Five-constant environment adaptive regulation and control system and method based on Internet of Things

Abstract

The invention discloses a five-constant environment adaptive regulation and control system and method based on the Internet of Things, particularly relates to the technical field of environment regulation and control of the Internet of Things, and is used for solving the problem that in existing five-constant environment regulation and control based on the Internet of Things, unified time base and complete environment state modeling are lacked. And comfort, stability, reasonable energy consumption and equipment service life cannot be considered under multi-room, multi-disturbance and multi-equipment constraints, and a user still needs frequent manual intervention. The method comprises the following steps: acquiring five constant indexes, occupancy information and an equipment operation state under a unified time service time base, and constructing an environment state vector carrying room purposes and building characteristics; a five-constant demand template, a disturbance type label, an equipment influence response table and a feasible adjustment space are combined to generate a target equipment adjustment vector, and building-level and room-level hierarchical control is implemented, so that five-constant indexes of each room are stably restrained to run around a target interval in different disturbance scenes such as meteorological change, windowing, bathing and gathering; environmental fluctuation is reduced; and the manual intervention frequency of the user is reduced.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) environmental control technology, specifically to an IoT-based adaptive control system and method for five constant environments. Background Technology

[0002] Currently, some IoT-based environmental control systems have emerged in residential and small public buildings for centralized management of indoor temperature, humidity, fresh air volume, and air cleanliness. Some high-end systems have even introduced the concept of a "five constant environments," achieving remote monitoring and scenario-based control by deploying temperature and humidity sensors, carbon dioxide sensors, and communication interfaces in various rooms. Existing solutions typically employ fixed setpoints plus simple time-based strategies, switching different temperature and humidity settings based on weekdays or weekends, and daytime or nighttime, and increasing air conditioning power or fresh air supply when indoor temperature or carbon dioxide concentration exceeds limits. Some solutions attempt to utilize cloud platforms to analyze historical data and provide energy-saving operation suggestions, but these often remain at the level of "alarm plus linkage" or "preset scene switching," lacking a unified control framework for multiple rooms, multiple devices, and multiple disturbance scenarios.

[0003] However, in real-world operation of constant environmental conditions, disturbances such as slow changes in outdoor weather, window ventilation, bathroom use, and large gatherings occur simultaneously. The uses and building characteristics of different rooms vary significantly, and users also have distinct preferences for comfort, noise levels, and energy efficiency. Existing technologies, on the one hand, often lack a unified time synchronization and quality correction mechanism for multi-channel data acquisition, making it difficult to construct an environmental state description that can be directly used for control decisions. They also rarely incorporate building characteristics such as room usage, building size, and building envelope into long-term record management. On the other hand, disturbance identification often relies on a single threshold or empirical rules, lacking the ability to traceably classify weather disturbances, window opening disturbances, bathroom disturbances, and gathering disturbances under a unified parameter table. This makes it difficult for control strategies to finely differentiate between different disturbance scenarios. Furthermore, existing systems generally lack equipment impact response tables and feasible adjustment space constraints based on commissioning and long-term operation records. This makes them insufficient for systematically optimizing and version-based correction of various equipment speed combinations under multiple constraints such as electricity prices, noise control, and equipment start-up and shutdown lifespan. They also lack mechanisms to incorporate frequent user changes to setpoints and proactive equipment shutdowns as negative feedback into strategy adjustments and model updates. Consequently, it is difficult to balance the stability of the five constant indicators, energy consumption levels, and equipment reliability during long-term operation.

[0004] Therefore, in IoT-based five-constant environmental control scenarios, especially for multi-room, multi-disturbance, and multi-constraint operating conditions in residential and small public buildings, there is still a need for an adaptive control technology that can, based on unified time synchronization and unified environmental state description, combine room usage and building characteristics to regularly identify different disturbance types, and form a closed-loop constraint relationship between equipment response models, feasible adjustment space, electricity price noise constraints, and user negative feedback, so as to more stably bring the five constant indicators close to the target range under complex operating conditions, while controlling energy consumption levels and limiting the frequency of equipment start-up and shutdown. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an adaptive control system and method for five constant environments based on the Internet of Things, in order to solve the problems mentioned in the background.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an adaptive control method for five constant environments based on the Internet of Things, comprising: S1. Collect the five constant indicators, occupancy and equipment status, construct the environmental status vector of each room according to the time anchor, and register the room use and building characteristics; S2. Based on the room's purpose, occupancy, and time, obtain the five constant target intervals and weight parameters from the demand template, analyze the state vector changes, and identify and label the disturbance types. S3. Based on debugging and operation data, establish equipment impact response tables and feasible adjustment space for each controlled device, taking into account electricity price, noise and start-stop constraints; S4. During the control cycle, a set of candidate control combinations is generated based on the state deviation, target range and feasible adjustment space. The five constant changes and costs of each combination are predicted using the response table. The equipment control vector that satisfies the constraints and has the lowest cost is selected. S5. Decompose the target adjustment vector into building-level and room-level control commands, send them to centralized equipment and room terminals, call the preset control strategy template according to the disturbance type, and adjust the feasible adjustment space according to the state deviation. S6. After the control ends, record the five constant recovery characteristics, energy consumption and number of start-stop cycles. Record user manual intervention as negative feedback. Periodically adjust the demand template, cost weight and impact response table parameters based on the operation records and negative feedback.

[0007] Furthermore, S1 includes: The control center follows the time synchronization provided by the unified time synchronization device and the acquisition cycle, observation window length, deviation threshold and continuous missing report duration in the parameter table; The five constant indicators, occupancy information and equipment operating status of each room are collected, time-aligned and quality-corrected, and the collection channels that meet the continuous missing reporting time are marked as invalid. At each time synchronization, the corrected data above is combined into an environmental state vector, which is then associated with the room use and building characteristics and stored in the room file. If the key five constant indicators and key equipment status of the room are not recorded within the predetermined time limit, the room will be marked as a monitoring downgraded state, and subsequent adjustments will be limited to a conservative adjustment strategy that does not increase the cooling level, fresh air level, and fan level.

