and operators/algorithms are defined, as shown in Figure 12.
  • the chilled water pump in the air conditioning chilled water system in the building is adjusted according to the end cooling condition.
  • the control requirement is to reduce the differential pressure setting value of the chilled water pump to reduce the pump when the end cooling demand is met. Energy consumption.
  • Network structure As shown in Figure 13, the left half of the figure shows the chilled water system control network.
  • the end of the water system is actually a network of multiple branches.
  • a riser branch is used to simplify the representation.
  • the space node network is also a multi-dimensional network structure.
  • the figure is simplified by a chain structure.
  • the chilled water system adopts a primary pump setting mode, two sets of cold machines correspond to three refrigerating pumps, and the chilled water flows through the water separator to the discs in the respective rooms at the end. tube.
  • the cold machine and the water pump respectively correspond to a computing node of a device type, and the end coil belongs to a computing node of a space type, and the cooling room where the cold machine is located is provided with a computing node of a spatial type, and the computing nodes are interconnected to form a function.
  • the subnet has a topology as shown in the right half of Figure 13.
  • Control logic If there is insufficient water at the end equipment, it will increase according to the number of ends that do not meet the requirements.
  • the differential pressure setting value increases the pump speed or increases the pump; if all the ends meet the requirements, if there is too much end water, reduce the differential pressure setting to reduce the pump speed or shut down the pump.
  • the whole problem can be decomposed into two relatively independent control loops.
  • One is to adjust the set value of the differential pressure across the chilled water pump according to the end supply condition, and the other is to adjust the number and frequency of the pump according to the set value of the differential pressure.
  • the pump has the lowest energy consumption.
  • the calculation process of the computing network system provided by the present invention is described herein for the first control loop, namely how to determine the differential pressure setpoint.
  • the task of automatically adjusting the differential pressure setpoint of the pump is accomplished by a combination of three events, as shown in Table 2, where events 1 and 2 can be calculated in parallel, and the results of events 1 and 2 are input as event 3, the entire process flow Going forward, no looping, once every time (for example 5 minutes).
  • the specific process includes the following steps:
  • the space node to which each end coil belongs is judged whether the cooling demand of the area is sufficient according to the temperature change in the area.
  • the basic logic is: the control accuracy of the current temperature measurement value in the area at the temperature set value. Within the range, and the valve opening (or duty cycle) of the coil in the area is not completely closed, the current cooling capacity is considered to meet the demand; if the current area temperature measurement is higher than the temperature setting value The range of control accuracy, and the valve opening or duty cycle of the coil in the area is fully open, or within a very high threshold, it is considered that the cooling capacity in the current area cannot meet the demand, and a "too hot” signal is issued; If the current temperature measurement is lower than the control accuracy range of the temperature set value, and the valve opening or duty cycle of the coil in the area is fully closed, or within a very low threshold, the current area is considered to have exceeded the actual amount of cooling.
  • the cold demand, issued a "too cold” signal can define two variables, variable 1 is "whether it is overheated", yes,
  • the end spatial node forms a spanning tree, starting from the first computing node, passing the value of the variable 1 or 2 to the neighboring node, and the neighboring node "locally sums" the number passed by the neighbor and the local data. Then pass this calculation result to the next node, and so on, the summation of all nodes is completed when passing to the end node; in the whole process, each node participates in the summation calculation, but none of the nodes know the whole How many nodes are there in the system. Event 1 that counts whether it is overheated and event 2 whose statistics are too cold can be calculated in parallel.
  • the "whether it is too cold" or “too hot” variable can be either a system variable provided by the computing network system or a custom variable when the user compiles the program;
  • the "local summation” It is a basic operator provided by the library module in the user module, which can be built in each computing node, and the user can directly call it, without having to specifically compile the implemented code;
  • the "spanning tree” algorithm is also calculated by The network provides, can be directly called; and the specific input variables, output variables, sequence, etc. of the above events 1, 2, 3 can be compiled by the user.
  • An infrared detector is installed at the doorway of the space-connected area.
  • the infrared detector sends a signal of the number-1 to the computing node of the area A, and the computing node of the area A receives After the signal, the number of people in the area will be -1, and a neighbor node B will be initiated.
  • the calculation makes the number of people in the B area +1, thereby simulating the physical process of the person entering the B area from the A area corresponding to the actual process on the system.
  • All infrared detectors installed at the junction of the communicating area will detect the transfer of personnel in the building in real time and send signals to the computing nodes in the corresponding area to trigger real-time calculation of the entire system, so that the number of personnel in all areas can be obtained. distributed.
  • This information can be used as the basic information to achieve optimal control of other functional subsystems (such as air conditioning systems, lighting systems).
  • Each electromechanical device of the system can record its own energy consumption through DCS or CPN, or DCS or CPN can calculate the power consumption of the device according to the running time of the electromechanical device, so the energy of all devices in each basic space unit The value of the consumption can be counted by the system. From the perspective of property management personnel, there is a need to collect energy consumption values for all regions for comparative analysis.
  • the entire network system first forms a spanning tree, and then returns its own energy consumption value to its neighbor nodes at the end node of each branch; then each computing node receives After the energy value passed by the neighbor is added, the energy consumption value of the corresponding area is also added, repackaged, and sent to other neighbors (ie, input other neighbor nodes other than the neighbor node of the energy consumption information); the computing network All the computing nodes in the system follow this principle. Finally, the energy consumption information of all nodes is packaged and sent to the computing node of the task initiator, that is, the task of collecting the global summation and collecting the energy consumption values of all basic space units is completed. .
  • the computing node (CPN) for distributed computing networks provided by the present invention has the following characteristics:
  • Each computing node is associated with a basic spatial unit or a regional control system of an electromechanical device.
  • a computing node is associated with the regional control system, the basic spatial unit or electromechanical
  • the location space information, relative positional relationship or topological relationship of the device is naturally embodied on the computing node (CPN), thus having the advantage of rapid deployment, and eliminating the massive and repeated field wiring of the original control system.
  • Adaptation, debugging, and definition work save a lot of manpower;
  • Standardization The relevant information of the basic space unit or electromechanical equipment is described in the form of a standard data table. After the computing node is associated with the regional control system, the computing node can automatically identify its location. Associated with the basic space unit or an electromechanical device, thus enabling the computing node (CPN) plug and play, automatic identification;
  • CPN computing node
  • Centerless computing The entire computing network system consisting of Computational Nodes (CPN) for distributed computing networks is flat and non-centralized. The status of each node is completely equal, and data interaction between nodes is achieved. Globally computing is performed distributedly, and various management control policies running on the system are embodied and completed by the distributed computing;
  • CPN Computational Nodes
  • the system consisting of computing nodes (CPN) for distributed computing networks provides an open and user-friendly programming platform that can be easily accomplished by the user using the operator/algorithm library provided by the system.
  • the definition of the event/task the system automatically compiles the underlying program code, so that the control management strategy is quickly software coded, which has the advantages of agile development; on the programming platform, a large number of applications can be developed with great compatibility and flexibility.

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computer And Data Communications (AREA)
  • Programmable Controllers (AREA)

Abstract

一种分布式计算网络系统和用于该系统的计算节点CPN,该分布式计算网络系统由若干个计算节点CPN构成,CPN为具有信息接收、处理、发送功能的计算机;每个CPN均与其拓扑上相邻的CPN进行数据交互;CPN具有空间属性,空间属性体现为CPN所在的绝对空间位置和/或CPN在所处的拓扑网络中的相对空间位置;分布式计算网络系统内嵌操作系统,操作系统的至少一部分分布于每个CPN内;操作系统提供API接口,将管理/控制的需求和/或策略转化为标准计算序列,并转化为操作系统可识别的指令序列;分布式计算网络系统中的CPN分布式地、自组织地共同完成所述计算序列。提供了一个开放的、扁平的、易编程的控制平台。

Description

分布式计算网络系统及用于该系统的计算节点
本申请要求申请日为2015年6月30日、申请号为201510377980.1和题目为“分布式计算网络系统”的中国专利申请和申请日为2015年6月30日、申请号为201510378076.2和题目为“用于分布式计算网络的计算节点”的中国专利申请的优先权,在此以参考的方式插入这些申请的所有内容。
技术领域
本发明涉及计算网络技术领域,具体涉及一种分布式计算网络系统及用于该系统的计算节点。
背景技术
20世纪80年代以来,人们开始利用信息技术实现建筑的智能化自动控制,例如面向特定服务功能的广播音响系统、IC卡管理系统、酒店客房管理系统、能源监测管理系统,以及面向特定机电设备的空调控制系统、安防系统、消防系统、冷站控制系统、用电安全系统等,然而问题是现有技术中的管理控制系统往往都不那么“智能”,超过半数的建筑自控系统只能实现在中央控制室远程监测建筑环境和系统设备的运行参数,以及通过中控人机界面手动地启停或调节机电设备运行状态。这样的系统实质上仍严重依赖于运行人员的手动操作,没有实现自动化和智能化。只有极少数的建筑,能够实现楼宇层面的自动化控制和管理,包括各个子系统内部的优化控制,和子系统之间的集成控制。
造成这种局面的根本原因在于自控系统的集中式的架构,如图1所示,现有技术中的自控系统采用集中式的组织架构,所有终端测控点(传感器、执行器、现场控制器)都通过总线通信网络连通,各个子系统(照明系统、空调系统、消防系统、安防系统)的末端的信息测控点虽然很大程度上分布在同一个建筑子空间内,但却是按照不同的子系统进行纵向垂直集成,这一集中式的自控系统具有如下主要缺点:
1、终端的测控点需要进行全局通信命名、定义物理属性并定义相互之间的关联关系,当测控点数较多时,这一现场配置和组态的工作变得工作量极大,难度很高;这一工作需要在建筑建造完成及机电设备就位后才能开展,可以利用的施工周期很短,因而时间仓促;在后期建筑布局或功能划分发生变化时,自控系统难以随之灵活改变;
2、各子系统之间难以实现信息的真正共享,要想实现跨系统信息共享,就要在现有若干系统的上层建立一个新系统,这就需要对全局重新进行组态和定义,难度和成本极高,不适应建筑控制智能化、信息化、前端化的需求;
3、自控平台封闭、通用性差,控制软件往往是针对个别建筑单独设计的,因而在对系统进行改造和扩展的过程中,新的控制策略难以在现有的自控平台上灵活简单地实现。
4、跨系统集成困难,现有的自控系统及平台通用性差、软硬件环境不友好,要求开发者具有较高的IT专业知识,而自控系统中运行的各种控制策略及控制逻辑(如暖通系统的控制策略、消防系统的控制策略、安防系统的控制策略)又往往是由各个领域的工程师(暖通工程师、消防安全工程师等)来制定的,这就导致各领域的工程师很难将其制定好的控制策略和控制逻辑转化为自控系统的控制软件,因而导致各个子系统应该实现的功能很难集成在现有的自控系统和自控平台当中。
发明内容
本发明旨在至少解决现有技术中存在的技术问题之一。
为此,本发明的一个目的在于提出一种分布式的计算网络系统和用于该系统的计算节点(CPN),该计算网络系统和计算节点(CPN)能够应用于建筑自控系统,提供一个开放的、扁平的、易编程的控制平台。另外需要说明的是虽然本发明提供的计算网络系统和计算节点(CPN)产生于并可以应用于楼宇自控领域,但这并不构成对本发明的保护范围的限定,本发明的计算网络系统作为一个基础性的计算网络还可以应用于其他领域。
为了实现上述目的,本发明的实施例公开了一种用于分布式计算网络的计算节点(CPN),所述计算节点(CPN)为具有信息接收、处理、发送功能的计算机,所述计算节点(CPN)具有中央处理器、存储器、通信接口;若干个所 述计算节点(CPN)构成分布式计算网络;每个计算节点(CPN)均与其拓扑上相邻的计算节点(CPN)进行数据交互;所述计算节点(CPN)具有空间属性,所述空间属性体现为所述计算节点(CPN)所在的绝对空间位置和/或所述计算节点(CPN)在所处的拓扑网络中的相对空间位置;所述计算节点内置操作系统,所述操作系统提供API接口,用户可通过所述API接口将各种管理/控制的需求和/或策略转化为标准计算序列;所述分布式计算网络中的计算节点(CPN)分布式地、自组织地共同完成所述计算序列。
为了实现上述目的,本发明的实施例公开了一种分布式计算网络系统,所述计算网络系统由若干个计算节点(CPN)构成,所述计算节点(CPN)为具有信息接收、处理、发送功能的计算机;每个计算节点(CPN)均与其拓扑上相邻的计算节点(CPN)进行数据交互;所述计算节点(CPN)具有空间属性,所述空间属性体现为所述计算节点(CPN)所在的绝对空间位置和/或所述计算节点(CPN)在所处的拓扑网络中的相对空间位置;所述分布式计算网络系统内嵌操作系统,所述操作系统或所述操作系统的至少一部分分布于每个所述计算节点(CPN)内;所述操作系统提供API接口,用户可通过所述API接口将各种管理/控制的需求和/或策略转化为标准计算序列;所述分布式计算网络系统中的计算节点(CPN)分布式地、自组织地共同完成所述计算序列。