[0008] Furthermore, S2 includes: In each control cycle, the control center reads the room's purpose, building characteristics, current time, and the environmental state vector sequence arranged in chronological order within the observation window from the room archives. Match a template record in the demand template library according to the room use, building characteristics and time period to obtain the five constant target intervals and weight parameters corresponding to temperature, relative humidity, air renewal index, air cleanliness index and noise sound pressure level index. The five constant target intervals, weight parameters, template identifiers, and template version numbers are written into the status record at that time.

[0009] Furthermore, after obtaining the five constant target intervals and weight parameters, the control center uses the observation window length, deviation threshold, number of triggers, disturbance type label retention time, slow change threshold, sudden change time threshold, small change threshold, and hold time threshold configured in the parameter table. The environmental state vector inside the observation window is jointly judged with the five constant target intervals, outdoor meteorological records, and door and window opening and closing status records. States that meet the criteria for weather disturbance, window opening disturbance, bathing disturbance, and gathering disturbance are labeled with the corresponding disturbance type tags; States that do not meet the four disturbance determination conditions are marked as unclassified states, and disturbance type labels are associated with and stored with the corresponding state records.

[0010] Furthermore, S3 includes: Based on the debugging and operation records, the control center establishes an equipment impact response table and feasible adjustment space for each controlled device; The Equipment Impact Response Table records the changes in the five constant indicators and response delay of each piece of equipment within the response time window at different gears and power levels. The feasible adjustment space links each device's gear and power level with energy consumption estimates, noise contribution values, and start-stop penalty factors; The parameter table configures the correction observation window length, deviation revision threshold, and revision trigger count; Compare the predicted changes with the actual changes within the calibration observation window; When the deviation exceeds the deviation revision threshold and the number of times reaches the revision trigger count, update the device impact response table and feasible adjustment space, and register the updated version number in the version index table.

[0011] Furthermore, S4 includes: During the control cycle, the control center reads the environmental state vector, the five constant target intervals, and the weight parameters from the room files. Based on the feasible adjustment space and equipment impact response table, candidate adjustment combinations are generated, including the operating settings of air conditioning indoor units, fresh air terminals, dehumidifiers, local exhaust systems, and terminal fans. Within the control time window, predict the changes in the five constant indicators for each candidate control combination based on the equipment impact response table; The control center eliminates candidate control combinations that do not meet the constraints based on the safety boundary, control improvement threshold, control degradation threshold, power consumption cost weight, noise cost weight, and start-up / shutdown cost weight. Calculate the comprehensive score for the remaining candidate control combinations and select the candidate control combination with the smallest comprehensive score as the target equipment control vector.

[0012] Furthermore, S5 includes: After generating the target equipment adjustment vector, the control center breaks down the target equipment adjustment vector into building-level and room-level control commands according to the control strategy template; Each control command is assigned a device identifier, timing time, policy version number, and idempotency identifier, and registered in the device command management table. The control command is then sent through the communication link. If a valid status code is not obtained within the command round-trip delay limit, the operating status of the corresponding controlled device will be limited to a safe combination of levels within the feasible adjustment space. The control center identifies oscillations based on the oscillation observation window length, oscillation amplitude threshold, overshoot amplitude threshold, gear switching frequency threshold, and boundary crossing frequency threshold, and tightens the feasible adjustment space corresponding to the room according to the gear step size threshold. Only retain adjustment combinations with gear change step sizes not exceeding the gear step size threshold, mark the remaining adjustment actions as unavailable, and register the adjustment convergence protection status in the room file.

[0013] Furthermore, S6 includes: At the end of the control time window, the control center generates a historical record containing the recovery characteristics of the five constants, power consumption statistics, and equipment start-up and shutdown statistics based on the environmental state vector and the five constant target intervals. Within the time range corresponding to the length of the negative feedback observation window, temperature setting changes, humidity setting changes, and equipment shutdown behaviors are compared with the threshold values ​​for the number of setting changes and the threshold values ​​for the number of equipment shutdowns to generate negative feedback records. The control center compares historical records and negative feedback records with the recovery time upper limit threshold, overshoot amplitude threshold, and negative feedback frequency threshold; In the room archive, mark the rooms that meet the judgment criteria as rooms that require strategy optimization; Adjust the five constant target ranges and cost weights in the room demand template; Update the typical change records in the Equipment Impact Response Table.

[0014] On the other hand, the present invention provides an adaptive control system for five constant environments based on the Internet of Things, comprising: The environmental status construction module is used to collect five constant indicators, room occupancy information and equipment operating status. Based on the time anchor, the collected data is time-aligned, an environmental status vector of each room is constructed, and the room use and building characteristics are registered. The demand and disturbance identification module is used to obtain the corresponding five constant target intervals and weight parameters from the demand template based on the room use, occupancy information and current time, analyze the changing characteristics of the environmental state vector, and identify and mark the current disturbance type. The equipment impact modeling module is used to establish an impact response table of various terminal equipment on the five constant indicators based on debugging data and operation data, determine the feasible adjustment space of each equipment under the current operating conditions, and introduce electricity price constraints, noise constraints and start-stop constraints in the modeling process. The adjustment combination solution module is used to generate candidate adjustment combinations based on environmental state deviation, five constant target ranges and feasible adjustment space within the control cycle. It uses the influence response table to predict the changes in five constant indicators and control costs caused by each candidate combination, and selects the target equipment adjustment vector based on the control costs. The hierarchical execution module is used to split the target device adjustment vector into building-level control commands and room-level control commands, which are then sent to the central device and the room terminal device, respectively. It also calls the preset control strategy template according to the disturbance type and adjusts the feasible adjustment space according to the environmental state deviation during the execution process. The self-learning correction module is used to record the recovery characteristics of the five constant indicators, energy consumption data, and the number of equipment start-ups and shutdowns after the control ends. It registers user manual interventions as negative feedback information and periodically corrects the demand template, control cost weight, and impact response table parameters based on the operation records and negative feedback information.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. By collecting five constant indicators, occupancy information and equipment operating status under a unified time base, an environmental state vector carrying room use and building characteristics is constructed. Combined with the five constant demand template, disturbance type label, equipment impact response table and feasible adjustment space, a target equipment adjustment vector is generated and a building-level and room-level hierarchical control is implemented. This ensures that the five constant indicators of each room operate around the target range under different disturbance scenarios such as weather changes, window opening, bathing, and gatherings, thereby reducing environmental fluctuations and the number of times users need to manually intervene.