本发明提供的分布式计算网络系统和用于该系统的计算节点具有如下特点:
1、面向空间:每个计算节点(CPN)均与一个基本空间单元或者某个机电设备的区域控制系统相互关联,在计算节点(CPN)关联到所述区域控制系统时,基本空间单元或机电设备所具有的位置空间信息、相对位置关系或拓扑关系便自然而然地体现在所述计算节点(CPN)上,因而具有快速部署的优点,能够免去原有控制系统的海量而反复的现场布线、适配、调试、定义的工作,节省大量人力;
2、标准化:所述基本空间单元或机电设备的相关信息全部以标准数据表的形式进行描述,计算节点与区域控制系统关联后计算节点能够自动识别出其所关联的是基本空间单元或者某个机电设备,因而能够实现计算节点(CPN)即插即用,自动识别;
3、无中心计算:整个计算网络系统是扁平化的、无中心化的,各个节点的 地位完全平等,通过各个节点之间的数据交互分布式地完成全局的计算,在所述系统上运行的各种管理控制策略是通过所述分布式的计算来体现和完成的;
4、快速友好的编程环境:本系统提供了一个开放、人性化的编程平台,用户利用本系统提供的算子/算法库能够轻松的完成事件/任务的定义,系统自动编译底层程序代码,从而实现控制管理策略迅速软件代码化,具有敏捷开发的优点;在所述编程平台上能够开发出海量的应用程序,具有极大的兼容性和灵活性。
附图说明
本发明的上述和/或附加的方面和优点从结合下面附图对实施例的描述中将变得明显和容易理解,其中:
图1是现有技术中的总控系统的组织架构图。
图2是本发明的实施例的分布式计算网络系统的结构示意图。
图3是本发明的实施例的计算节点与DCS系统的通讯方式的第一种实现方式示意图。
图4是本发明的实施例的计算节点与DCS系统的通讯方式的另一种实现方式示意图。
图5是本发明的实施例的计算节点(CPN)的结构示意图。
图6是本发明的实施例的操作系统的用户模块之库模块的示意图。
图7是本发明的实施例的操作系统的用户模块之用户交互模块的示意图。
图8是本发明的实施例的操作系统的内核模块的示意图。
图9是本发明的实施例的分布式计算网络系统的拓扑及命名规则示意图。
图10是本发明的实施例的用户模块与内核模块之间的通信帧的结构示意图。
图11是从三个维度表示信息的原理示意图。
图12是本发明的实施例的一项事件的组成示意图。
图13是本发明的实施例的算例1的示意图。
图14是本发明的实施例的算例1的另一个示意图。
具体实施方式
本发明实施例的分布式计算网络系统的基本结构:
如图2所示,示出了本发明实施例的分布式计算网络系统的结构,计算网络系统由若干个计算节点(CPN,computing node)构成,所有计算节点共同组成了一个扁平化的无中心的计算网络,每个计算节点的地位均是平等的。
计算节点为具有信息接收、处理、发送功能的计算机(如小型计算机),计算节点的结构及组成如图5所示,每个计算节点均具有处理器1、存储器2及通信接口3。
每个计算节点均与其拓扑上相邻的计算节点进行数据交互,数据交互为一跳通信,经过信息的加工处理后再与其拓扑上相邻的计算节点进行数据交互,所述数据交互也是一跳通信,所有的计算节点分布式地共同完成计算任务;也就是说本发明提供的分布式的计算网络将任务拆解为典型的、可以复制的基本的计算,而基本的计算是通过每个计算节点从邻居节点处获取输入信息,完成本地计算,再将计算结果传送给邻居节点完成的,整个网络没有中心或首脑的概念,各个节点之间以一种自组织的协作机制相互配合来完成系统的计算。
计算节点(CPN)具有空间属性,空间属性体现为计算节点(CPN)所在的绝对空间位置和/或计算节点(CPN)在所处的拓扑网络中的相对空间位置。
再次结合图5,分布式计算网络系统内嵌操作系统,操作系统包括内核模块和用户模块,操作系统的内核模块分布于每一个计算节点,且每个计算节点中的内核模块是完全相同的。
由此可知由计算节点(CPN)互连形成的网络形成了一个大的计算机,计算过程分布于每个计算节点(CPN)中;单个计算节点自己也可以是一个具有全面的信息处理、运行计算能力的计算机,但是处理信息和计算的能力较弱,多个计算节点互联后所组成的系统具备更强的信息存储、计算及通信的能力。
计算节点(CPN)的存储器中存在若干个缓存模块,用来存储任务和/或事件的状态,缓存模块具体可以为1024个或任意多个,系统能够同时并行处理多个任务或事件,能够适应复杂多变的控制任务需求。
结合图9,每个计算节点(CPN)在接入网络时,均采用局部化的命名方式,即每一个计算节点(CPN)的命名只需要区别相邻节点即可,并不需要进行全局命名或经过路由配置,如图9所示,名称为1、4或5的节点在网络中多次出现,而这并不会影响计算网络的正常运行,每个计算节点(CPN)可直接、且只与其拓扑上直接相邻的计算节点进行数据交互,因而相邻的节点名称不同 即可,因而能够有效规避海量现场命名和系统组态的工作。
每个计算节点(CPN)均具有若干个通信接口(即API接口),通信接口分为A类和B类两种,计算节点通过A类通信接口与其拓扑上相邻的计算节点互联,通过B类通信接口与区域控制系统(District Control System)互联。而所述区域控制系统与某个基本空间单元(如建筑的子空间:一间办公室、一个走廊等)或某个机电设备(如冷机、水泵等)相关联,区域控制系统用于收集所述子空间或机电设备的所有测控信息或用于控制与子空间或机电设备相关的控制器或执行器,计算节点(CPN)与相应的DCS之间按照标准信息集实现信息交互,从而能够从DCS处获取计算所需要的相关信息,或者将计算得到的结果发送给相应DCS以便其执行相关的控制。
计算节点(CPN)与区域控制系统(DCS)之间的通讯方式包括但不限于以下两种:
第一种通讯方式结合图3所示:
计算节点与DCS中的区域控制器(DCU)直接进行数据交互,DCU再与各个控制器、传感器或执行器进行数据交互,所述DCU和各控制器、传感器或执行器构成主从关系。
第二种通讯方式结合图4所示:
DCS中的区域控制器(DCU)及各个控制器、传感器、执行器或灯、百叶、FCU智能设备为共同位于一个局部网络系统之上(可以是一个局域网),计算节点与局部网络系统互联,所述CPN、DCU和各个控制器、传感器、执行器或灯、百叶、FCU智能设备之间均可以进行信息交互,上述各部件之间无主从关系。
CPN与DCS之间的通讯方式支持多种通讯协议,从而使得CPN可以与目前现有技术中的各类DCU产品兼容。
此外还可以将DCU的功能整合到CPN中,即在CPN中增加一个相当于DCU的模块,该模块与建筑的子空间或机电设备相对应,收集相关信息或者执行相关的控制。
拓扑关系与功能子网:
各个CPN互联所形成的网络属于一个物理意义上的有形的网络,在该网络之上可以定义出多个虚拟的功能子网,例如室内交通网络、空气流动网络、传热网络、配电系统、空调水系统,自来水系统,生活热水系统,空调风系统, 排烟通风系统,新风系统,供燃气系统,冷站,热力站等,在单个功能子网内部可以进行计算,也可以将多个功能子网进行关联计算,因而能够实现信息在多个子网中的真正意义上的共享。
功能子网由具体应用的内容决定,是本发明提供的分布式的计算网络之上的虚拟网,其拓扑结构与实体网络一致。
由于本发明的计算网络之上可以定义多个虚拟子网,因此同一个计算节点(CPN)可能会隶属于多个功能子网;多个功能子网之间互不影响,因而计算节点能够同时地、并行地处理多个功能子网的计算任务。
因应用功能的不同,每个计算节点(CPN)在功能子网中起到的作用也有差异,例如空气流动子网中所有空间单元的CPN都是相互平等的;而在空调箱风系统中空调箱的CPN与下属空间的CPN则构成主从关系。
如此,通过在本发明提供的分布式的计算网络上定义功能子网就自然而然地完成了系统组态的工作,通过功能子网进一步定义计算程序并在子网上运行计算程序就可以完成各种功能子网所需的管理、控制任务。
术语解释:
本发明提供的计算网络系统所涉及到的术语具有如下的含义:
标准计算序列:用户通过分布式计算网络所提供的操作系统中的API接口来将所有控制管理的需求及任务转化成为标准计算序列。
标准计算序列包括若干个计算单元,计算单元具体为一个任务或事件,因此标准计算序列的定义包括如下内容:
若干个计算单元间的逻辑流程图;
每个计算单元涉及到的算子和/或算法,输入变量、输出变量,计算流程和/或步骤。
下面具体介绍任务和事件的定义。
任务:是一段由多个计算节点(CPN)协作完成的控制管理策略,由若干个事件按照一定的序列构成。
事件:是计算节点(CPN)完成的一项无中心计算,通过定义算子/算法、输入变量、输出变量、中间变量来确定。
变量:基本空间单元相关的各个物理信息和/或机电设备相关的各个物理信息和/或用户定义的临时变量;例如:区域基本信息中的区域面积,环境及系统测量参数中的室内温度,智能设备运行参数中的照明设备开关状态的反馈,区 域内风机盘管运行状态的设定值,室内环境温度设定值,消防设备的故障报警,过去10分钟内区域总平均用电功率等等。
算法:系统提供的算子(属于算法的下位概念)、基础算法、高级算法或者用户自定义的算法;支持用户将算法下载到计算节点(CPN),支持系统采用类似传输计算的机制,将算法传递给每个相关的计算节点(CPN);可以通过定义算法的有效范围和/或有效时间来删除用户自定义的算法。
下文还会结合具体的算例来对上述概念做出进一步的解释和说明。
标准信息集:为何计算节点(CPN)能自动识别,且即插即用。
运行管理控制策略时涉及到的各个信息点一般具有三个维度的信息,如图11所示,第一个维度解决的是在哪里/是什么,即所述信息点位于哪里(包括所处空间、位置、靠近干路还是支路等带有空间属性的信息)、是什么(是一个房间还是一台水泵或冷机),第二个维度解决的是所述信息点所传递的数据所代表的具体含义(表示温度还是湿度还是能耗等具体内容信息),第三个维度是随着时间的流逝,数据发生的变化等。
通过本发明提供的分布式的计算网络系统,在计算节点(CPN)与DCS结合或通讯的那一刻,DCS自身所带有的空间属性(DCS要么是一个基本空间单元,即一个房间或走廊,要么是一台机电设备,基本空间单元或者机电设备天然具有位置或空间属性)便自然而然地映射在在了CPN当中,CPN组网后形成的网络也就形成了一个带有空间属性的物联网;因而解决了第一个维度的表征问题。
而第二个维度的表征是通过数据标准化实现的:
基本空间单元及各类机电设备所涉及到的所有信息点进行标准化处理,即形成一套标准信息集,标准信息集是由若干个标准数据表组成的,每一个基本空间单元或机电设备都对应一张标准数据表(如表1所示),该标准数据表定义了需要采集哪些变量、变量名称、变量的标准格式等等,所有的基本空间单元或机电设备都按照这个统一的标准信息集来进行描述,这样CPN在与DCS进行数据交互之后便可以获得对应区域或设备的信息,例如,标准信息集中的5号数据是“房间空气温度”,若DCS所连的设备有温度测点,那么DCS的数据库中5号数据便对应该传感器的测量值,从而映射到CPN数据库中的5号数据;而当底层没有测点时,DCS数据库中的5号数据处于默认值状态,映射到CPN中5号数据也处于默认值。即便测控点发生改变,如更换设备、增减测控点时, 只要标准数据表不变,就不会对计算节点(CPN)造成任何影响,采用这种形式,无论底层采用什么设备、什么通讯协议,CPN都能够从DCS处获取信息;也是通过这种方式,当CPN与现场的DCS相连后,通过信息交互CPN便能够识别该DCS所对应的是建筑空间还是某种机电设备,如冷机、或水泵等,从而能够实现CPN的即插即用、自动识别。
表1为所述标准数据表的示例。
表1
Figure PCTCN2016084332-appb-000001
第三个维度的表征是通过DCS或CPN的记录存储功能实现的,能够将所述信息点按照时间流逝的变化发展情况记录下来,不再赘述。
由此可知,本发明提供的分布式的计算网络系统中的计算节点能够即插即 用,智能识别,具有极强的鲁棒性和灵活性。
用户模块的描述:
用户模块具体为一种基于通信协议的API,所述API包括库模块及用户交互模块,即操作系统提供从简单数学计算到专业应用算法的各级算法库,用户在编写标准计算序列时可调用算法库中的算法,操作系统自动形成底层程序代码,从而实现敏捷化编程;结合图6,库模块内置三级算子/算法库,分别为算子库、基础算法库及高级级算法库。
算子库包括:加,减,乘,除,加权求和、积,逻辑运算(求与、或、非等),求最大值、最小值,集合运算(求交集、并集等),生成树,Jacobi/高斯赛德尔迭代,及其他常见的基本数学运算。
基础算法库包括:矩阵计算算法,最速下降法,牛顿法,遗传算法,神经元算法,其他常见的基础数学算法。
高级算法库包括:传感器故障诊断算法,人数分布校核算法,火灾反演算法,基于区域的CFD算法,其他应用于各种专业领域的高级算法。
结合图7,用户交互模块给用户提供如下界面:任务定义界面、事件定义界面、算法定义界面、变量定义界面;
在任务定义界面下,用户可以定义下述各项中的一种或几种:任务名称、任务执行条件、任务步骤总数、任务包括的事件及事件的执行步骤、任务返回结果中包括的计算结构。
在事件定义界面下,用户可以定义下述各项中的一种或几种:事件名称、事件隶属于哪个任务、事件执行条件、事件的输入变量、输出变量、中间变量的名称、事件调用的算子/算法。
在算法定义界面下,用户可以定义下述各项中的一种或几种:算法作用于哪个任务/事件、算法名称、算法的具体内容;
在变量定义界面下,用户可以定义下述各项中的一种或几种:变量名称、变量隶属于哪个任务和/或事件、变量字节长度、变量初值。
可以理解的是上述列举仅为示意性的,没有穷尽也不可能穷尽列举。
用户模块可以采用计算节点为载体,也可以采用区域控制系统为载体,还可以采用另外的独立的软硬件为载体,设置方式灵活多变。
内核模块的描述:
结合图8:所述内核模块进一步包括:
计算处理模块:用于执行一项基本的计算;
并行计算协作服务模块:用于支持和服务于并行计算;令所述系统能够进行多个标准计算序列的并行计算。
通讯模块:用于完成通讯任务。
并行计算协作服务模块进一步包括如下子模块:任务管理模块、事件管理模块、算子/算法管理模块、变量管理模块。
通讯模块进一步包括如下子模块:底层通讯接口驱动模块、通讯协议解析模块、通信帧编辑模块。
模块之间具有如下协同关系:
底层通信接口驱动模块从通信接口处接收信息并将接收到的信息存储于接收缓存;接收缓存位于计算节点(CPN)的内存中。
通信协议解析模块从接收缓存中调取信息并经过解析处理后将信息处理为任务相关信息、事件相关信息、算子/算法相关信息、变量相关信息四个子类并将所述处理后的信息按类别分别存储于任务管理空间、事件管理空间、算子/算法管理空间、变量管理空间,所述任务管理空间、事件管理空间、算子/算法管理空间、变量管理空间均位于计算节点(CPN)的内存中。
任务管理模块对任务管理空间中的任务相关信息进行管理和维护。
事件管理模块对事件管理空间中的事件相关信息进行管理和维护。
算子/算法管理模块对算子/算法管理空间中的算子/算法相关信息进行管理和维护。
变量管理模块对变量管理空间中的变量相关信息进行管理和维护。
计算处理模块从任务管理空间、事件管理空间、算子/算法管理空间及变量管理空间中调用相关信息并进行计算,然后把计算得出的结果再按照类别分别存储于任务管理空间、事件管理空间、算子/算法管理空间、变量管理空间。
通信帧编辑模块从任务管理空间、事件管理空间、算子/算法管理空间、变量管理空间中调出相关信息并编辑成通信帧并存储于发送缓存中,发送缓存位于计算节点的内存中。
底层通信接口驱动模块将发送缓存中的通信帧通过通信接口发送出去。
内核模块与用户模块的通信:
内核模块与用户模块之间在某种通信协议下进行信息交互,某种通信协议可以是现有技术中存在的各种通讯协议,也可以是自定义的通讯协议,例如结 合图10所示,用户模块向内核模块发送的通讯帧可以采用三层结构,底层为物理链路层协议,包括同步码、长度、校验部分及数据部分,这一物理链路层可以采用Ethernet、Wifi、Zigbee等各种成熟的通信技术;第二层为应用底层,包括帧首和数据部分,用于处理通信应答、校验、分片传输等通信问题;第三层为应用上层包括报文类型及报文内容部分,报文内容部分用于处理任务定义、事件定义、算法下载、变量赋值以及系统升级等问题。应用层的结构还可以采用其他的方式。用户模块就是利用图10所示的通信帧的形式将用户定义好的任务、事件、算法、变量中的一种或几种形式发送给内核模块。
算例解释:
本发明提供的分布式计算网络系统上运行的各种运行管理任务及策略均是通过标准计算序列来体现和完成的,当网络完成标准计算序列之后就基本上完成了运行管理的任务及策略,下面结合算例来介绍本系统的计算节点之间是如何协作来完成分布式计算的。
标准计算序列是由若干计算单元组成的,计算单元指的是任务或事件,而任务又是由若干事件按照一定的序列组成的,事件是由一个计算节点完成的一项基本的计算,其通过输入变量、输出变量、中间变量(在某些案例下可能没有)及算子/算法来定义,如图12所示。
算例1:
水泵压差设定值的自动寻优:
问题描述:建筑中空调冷冻水系统中的冷冻水泵根据末端供冷情况进行运行调节,控制要求在满足每个末端冷量需求的前提下,尽量降低冷冻水泵的压差设定值,以降低水泵能耗。
网络结构:如图13所示,该图的左半部分示出了冷冻水系统控制网络,图中水系统末端实为多个支路的网络,图中用一个立管支路简化表示,对应空间节点网络也是多维网络结构,图中以链状结构简化表示,该冷冻水系统采用一次泵设置方式,2台冷机对应3台冷冻泵,冷冻水经分水器流向末端各个房间内的盘管。冷机和水泵各自对应一个设备类型的计算节点,末端盘管隶属于一个空间类型的计算节点,冷机所在的制冷机房设置一个空间类型的计算节点,所述的计算节点之间互联形成一个功能子网,其拓扑结构如图13的右半部分所示。
控制逻辑:如果有末端设备水量不够,则根据不满足要求的末端数量提高 压差设定值,增加水泵转速或加开水泵;在所有末端都满足要求的情况下,如果有末端水量过大,则降低压差设定值,从而降低水泵转速或关闭水泵。