[0016] 2. By combining the negative feedback generated by monitoring degraded status, safety mode, and user manual intervention with the five constant recovery characteristics, energy consumption statistics, and equipment start-up and shutdown times, the demand template, cost weight, and equipment impact response table are versioned and self-learned for correction. In addition, with the cooperation of idempotent control command management, oscillation identification, and adjustment convergence protection mechanisms, the control strategy and equipment model are continuously optimized under the premise of ensuring operational safety and equipment lifespan, thereby improving the energy efficiency, operational stability, and full-process traceability of the five constant environmental control. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the adaptive control method for five constant environments based on the Internet of Things according to the present invention. Figure 2 This is a schematic diagram of the five constant environment adaptive control system based on the Internet of Things of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: Figure 1 A flowchart illustrating the IoT-based adaptive control method for five constant environments is provided. The IoT-based adaptive control method for five constant environments includes: S1. Collect the five constant indicators, occupancy and equipment status, construct the environmental status vector of each room according to the time anchor, and register the room use and building characteristics; S2. Based on the room's purpose, occupancy, and time, obtain the five constant target intervals and weight parameters from the demand template, analyze the state vector changes, and identify and label the disturbance types. S3. Based on debugging and operation data, establish equipment impact response tables and feasible adjustment space for each controlled device, taking into account electricity price, noise and start-stop constraints; S4. During the control cycle, a set of candidate control combinations is generated based on the state deviation, target range and feasible adjustment space. The five constant changes and costs of each combination are predicted using the response table. The equipment control vector that satisfies the constraints and has the lowest cost is selected. S5. Decompose the target adjustment vector into building-level and room-level control commands, send them to centralized equipment and room terminals, call the preset control strategy template according to the disturbance type, and adjust the feasible adjustment space according to the state deviation. S6. After the control ends, record the five constant recovery characteristics, energy consumption and number of start-stop cycles. Record user manual intervention as negative feedback. Periodically adjust the demand template, cost weight and impact response table parameters based on the operation records and negative feedback.

[0020] The technical connections and implementation logic of the six steps are as follows: In step S1, the system first collects the five constant indicators, occupancy information, and equipment operating status of each room under unified time constraints, constructs an environmental state vector with room usage and building characteristic labels, and stores it in the room archive to provide a complete state baseline for subsequent analysis. In step S2, the control center matches the corresponding five constant target intervals and weight parameters from the demand template library based on room usage, occupancy status, and current time, and identifies and labels disturbance types such as meteorological disturbances, window opening disturbances, bathing disturbances, and gathering disturbances by combining the change process of the environmental state vector within the observation window, thereby transforming the original state sequence into control requirements with target and disturbance labels. In step S3, the control center uses debugging records and long-term operation records to establish equipment impact response tables and feasible adjustment spaces for each controlled equipment, and associates equipment levels with changes in the five constant indicators, electricity price, noise, and start-stop constraints, providing a set of feasible actions and their cost models for subsequent combinatorial optimization. In step S4, the control center reads the environmental state vector and the five constant indicators in each control cycle. The target range, weight parameters, feasible adjustment space, and equipment impact response table are used to generate candidate adjustment combinations. The changes in the five constants and the comprehensive cost of each candidate combination within the control time window are predicted. Under the premise of meeting the safety boundary and control improvement constraints, the target equipment adjustment vector with the minimum comprehensive cost is selected. In step S5, the control center splits the target equipment adjustment vector into building-level control commands and room-level control commands according to the equipment level. Combined with the disturbance type, the corresponding control strategy template is called and issued to the centralized equipment and room end. During the execution process, the feasible adjustment space is tightened or loosened according to the oscillation and deviation to keep the field control stable. In step S6, after the end of each control time window, the control center records the five constants recovery characteristics, energy consumption statistics, and equipment start-up and shutdown times. User manual intervention behavior is registered as negative feedback. During the analysis period, the five constants target range, cost weight, and equipment impact response table parameters in the demand template are corrected accordingly. This makes steps S1 to S5 form a closed-loop adaptive control logic based on continuous correction of operation records and negative feedback.

[0021] S1. Collect five constant indicators, occupancy and equipment status, construct environmental status vectors for each room according to the time synchronization anchor, and register the room's purpose and building characteristics. The specific implementation is as follows: The system is deployed in residential or small public buildings, equipped with environmental sensing and status recording capabilities for each physical room. It continuously acquires five constant indicators (temperature, humidity, air quality, and temperature), occupancy information, and the operating status of related equipment within a predetermined time resolution and accuracy range, and constructs an environmental status vector accordingly. In this solution, the five constant indicators are defined as a set of indoor temperature, indoor relative humidity, airborne carbon dioxide or oxygen concentration, cleanliness indicators represented by airborne particulate matter concentration, and noise level. The key five constant indicators are limited to temperature and relative humidity, used to determine whether the room meets the conditions for implementing fine-grained control strategies.