整个问题可以分解为两个相对独立的控制环路,一个是根据末端供应情况调整冷冻水泵两端压差设定值,另一个是水泵组根据压差设定值调节运行的水泵台数和频率使得水泵能耗最低。此处针对第一个控制环路——即如何确定压差设定值的问题来介绍本发明提供的计算网络系统的计算过程。
计算过程:
水泵的压差设定值的自动调节这一任务由三个事件组合完成,如表2所示,其中事件1和2可并行计算,事件1和2的结果作为事件3的输入,整个流程单向进行,无循环,每过一段时间(例如5分钟)执行一次。
表2
Figure PCTCN2016084332-appb-000002
具体过程包括如下步骤:
1、每个末端盘管所隶属的空间节点根据区域内温度的变化情况,判断出本区域的冷量需求是否足够,基本逻辑为:该区域当前的温度测量值在温度设定值的控制精度范围内,且区域内盘管的阀门开度(或占空比)没有完全关闭,则认为当前的冷量可以满足需求;如果当前区域内温度测量值高于温度设定值 的控制精度范围,且区域内盘管的阀门开度或占空比全开,或者在一个很高的阈值内,则认为当前区域内的冷量不能满足需求,发出“太热”信号;如果当前温度测量值低于温度设定值的控制精度范围,且区域内盘管的阀门开度或者占空比全关,或者在一个很低的阈值内,则认为当前区域的冷量超过了实际的冷量需求,发出“太冷”信号;可以定义两个变量,变量1为“是否过热”,是则赋值为1,否则赋值为0,变量2为“是否过冷”,是则赋值为1,否则赋值为0。
2、末端的空间节点形成一个生成树,从第一个计算节点起始,将变量1或2的值传递给邻居节点,邻居节点将邻居传递来的数和本地的数据“局部求和”,再将这个计算结果传递给下一个节点,如此反复,在传递到末端节点时就完成了所有节点的求和;在整个过程中,每个节点都参与了求和计算,但没有一个节点知道整个系统中究竟有多少个节点。统计是否过热的事件1和统计是否过冷的事件2可以并行计算。
3、事件1和事件2计算完成之后,上述两个事件的计算结果传递给任意一个计算节点(可以是冷冻泵所在的计算节点),该计算节点根据如下算法计算新的压差设定值——即事件3。
结合图14。若末端区域的“太热”比例高于5%,则新的压差设定值=原来的压差设定值+1;若末端区域的“太热”比例低于或等于5%但“太冷”比例高于5%,则新的压差设定值=原来的压差设定值-1;若末端区域的“太冷”比例及“太热”比例均低于或等于5%,则压差设定值不变。
在本算例中,“是否太冷”或“是否太热”变量既可以是由计算网络系统提供的系统变量,也可以是用户在编译程序时的自定义变量;所述“局部求和”是由用户模块中的库模块提供的基本算子,可以内置于每个计算节点中,用户可以直接进行调用,不需要再去具体编译实现的代码;所述“生成树”的算法也是由计算网络提供的,可以直接调用;而上述事件1、2、3的具体输入变量、输出变量、先后顺序等可以由用户进行编译。
算例2:
统计人员分布:
在本发明所提供的计算网络系统之上,还可以统计各个区域内的人员分布。在空间相连的区域的门口处安装红外探测器,当有人从A区域离开,进入B区域时,该红外探测器会发送一个人数-1的信号给A区域的计算节点,A区域的计算节点收到信号后会将本区域的人数-1,同时会向邻居计算节点B发起一个 计算,使B区域内的人数+1,由此在所述系统上仿真完成了对应实际过程中人员从A区域进入B区域的物理过程。
所有安装在相通区域连接处的红外探测器,会实时地检测建筑内的人员转移情况,并将信号发送给对应的区域的计算节点,触发整个系统的实时计算,从而可以得到所有区域的人员数目分布。这一信息可以作为基础信息实现其他功能子系统(如空调系统、照明系统)的优化控制。
算例3:
收取所有房间的能耗值:
系统的各个机电设备均可通过DCS或CPN记录自己的能耗情况,或者由DCS或CPN根据机电设备的运行时间,计算出该设备的功耗,因而每个基本空间单元内的所有设备的能耗值是可以由所述系统统计出来的。在物业管理人员的角度,会有将所有区域的能耗值收集起来进行比较分析的需求。
从数学运算的角度来看,收集所有房间的能耗值本质上是一个全局求并集的过程。在实际的计算过程中,由物业管理人员接入的计算节点发起该能耗统计任务,全局求并集,CPN操作的变量为每个基本空间单元的所涉及到设备的总能耗值。任务发起后,由发起任务的计算节点向全局扩散,整个网络系统先形成一个生成树,然后在每个树枝的末端节点将自己的能耗值回传给其邻居节点;之后每个计算节点收到邻居传递过来的能耗值后,将自己所对应区域的能耗值也加进去,重新打包,发给其他邻居(即输入能耗信息邻居节点之外的其它邻居节点);所述计算网络系统中所有的计算节点都遵循该原则,最终所有节点的能耗信息都会打包发送到任务发起者的计算节点处,即完成了全局求并集,收取所有基本空间单元的能耗值这一任务。
本发明提供的用于分布式计算网络的计算节点(CPN)具有如下特点:
1、面向空间:每个计算节点(CPN)均与一个基本空间单元或者某个机电设备的区域控制系统相互关联,在计算节点(CPN)关联到所述区域控制系统时,基本空间单元或机电设备所具有的位置空间信息、相对位置关系或拓扑关系便自然而然地体现在所述计算节点(CPN)上,因而具有快速部署的优点,能够免去原有控制系统的海量而反复的现场布线、适配、调试、定义的工作,节省大量人力;
2、标准化:所述基本空间单元或机电设备的相关信息全部以标准数据表的形式进行描述,计算节点与区域控制系统关联后计算节点能够自动识别出其所 关联的是基本空间单元或者某个机电设备,因而能够实现计算节点(CPN)即插即用,自动识别;
3、无中心计算:由用于分布式计算网络的计算节点(CPN)构成的整个计算网络系统是扁平化的、无中心化的,各个节点的地位完全平等,通过各个节点之间的数据交互分布式地完成全局的计算,在所述系统上运行的各种管理控制策略是通过所述分布式的计算来体现和完成的;
4、快速友好的编程环境:由用于分布式计算网络的计算节点(CPN)构成的系统提供了一个开放、人性化的编程平台,用户利用本系统提供的算子/算法库能够轻松的完成事件/任务的定义,系统自动编译底层程序代码,从而实现控制管理策略迅速软件代码化,具有敏捷开发的优点;在所述编程平台上能够开发出海量的应用程序,具有极大的兼容性和灵活性。

Claims (25)

  1. 一种用于分布式计算网络的计算节点(CPN),其特征在于,所述计算节点(CPN)为具有信息接收、处理、发送功能的计算机,所述计算节点(CPN)具有中央处理器、存储器、通信接口;
    若干个所述计算节点(CPN)构成分布式计算网络;每个计算节点(CPN)均与其拓扑上相邻的计算节点(CPN)进行数据交互;
    所述计算节点(CPN)具有空间属性,所述空间属性体现为所述计算节点(CPN)所在的绝对空间位置和/或所述计算节点(CPN)在所处的拓扑网络中的相对空间位置;
    所述计算节点内置操作系统,所述操作系统提供API接口,用户可通过所述API接口将各种管理/控制的需求和/或策略转化为标准计算序列;
    所述分布式计算网络中的计算节点(CPN)分布式地、自组织地共同完成所述计算序列。
  2. 根据权利要求1所述的计算节点(CPN),其特征在于,所述操作系统支持多个标准计算序列的并行计算。
  3. 根据权利要求1或2所述的计算节点(CPN),其特征在于,在所述分布式计算网络之上可定义多个功能子网,每个所述计算节点(CPN)可隶属于不同的功能子网,所述功能子网之间互不影响。
  4. 根据权利要求1至3任一项所述的计算节点(CPN),其特征在于,所述计算节点(CPN)与某个基本空间单元相关联或者与某个机电设备相关联,所述基本空间单元或机电设备的相关信息均以标准数据表的形式描述,所述标准数据表形成一套标准信息集。
  5. 根据权利要求4所述的计算节点(CPN),其特征在于,所述计算节点(CPN)基于所述标准信息集自动辨识其所关联的具体为某个基本空间单元或某个机电设备,从而实现计算节点(CPN)的即插即用。
  6. 根据权利要求1至5任一项所述的计算节点(CPN),其特征在于,所述操作系统提供从简单数学计算到专业应用算法的各级算法库,用户在编写所述标准计算序列时可调用所述算法库中的算法,所述操作系统自动形成底层程序代码,从而实现敏捷化编程。
  7. 根据权利要求6所述的计算节点(CPN),其特征在于,
    所述算子库包括:加,减,乘,除,加权求和、积,逻辑运算,求最大值、最小值,集合运算,生成树,Jacobi/高斯赛德尔迭代,及其他常见的基本数学运算;
    所述基础算法库包括:矩阵计算算法,最速下降法,牛顿法,遗传算法,神经元算法,其他常见的基础数学算法;
    所述高级算法库包括:传感器故障诊断算法,人数分布校核算法,火灾反演算法,基于区域的CFD算法,其他用于各种专业领域的高级算法。
  8. 根据权利要求1至7任一项所述的计算节点(CPN),其特征在于,所述计算节点(CPN)具体是通过某个区域控制系统与某个基本空间单元或者某个机电设备相关联,所述区域控制系统用于收集所述基本空间单元或机电设备的相关信息或用于控制与所述基本空间单元或机电设备相关的执行器。
  9. 根据权利要求1至8任一项所述的计算节点(CPN),其特征在于,所述计算节点(CPN)具有若干个所述通信接口,所述通信接口分为A类和B类;
    所述计算节点(CPN)通过所述A类通信接口与其拓扑上相邻的计算节点(CPN)进行数据交互;
    所述计算节点(CPN)通过所述B类通信接口与所述区域控制系统(DCS)进行数据交互。
  10. 根据权利要求1至9任一项所述的计算节点(CPN),其特征在于,所述计算节点(CPN)在接入网络时采用局部命名方式,其名称与其拓扑上相邻的计算节点(CPN)不同,拓扑上不相邻的计算节点(CPN)可以具有相同的名称。
  11. 根据权利要求1至10任一项所述的计算节点(CPN),其特征在于,所述API接口具体为一种基于通信协议的API接口或其他形式的普通接口。
  12. 根据权利要求1所述的计算节点(CPN),其特征在于,
    所述标准计算序列包括若干个计算单元,所述标准计算序列的定义包括如下内容:
    所述若干个计算单元间的逻辑流程图;
    每个所述计算单元涉及到的算子和/或算法,输入变量、输出变量,计算流程和/或步骤。
  13. 一种分布式计算网络系统,其特征在于,所述计算网络系统由若干个 计算节点(CPN)构成,所述计算节点(CPN)为具有信息接收、处理、发送功能的计算机;
    每个计算节点(CPN)均与其拓扑上相邻的计算节点(CPN)进行数据交互;
    所述计算节点(CPN)具有空间属性,所述空间属性体现为所述计算节点(CPN)所在的绝对空间位置和/或所述计算节点(CPN)在所处的拓扑网络中的相对空间位置;
    所述分布式计算网络系统内嵌操作系统,所述操作系统或所述操作系统的至少一部分分布于每个所述计算节点(CPN)内;
    所述操作系统提供API接口,用户可通过所述API接口将各种管理/控制的需求和/或策略转化为标准计算序列;
    所述分布式计算网络系统中的计算节点(CPN)分布式地、自组织地共同完成所述计算序列。
  14. 根据权利要求13所述的系统,其特征在于,所述操作系统支持多个标准计算序列的并行计算。
  15. 根据权利要求13或14所述的系统,其特征在于,在所述分布式计算网络系统之上可定义多个功能子网,每个所述计算节点(CPN)可隶属于不同的功能子网,所述功能子网之间互不影响。
  16. 根据权利要求13至15任一项所述的系统,其特征在于,每个所述计算节点(CPN)与某个基本空间单元相关联或者与某个机电设备相关联,所述基本空间单元或机电设备的相关信息均以标准数据表的形式描述,所述标准数据表形成一套标准信息集。
  17. 根据权利要求16所述的系统,其特征在于,所述计算节点(CPN)基于所述标准信息集自动辨识其所关联的具体为某个基本空间单元或某个机电设备,从而实现计算节点(CPN)的即插即用。
  18. 根据权利要求13至17任一项所述的系统,其特征在于,所述操作系统提供从简单数学计算到专业应用算法的各级算法库,用户在编写所述标准计算序列时可调用所述算法库中的算法,所述操作系统自动形成底层程序代码,从而实现敏捷化编程。
  19. 根据权利要求18所述的系统,其特征在于,
    所述算子库包括:加,减,乘,除,加权求和、积,逻辑运算,求最大值、 最小值,集合运算,生成树,Jacobi/高斯赛德尔迭代,及其他常见的基本数学运算;
    所述基础算法库包括:矩阵计算算法,最速下降法,牛顿法,遗传算法,神经元算法,其他常见的基础数学算法;
    所述高级算法库包括:传感器故障诊断算法,人数分布校核算法,火灾反演算法,基于区域的CFD算法,其他用于各种专业领域的高级算法。
  20. 根据权利要求13至19任一项所述的系统,其特征在于,每个所述计算节点(CPN)具体是通过某个区域控制系统(DCS)与某个基本空间单元或者某个机电设备相关联,所述区域控制系统用于收集所述基本空间单元或机电设备的相关信息或用于控制与所述基本空间单元或机电设备相关的执行器。
  21. 根据权利要求13至20任一项所述的系统,其特征在于,每个所述计算节点(CPN)具有若干个所述通信接口,所述通信接口分为A类和B类;
    所述计算节点(CPN)通过所述A类通信接口与其拓扑上相邻的计算节点(CPN)进行数据交互;
    所述计算节点(CPN)通过所述B类通信接口与所述区域控制系统(DCS)进行数据交互。
  22. 根据权利要求13至21任一项所述的系统,其特征在于,所述计算节点(CPN)在接入网络时采用局部命名方式,其名称与其拓扑上相邻的计算节点(CPN)不同,拓扑上不相邻的计算节点(CPN)可以具有相同的名称。
  23. 根据权利要求13至22任一项所述的系统,其特征在于,所述API接口具体为一种基于通信协议的API接口或其他形式的普通接口。
  24. 根据权利要求13所述的系统,其特征在于,
    所述标准计算序列包括若干个计算单元,所述标准计算序列的定义包括如下内容:
    所述若干个计算单元间的逻辑流程图;
    每个所述计算单元涉及到的算子和/或算法,输入变量、输出变量,计算流程和/或步骤;
  25. 根据权利要求13至24任一项所述的系统,其特征在于,每个所述计算节点(CPN)具有相同的结构,嵌于每个所述计算节点(CPN)的操作系统相同。
PCT/CN2016/084332 2015-06-30 2016-06-01 分布式计算网络系统及用于该系统的计算节点 WO2017000738A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
JP2017568390A JP6742353B2 (ja) 2015-06-30 2016-06-01 分布式計算ネットワークシステムおよび当該システムに用いられる計算ノード
EP16817102.3A EP3318938A4 (en) 2015-06-30 2016-06-01 DISTRIBUTED COMPUTER NETWORK SYSTEM AND DATA NODES FOR IT
US15/740,146 US10732588B2 (en) 2015-06-30 2016-06-01 Decentralized computing network system and computing processing node used for the same