[0022] The carbon dioxide concentration or oxygen content in the air is used to reflect the degree of air renewal. The system preferably configures at least one of these indicators. Whether carbon dioxide concentration, oxygen content, or both are used is determined by the configuration items in the parameter dictionary, which records the selected indicator type and corresponding unit. Occupancy information refers to the current room occupancy status identified by a human presence detection device, including at least whether there are people present and the preset number of people within that range.

[0023] The equipment operating status includes the current on / off status and speed setting of air conditioners, fresh air units, dehumidifiers, exhaust systems, and terminal fans. The critical equipment status is limited to the operating status of air conditioners and fresh air units, used to determine whether controls involving cooling, heating, and fresh air volume adjustments are permitted. The environmental state vector is a record composed of five constant indicators, occupancy information, and equipment operating status combined in a fixed field order at the same time interval, reflecting the comprehensive environmental state of the room at that time. Room usage refers to the pre-registered functional category for each room, indicating its primary use; preferably, it may include types such as bedroom, living room, kitchen, bathroom, and study.

[0024] Building characteristics are a set of parameters related to the thermal and humidity inertia of the room. Preferably, these include room volume, the main material type of the building envelope, room orientation, and the window-to-wall area ratio. They can also be expanded to include indicators such as insulation layer construction and the heat transfer performance rating of external windows, depending on project requirements. During the design phase, the system uniformly registers the names, meanings, units, and value ranges of the above-mentioned terms in the parameter dictionary and records the effective date in the configuration version to ensure consistency in terminology throughout subsequent stages.

[0025] To achieve the above capabilities, the system is equipped with temperature sensors, humidity sensors, carbon dioxide or oxygen concentration detectors, particulate matter concentration detectors, noise detectors, human presence detectors, and status acquisition devices for collecting data on the operating status of air conditioners, fresh air units, dehumidifiers, exhaust systems, and terminal fans in each room. During installation and commissioning, the measurement units, ranges, and allowable measurement errors of each acquisition channel are uniformly agreed upon. Preferably, temperature can be set to degrees Celsius, humidity can be set to a percentage, carbon dioxide concentration can be set to a volume fraction or mass concentration, and noise can be set to a sound pressure level. The default values ​​for the above ranges and allowable errors are registered in the parameter dictionary and can be adjusted with authorization according to specific projects.

[0026] The system initiates data collection for each room one by one according to the preset collection cycle. The collection cycle and minimum observation time step can be set to fall within a range of several seconds to several minutes, and are recorded in the parameter table as configuration items. The specific values ​​can be determined and written into the parameter table during the engineering design stage according to the building scale, network conditions and control cycle requirements. The parameter table also records the observation window length, which is used for subsequent judgment of data quality.

[0027] Each acquired record is accompanied by a time synchronization provided by a unified time synchronization device. After receiving records from multiple sensor channels, the control center performs time alignment on the multi-source records based on the time synchronization. Records with time synchronization times within the same time interval are grouped into the same time window. In the case of multiple duplicate records in the same room within the same time window, preferably, one record can be selected as a valid record according to the order of record arrival, signal quality mark, or integrity mark, and the selection rule is recorded in the log.

[0028] To ensure that the judgment criteria for handling abnormal measurements are unique, the parameter table is pre-configured with deviation thresholds, continuous missing report durations, and the number of adjacent observation windows used to calculate statistical values. The deviation threshold is the upper limit of the allowable deviation of a certain indicator relative to the statistical value within the adjacent observation windows. The continuous missing report duration is the maximum length of time that a channel is allowed to have missing records within a certain number of consecutive acquisition cycles. The number of adjacent observation windows is used to limit the number of environmental state vectors participating in the calculation of statistical values. The parameter table also configures the statistical type as either median or mean, and locks it through configuration version. The statistical value is calculated using the median or mean within the adjacent observation windows according to the statistical type registered in the parameter table.

[0029] For isolated measurements that exceed the deviation threshold within a short period of time, the control center replaces the measurement with the median or mean of the environmental state vector in the corresponding adjacent observation window, according to the statistical type and the number of adjacent observation windows registered in the parameter table, and internally marks this correction behavior. For cases where some channel records are missing in a single acquisition cycle but do not exceed the continuous missing reporting time, the system uses the valid records in the previous observation window to fill in the missing channel value and marks this value as the filler value in the environmental state vector. When a channel still has no valid records after the continuous missing reporting time exceeds the continuous missing reporting time, the control center marks the channel as in a failed state and accordingly restricts the subsequent control strategy for the corresponding room.

[0030] After completing the above time alignment and quality correction, the control center combines the temperature, relative humidity, carbon dioxide concentration or oxygen content, particulate matter concentration, noise pressure level, human presence status, and various equipment operating status of the same room into an environmental state vector in each time synchronization window according to the pre-defined field order. The control center then associates this environmental state vector with the pre-registered room use and building characteristics according to the room identifier and adds it to the room file maintained by the control center.

[0031] Room records are preferably organized chronologically, enabling retrieval of corresponding environmental state vector sequences based on room identifiers and time ranges. These sequences are then sequentially accessed for subsequent demand identification, equipment modeling, and adjustment decision-making. Each data acquisition device communicates with the control center via fieldbus or local area network. The communication protocol is uniformly agreed upon before system commissioning. The round-trip delay for a single request and response is configured with a predetermined time upper limit in the parameter table. This predetermined time upper limit can be set to fall within a range of several seconds to tens of seconds. The specific value can be determined and written into the parameter table during the engineering design phase based on building scale, network conditions, and real-time requirements.