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN201510378076.2 2015-06-30
CN201510378076.2A CN106331037B (zh) 2015-06-30 2015-06-30 用于分布式计算网络的计算节点
CN201510377980.1A CN106325229B (zh) 2015-06-30 2015-06-30 分布式计算网络系统
CN201510377980.1 2015-06-30

Publications (1)

Publication Number Publication Date
WO2017000738A1 true WO2017000738A1 (zh) 2017-01-05

Family

ID=57607775

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/084332 WO2017000738A1 (zh) 2015-06-30 2016-06-01 分布式计算网络系统及用于该系统的计算节点

Country Status (4)

Country Link
US (1) US10732588B2 (zh)
EP (1) EP3318938A4 (zh)
JP (1) JP6742353B2 (zh)
WO (1) WO2017000738A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107483541A (zh) * 2017-07-17 2017-12-15 广东工业大学 一种基于滚动时域的在线任务迁移方法

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107992478A (zh) * 2017-11-30 2018-05-04 百度在线网络技术(北京)有限公司 确定热点事件的方法和装置
CN113168160A (zh) * 2018-12-11 2021-07-23 三菱电机株式会社 管理指标计算系统及管理指标计算方法
CN112521094A (zh) * 2020-12-11 2021-03-19 青岛理工大学 装配式建筑用c35免蒸汽养护混凝土及其制备方法
US20220397889A1 (en) * 2021-06-14 2022-12-15 Transportation Ip Holdings, Llc Facility control and communication system

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101102225A (zh) * 2007-07-26 2008-01-09 北京航空航天大学 无线传感器网络节点管理方法
CN201477439U (zh) * 2009-09-02 2010-05-19 沙毅 网络化控制器
CN102130950A (zh) * 2011-03-14 2011-07-20 中国科学技术大学苏州研究院 基于Hadoop集群的分布式监控系统及其监控方法
CN104620183A (zh) * 2012-05-21 2015-05-13 泰塔制造有限责任公司 使用分布式控制模型的自动化和运动控制系统

Family Cites Families (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS62212763A (ja) 1986-03-13 1987-09-18 Fujitsu Ltd 計算機ネツトワ−クにおけるジヨブ実行方式
JP2533282B2 (ja) 1992-05-22 1996-09-11 インターナショナル・ビジネス・マシーンズ・コーポレイション 並列アレイ・プロセッサ
JPH06301655A (ja) 1993-04-14 1994-10-28 Hitachi Ltd 分散処理システム
US6671737B1 (en) * 1999-09-24 2003-12-30 Xerox Corporation Decentralized network system
US7031945B1 (en) * 2000-07-24 2006-04-18 Donner Irah H System and method for reallocating and/or upgrading and/or rewarding tickets, other event admittance means, goods and/or services
JP4500484B2 (ja) 2002-07-19 2010-07-14 株式会社日立製作所 コントローラおよび情報通信サービス方法
US7656822B1 (en) * 2003-12-22 2010-02-02 Sun Microsystems, Inc. Method and apparatus for decentralized device and service description and discovery
DE102004021385A1 (de) 2004-04-30 2005-11-17 Daimlerchrysler Ag Datenkommunikationsnetzwerk mit dezentralem Kommunikationsmanagement
EP1613123B1 (en) * 2004-07-02 2006-06-14 Alcatel Transport network restoration method supporting extra traffic
JP3987512B2 (ja) 2004-07-15 2007-10-10 株式会社東芝 メッセージ配送方法、計算機及びプログラム
US7765307B1 (en) * 2006-02-28 2010-07-27 Symantec Operating Corporation Bulk network transmissions using multiple connections primed to optimize transfer parameters
US7945346B2 (en) * 2006-12-14 2011-05-17 Palo Alto Research Center Incorporated Module identification method and system for path connectivity in modular systems
US20140351010A1 (en) * 2008-11-14 2014-11-27 Thinkeco Power Inc. System and method of democratizing power to create a meta-exchange
GB2486016A (en) 2010-12-02 2012-06-06 Sony Corp Control of storage devices in an electric power network
US20130325525A1 (en) * 2012-05-21 2013-12-05 Boost3, Llc Systems and methods for an integrated online portal and marketplace for event-related items
US8983669B2 (en) * 2012-07-31 2015-03-17 Causam Energy, Inc. System, method, and data packets for messaging for electric power grid elements over a secure internet protocol network
GB201306891D0 (en) * 2013-04-16 2013-05-29 Truphone Ltd International converged mobile services
US20170270157A1 (en) * 2016-03-21 2017-09-21 Virtual Network Element, Inc. TCP/IP Network Automation and Orchestration Tools

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101102225A (zh) * 2007-07-26 2008-01-09 北京航空航天大学 无线传感器网络节点管理方法
CN201477439U (zh) * 2009-09-02 2010-05-19 沙毅 网络化控制器
CN102130950A (zh) * 2011-03-14 2011-07-20 中国科学技术大学苏州研究院 基于Hadoop集群的分布式监控系统及其监控方法
CN104620183A (zh) * 2012-05-21 2015-05-13 泰塔制造有限责任公司 使用分布式控制模型的自动化和运动控制系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3318938A4 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107483541A (zh) * 2017-07-17 2017-12-15 广东工业大学 一种基于滚动时域的在线任务迁移方法

Also Published As

Publication number Publication date
US10732588B2 (en) 2020-08-04
US20180188702A1 (en) 2018-07-05
EP3318938A4 (en) 2019-02-20
JP2018523226A (ja) 2018-08-16
JP6742353B2 (ja) 2020-08-19
EP3318938A1 (en) 2018-05-09

Similar Documents

Publication Publication Date Title
WO2017000738A1 (zh) 分布式计算网络系统及用于该系统的计算节点
CN106325229B (zh) 分布式计算网络系统
US9625885B2 (en) Application-generated function block for data exchange between control programs and building automation objects
Sembroiz et al. Planning and operational energy optimization solutions for smart buildings
US11619410B2 (en) Building HVAC control system, method and wireless mesh device
Taneja et al. Enabling advanced environmental conditioning with a building application stack
CN106331037B (zh) 用于分布式计算网络的计算节点
EP2631857A1 (en) Building automation and control system and method for operating the same
Martirano et al. Building automation and control systems (BACS): a review
CN1658601B (zh) 网关设定工具
CN205864470U (zh) 一种用于分布式计算网络的计算节点及分布式计算网络
US20200209820A1 (en) System and method for improving the energy management of hvac equipment
Sharma Design of Wireless Sensors Network for Building Management Systems
CN113535233B (zh) 用于暖通云边协同的人工智能系统
CN112348347B (zh) 楼宇管理系统及其处理方法、装置、设备
Penya et al. Smart buildings and the smart grid
Soucek et al. Current developments and challenges in building automation
CN105450438A (zh) 多区域无线管理与通讯网络系统及其管理方法
CN101661502B (zh) 利用目标的网关装置
McGibney et al. A systematic engineering tool chain approach for self-organizing building automation systems
Mahdavi A combined product-process model for building systems control
Hakiri et al. A SDN-based IoT architecture framework for efficient energy management in smart buildings
US11892185B1 (en) HVAC system having learning and prediction modeling
Merino-Córdoba et al. Towards concepts for climate and energy-oriented digital twins for buildings
Doellner et al. Towards Concepts for Climate and Energy-Oriented Digital Twins for Buildings

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16817102

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2017568390

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2016817102

Country of ref document: EP