[0032] After initiating a data acquisition request, if the control center does not receive a record from a certain channel within the predetermined time limit, it can automatically initiate a retry according to the number of retry attempts and retry intervals recorded in the parameter table. If the key five constant indicators or key equipment status of the room cannot be obtained after the specified number of retry attempts, the control center marks the room's operating status as a monitoring downgraded state. The monitoring downgraded state is limited in the parameter dictionary to a state in which only conservative adjustment strategies are allowed in subsequent control links. The conservative adjustment strategy is limited to not increasing the air conditioning cooling power, not increasing the fresh air volume, and not adding a setting that would cause the noise sound pressure level to exceed the noise control limit in the parameter table. Only the control combination of maintaining or reducing the existing cooling, fresh air, and fan settings is allowed.

[0033] Upon entering the monitoring degradation state, the control center records the room identifier, occurrence time, failed channel identifier, current configuration version number, and a brief description of the reason in the operation and maintenance log. This information is used by operation and maintenance personnel to investigate and form a chain of evidence for configuration version adjustments. Through the above configuration, this invention clearly discloses, at the level of capability and scope, the object scope, time rhythm, judgment rules, and boundary conditions of multi-channel acquisition, time alignment, quality correction, and environmental state vector construction in a five-constant environmental control scenario. This enables those skilled in the art to perform equivalent implementations under different building scales, sensor models, and network conditions, provided that the parameter tables and parameter dictionary configurations are known, without changing the technical essence of this invention.

[0034] S2. Based on room usage, occupancy, and time, obtain the five constant target intervals and weight parameters from the demand template, analyze the state vector changes, identify and label the disturbance types, specifically as follows: After completing the construction of the environmental state vector, the control center reads the room's purpose, current occupancy status, current time of the room, and the environmental state vector sequence stored in chronological order within the preset observation window length from the room file for each room's control cycle at each time interval. Based on this, it completes the matching of the five constant requirements template and the identification of disturbance types.

[0035] The requirement template library is a set of pre-established configuration records categorized according to different room uses, time periods, and optional building characteristics. Each template record is registered in the template table with a unique template identifier and includes at least the target range, weight parameters, and priority information for temperature, relative humidity, carbon dioxide concentration or oxygen content (indicating air renewal), air cleanliness, and noise pressure level. The target range is preferably given in the form of upper and lower limits. The weight parameters are used to characterize the importance of each indicator in the comprehensive evaluation under the current scenario. The priority is used to determine the order of indicators to be satisfied when there are constraint conflicts. The template table also records the unit, default target range, and allowable fluctuation range of each of the above indicators and marks the effective version of the template set with the template version number.

[0036] The control center searches for template records that match the room's purpose and time period in the demand template library based on the room's registered use, building characteristics, and the current time period. Preferably, the time period can be divided into preset intervals such as early morning, daytime, evening, and nighttime. The specific division method and template matching rules are uniformly registered in the parameter dictionary and template table. When there are multiple candidate templates, one can be selected as the target template for the current period according to the building characteristics or the user-defined priority rules. The five constant target intervals and weight parameters recorded in it are used as the target control requirements for the room in this period.

[0037] To ensure a unified standard for determining disturbance types, the parameter table pre-configures parameters for each room, including observation window length, deviation threshold, trigger count, disturbance type label retention time, slow change threshold, abrupt change time threshold, small change threshold, and hold time threshold. The observation window length limits the time range for analyzing environmental state changes, preferably configured in units of several acquisition cycles. The deviation threshold is the upper limit of permissible deviation for each of the five constant indicators relative to its target interval or relative to the statistical values ​​within adjacent observation windows. The statistical values ​​are calculated using the median or mean within adjacent observation windows, based on the statistical type and the number of adjacent observation windows registered in the parameter table. The following parameters are calculated: the trigger count is the threshold number of times a deviation behavior needs to be accumulated within the observation window; the disturbance type label retention time is the shortest time that the label must remain valid in subsequent control cycles once it is determined to be a certain disturbance type; the slow change threshold is used to limit the situation where the rate of change of outdoor meteorological quantities per unit time is lower than the threshold and the cumulative change within the observation window exceeds the deviation threshold; the sudden change time threshold is used to limit the maximum duration of significant changes in indoor indicators; the small change threshold is used to limit the range of fluctuations of indicators that need to be determined to have small changes; and the hold time threshold is used to limit the shortest time that a deviation state of an indicator needs to be maintained.

[0038] The occupancy information is preferably divided into several number intervals in terms of the number of people. For example, number interval 1 corresponds to zero or a small number of people, number interval 2 corresponds to a medium number of people, and number interval 3 corresponds to a large number of people. The lower limit and upper limit of the number of people corresponding to each number interval are recorded in the parameter dictionary. In this scheme, the small number of people interval is limited to number interval 1, and the large number of people interval is limited to number interval 3.

[0039] The "Room Usage" field is used to characterize the room's functional category. Bathroom-type rooms are limited to those marked as "toilet" or "shower room" in the "Room Usage" field, while bedroom or living room-type rooms are limited to those marked as "bedroom," "living room," or "family room" in the "Room Usage" field. The meteorological module is used to acquire outdoor meteorological records corresponding to the building's location. Preferably, this can be achieved by accessing data from a meteorological service platform or by deploying local outdoor temperature, humidity, and wind speed detection devices. Outdoor temperature, outdoor relative humidity, outdoor wind speed, and outdoor weather type are written into the meteorological records according to the time of the data transmission.

[0040] Within each control cycle, the control center extracts the environmental state vector sequence of the room within the current observation window length from the room archive. It compares the current values ​​of the five constant indicators at each time interval with the target interval recorded in the corresponding template. It then combines this with deviation threshold statistics to determine the magnitude of deviation from the target interval and the duration of continuous deviations. Finally, it integrates outdoor weather records provided by the meteorological module, door and window opening / closing status records, and changes in the operational status of relevant equipment to comprehensively assess the trend of environmental state changes. Door and window opening / closing status is preferably provided by door and window magnetic detection devices, and their opening / closing status records are aligned with the environmental state vector as an additional status field in the room archive, according to the time interval.

[0041] According to the preset judgment rules in the parameter table, when the rate of change of outdoor temperature or outdoor relative humidity within a unit time is less than the slow change threshold and the cumulative change within the entire observation window exceeds the deviation threshold, and the direction of indoor temperature and relative humidity deviation from the target range is consistent with the outdoor change direction, the deviation magnitude exceeds the deviation threshold, and the number of consecutive deviations reaches the trigger count, the control center classifies the current state as a meteorological disturbance; when the door / window opening / closing status changes from closed to open within the observation window, and at least one of the indoor temperature, relative humidity, or noise sound pressure level changes relative to the target range by a magnitude exceeding the deviation threshold within a time period not exceeding the sudden change time threshold, the current state is classified as a window opening disturbance; when the room usage field indicates that the room is a bathroom room, and the relative humidity within the observation window increases by a magnitude exceeding the deviation threshold within a time period not exceeding the sudden change time threshold, and at least... If the concentration of particulate matter and the noise level remain above the upper limit of the target range for a specified duration, and the fluctuations in these values ​​within the observation window are both below the threshold for smaller changes, while the ventilation system's operating status changes from off to on or its speed increases within the observation window, the control center classifies the current state as a bathing disturbance. If the room usage field indicates that the room is a bedroom or living room, and the occupancy information within the observation window changes from an empty or low-occupancy range to a high-occupancy range, and the carbon dioxide concentration or oxygen content deviates from the target range within the observation window by more than the deviation threshold and the number of consecutive deviations reaches the trigger count, and the noise level remains above the upper limit of the template record's allowable fluctuation range for at least the specified duration, the control center classifies the current state as a gathering disturbance. If none of the above disturbance type determination conditions are met, the system marks the current state as a pending classification state.

[0042] The judgment rules for various disturbance types are given in the parameter table in the form of condition combinations, including the deviation direction, deviation magnitude, duration, door and window opening and closing status change pattern, outdoor weather change trend, and equipment status change combination constraints. The parameter table configures a set of rules for each disturbance type and locks it through the parameter table version number to ensure that the judgment rules used in different operation stages are traceable.

[0043] If a room is already in a downgraded monitoring state, the control center can still perform template matching of the five constant target intervals and weight parameters. However, it is preferable to restrict the disturbance judgment rules that rely on the failed channel data in the switch options configured for this state in the parameter table, so that the disturbance type label remains in the unclassified state when key five constant indicators or key equipment status are missing. Alternatively, according to the configuration item in the parameter table regarding whether to uniformly downgrade to meteorological disturbance when key information is missing, the current state can be directly marked as meteorological disturbance.

[0044] After completing the matching of the five constant requirements template and the identification of disturbance types, the control center generates or updates a status record in the room file for the current time. It associates and stores the five constant target intervals, weight parameters and disturbance type labels used in this cycle with the corresponding environmental state vectors. The status record also includes the template identifier, template version number and parameter table version number referenced in this cycle. This allows the subsequent adjustment combination solution to directly read this information as constraints and weight sources, while providing a complete version context for subsequent operation analysis and evidence chain tracing.

[0045] Through the above configuration, this invention clearly discloses, at the level of capability and scope, the organization method, matching rules, disturbance type classification criteria, configuration range and version management method of the five constant requirements template, enabling those skilled in the art to adapt and equivalently implement the template library and disturbance judgment rules according to different building uses, climate conditions and user preferences, given the known room file structure, environmental state vector criteria and parameter table configuration, without changing the subsequent adaptive control approach of this invention based on the five constant requirements template and disturbance type label driven by the present invention.

[0046] S3. Based on commissioning and operational data, establish equipment impact response tables and feasible adjustment spaces for each controlled device, considering electricity price, noise, and start-up / shutdown constraints. The specific implementation is as follows: After completing the matching of the five constant requirements template and the identification of disturbance types, the control center, based on the trial operation records from the early commissioning phase and long-term operation records, establishes an equipment impact response table and feasible adjustment space for various controlled equipment within the building. This supports the subsequent optimization of adjustment combinations and tiered execution. The controlled equipment preferably includes room-level air conditioning indoor units, fresh air terminals, dehumidifiers, local exhaust ventilation systems, and terminal fans, as well as building-level fresh air handling units, cooling source equipment, and main exhaust ventilation systems, collectively referred to as controlled equipment.

[0047] The Equipment Impact Response Table is used to depict the typical effects of changes in the gear or power level of controlled equipment on the five constant indicators. Each record is grouped according to equipment type, equipment identification, equipment gear or power level, room or area where the equipment is located, room purpose, and building characteristics. Within a specified response time window, it records the typical changes in indoor temperature, relative humidity, carbon dioxide concentration or oxygen content (indicating air renewal), air cleanliness, and noise pressure level after the equipment is adjusted from one gear or power level to another, as well as the response delay time from when the control center issues the adjustment command to when the above changes begin to appear.

[0048] The response time window is preferably configured in the parameter table in units of several control cycles or several acquisition cycles. Different devices can be set with different response time window lengths to adapt to different thermal inertia and airflow organization characteristics. The typical change can be set as the statistical result of the difference between the corresponding index in the environmental state vector within the response time window and the baseline state before adjustment. Preferably, it is registered based on the statistical values ​​of multiple sets of samples collected under the same adjustment conditions to reduce the impact of accidental disturbances.

[0049] For equipment sensitive to seasonality and outdoor weather conditions, the impact response table should preferably be established by season or by outdoor temperature and humidity range, with the relevant seasonal and meteorological range division methods clearly defined in the parameter dictionary. The feasible adjustment space is used to limit the range of settings or power that each controlled device can use within the current operating cycle, and to provide energy consumption and noise cost information for subsequent adjustment combination scoring.

[0050] For each controlled device, the control center, taking into account factors such as electricity price range, noise limits, maximum number of device start-stop cycles, and rated device capacity, provides a set of permissible power levels or power ranges within the feasible adjustment space. The corresponding records include the estimated energy consumption, noise contribution, and start-stop penalty factor for that power level or power range within a unit control time window.

[0051] Electricity price information is preferably provided by the building energy management system by time period, recorded as electricity price ranges corresponding to different time periods. The control center generates corresponding electricity price weights in the parameter table based on this. Noise limits can be formed according to contractual agreements, specification requirements, or owner settings. The parameter table records the noise limit and the allowable range and duration of exceeding the limit within a short period of time. The upper limit of equipment start-stop frequency is preferably determined comprehensively based on the recommended start-stop frequency, maintenance cycle, and historical fault statistics given in the equipment manual. The maximum allowable start-stop frequency or start-stop number is registered in the parameter table on a daily, weekly, or monthly basis. The start-stop penalty factor is used to reflect the impact of frequent start-stop on equipment lifespan and maintenance costs, and can be set to different penalty levels corresponding to different start-stop modes.

[0052] To ensure the executability and traceability of the equipment impact response table and feasible adjustment space, the parameter table pre-configures parameters such as the model calibration observation window length, meteorological stability threshold, response deviation revision threshold, model revision trigger count, and upper limit of the number of trial operation equipment. The model calibration observation window length defines the time window for assessing the deviation between predicted and actual changes, preferably recorded as several control cycles or a fixed time length. The meteorological stability threshold defines the upper limit of the rate of change of outdoor temperature and outdoor relative humidity per unit time. When the rate of change of outdoor temperature and outdoor relative humidity is lower than the meteorological stability threshold within a certain time period, that time period is considered a meteorological stable condition and can be used for equipment trial operation modeling. The response deviation revision threshold defines the maximum allowable deviation range between predicted and actual changes. The model revision trigger count defines the number of times the deviation exceeds the response deviation revision threshold within the model calibration observation window before triggering the model revision process. The upper limit of the number of trial operation equipment limits the number of devices allowed to participate in the stepped adjustment simultaneously in the same trial operation scenario.

[0053] When constructing the initial equipment impact response table, the control center preferably performs step-by-step gear or power adjustments on individual devices or devices whose total number does not exceed the upper limit of the number of devices in trial operation within a time period that meets the meteorological stability threshold conditions, while keeping the operating status of other controlled devices and the status of room doors and windows stable. By comparing multiple sets of environmental state vectors before and after adjustment, the typical changes and response delays corresponding to each adjustment action are extracted within the predetermined response time window to form the initial impact response record.

[0054] During long-term operation, under normal business or residential conditions, the control center associates the control decision records generated in each control cycle with the subsequent environmental state vector sequence. Samples are extracted according to the model calibration observation window length. The changes in the five constant indicators under each adjustment action are predicted using the current version of the equipment impact response table and compared with the actual changes. When the deviation between the predicted and actual changes is continuously observed to exceed the response deviation revision threshold within the model calibration observation window and the number of deviations reaches the model revision trigger number, the control center selects samples from the most recent continuous time range within the predetermined maintenance window and adjusts the entries in the impact response table of the relevant equipment. If necessary, the estimated power consumption and noise contribution values ​​of the corresponding gear or power range in the feasible adjustment space are adjusted simultaneously. The updated entries are stored in the version index table with the new version number, and the old version records are retained in the version index table for retrospective purposes.

[0055] The aforementioned tolerance ranges are preferably set according to the five constant indicators, consistent with the aforementioned deviation threshold. The maximum allowable deviation range of the predicted value from the actual value and the statistical judgment method used are uniformly registered in the parameter table. The equipment impact response table and feasible adjustment space are stored in the control center as records with version numbers. The version number is associated with the generation time, effective time, applicable equipment scope, and source description in the version index table. Each version update is recorded in the evidence chain log as a summary of the differences between the records before and after the modification, the modification time, the operator's identifier, and the triggering reason.

[0056] Subsequently, in the adjustment combination solution and hierarchical execution phase, the control center directly references the currently effective version of the equipment impact response table and feasible adjustment space. Based on the recorded typical changes, estimated power consumption, noise contribution, and start-stop penalty factors, the candidate adjustment combinations are predicted and scored. The version numbers of the equipment impact response table and feasible adjustment space involved in the decision-making are archived together with the control decision records generated in each control cycle, so that the equipment model and constraints used behind each control decision can be traced back to the record content at that time through the version number chain.

[0057] In some implementations, when a certain type of controlled equipment has a short operating time or has not yet accumulated enough samples to form a local response record for the building, it is preferable to use the performance curve provided by the equipment manufacturer or the experience response table of the same type of building as the initial version. After running for a period of time, it can be gradually replaced with the local version of the building according to the above-mentioned model correction and version update mechanism. As long as the equipment type, gear or power level, the change of the five constant indicators and the definition of the response time window are consistent, it is considered to be an equivalent implementation of the present invention.

[0058] Through the above configuration, this invention fully discloses the object scope, time scale, parameter source, version management, and adaptive revision rules for equipment impact response modeling and feasible adjustment space limitation. This enables those skilled in the art to construct equipment behavior models and constraint sets adapted to specific equipment types and building characteristics based on existing environmental state vectors, five constant requirements templates, and disturbance type labels. This provides reliable support for subsequent optimization of adjustment combinations based on five constant objectives and cost weights.

[0059] S4. Within the control cycle, a set of candidate control combinations is generated based on the state deviation, target range, and feasible adjustment space. The five constant changes and costs of each combination are predicted using the response table. The equipment control vector that satisfies the constraints and has the lowest cost is selected. The specific implementation is as follows: At the beginning of each control cycle, the control center reads the environmental state vector corresponding to the current time from the room file, as well as the five constant target intervals, weight parameters, and disturbance type labels written in the aforementioned five constant requirements template matching stage, for rooms that are in normal monitoring or monitoring degraded state. Combined with the feasible adjustment space and equipment impact response table, the control center generates target equipment adjustment vectors for each room within the preset control time scale.

[0060] The control cycle length is used to limit the refresh rhythm of control commands, and is preferably configured in the parameter table in units of several seconds to several minutes; the control time window is used to limit the time range for predicting and evaluating the effect of this adjustment, and can preferably be set to one control cycle or an integer multiple of several control cycles. The specific value is registered in the parameter table and locked by the configuration version number.

[0061] The feasible adjustment space is the set of allowable gears or power ranges formed for the controlled equipment in the previous stage. In the current control cycle, the control center first filters the available gears or power ranges of the controlled equipment in each room based on the feasible adjustment space, room use and disturbance type labels, and removes gears or power ranges that are marked as unusable under the current electricity price range, noise limit and start / stop number constraints.

[0062] Based on this, the control center, according to the candidate combination generation rules recorded in the parameter table, combines the air conditioning speed, fresh air volume level, dehumidification level, exhaust speed, and terminal fan speed or speed within the feasible adjustment space according to a preset step size, forming a set of candidate adjustment combinations. In this scheme, the candidate adjustment combination is defined as an ordered set of equipment speeds or power levels to be used in a certain room within the current control time window, including at least the air conditioning indoor unit operating speed, fresh air terminal supply air volume level, dehumidification device operating speed, local exhaust device operating speed, and terminal fan speed or speed. If necessary, it can be expanded to include the adjustment of building-grade fresh air units and cold source equipment.

[0063] The upper limit of the number of candidate combinations is configured in the parameter table in the form of a positive integer. It is used to limit the number of combinations that can participate in the scoring in each control cycle. When the control center generates candidate adjustment combinations, it stops expanding when the number of combinations reaches the upper limit. Preferably, it can combine the deviation degree of the current five constant indicators and the disturbance type label to pre-screen combinations that are obviously unlikely to bring effective improvement, so as to balance the combination coverage and the consumption of computing resources within the constraint of the upper limit of the number of candidate combinations.

[0064] For each candidate adjustment combination, the control center infers the trend of change of the combination on temperature, relative humidity, carbon dioxide concentration or oxygen content, air cleanliness and noise pressure level within the control time window length based on the typical changes and response delays registered in the equipment impact response table. Preferably, the response time window is aligned with the control time window or nested in integer multiples, and the correspondence between the two is recorded in the parameter table.

[0065] After obtaining the predicted values ​​of the five constant indicators, the control center compares the predicted values ​​with the current environmental state vector and the target range of the five constant indicators, calculates the improvement amount of each indicator deviation and the residual deviation after adjustment, and pre-configures the control improvement threshold and control degradation threshold in the parameter table. These are used to limit the improvement conditions that a candidate adjustment combination must meet and the degradation boundary that it must not exceed when it is considered an effective combination. The control improvement threshold can be set as the minimum improvement ratio or minimum absolute value of each five constant indicator's predicted deviation relative to the current deviation. The control degradation threshold can be set as the maximum allowable slight degradation of individual indicators within an acceptable range. The specific numerical range is registered separately for each indicator and locked through the parameter table version number.

[0066] To describe the overall cost of equipment regulation, the control center assigns cost weights to power consumption, noise level, and equipment start-up and shutdown behavior in the parameter table. The power consumption cost is calculated based on the estimated power consumption per unit control time window registered in the feasible adjustment space. The noise cost is calculated based on the noise contribution value of each level and the degree of closeness of the predicted noise level to the noise limit and allowable fluctuation range. The start-up and shutdown cost is calculated based on the start-up and shutdown penalty factor, the existence of start-up and shutdown actions in the current cycle, and the actual start-up and shutdown frequency in the model correction observation window. The default values ​​for the above cost weights are configured in the parameter table and can be adjusted as needed according to time period, room use, or disturbance type label.

[0067] In this scheme, the comprehensive score is defined as a scalar index that reflects the balance between the degree of proximity of a candidate control combination to the five constant objectives, power consumption constraints, noise control, and equipment life protection. When scoring, the control center combines the residual deviation of each of the five constant indicators with the aforementioned weight parameters, and combines the power consumption cost, noise cost, and start-up and shutdown cost with the cost weights in the parameter table. Preferably, the comprehensive score value can be calculated by weighted summation of each cost and deviation or other monotonically increasing functions. The specific scoring method is disclosed in the parameter dictionary in the form of a text description.

[0068] During the comprehensive scoring process, the system first eliminates candidate adjustment combinations that, according to the equipment impact response table, would cause any of the five constant indicators to exceed the safety boundary. In this scheme, the safety boundary is defined as the upper and lower limits that each of the five constant indicators cannot exceed in terms of safe operation and health and comfort. It can be uniformly configured in the parameter table during the design phase based on the standard limits, health and comfort requirements, and equipment operation restrictions.

[0069] Within the set of combinations that meet the safety boundary conditions, the system further eliminates combinations that do not meet the control improvement threshold or exceed the control degradation threshold, retaining only combinations that can bring the five constant indicators closer to the target range as a whole and prevent any indicator from deviating beyond the acceptable degradation range, and sorting them from best to worst according to the comprehensive score.

[0070] To handle situations where scores are very close, a score similarity threshold is set in the parameter table. This threshold is used to define two candidate control combinations as having similar scores if the difference in their overall scores is less than this threshold. Among combinations with a score difference less than the score similarity threshold, the control center preferably makes a decision based on the following order: lower power consumption first; lower noise first when power consumption is the same or the score difference is still within the score similarity threshold; and fewer start-stop actions first when the first two are the same. Finally, the candidate control combination with the best overall score is selected as the target equipment control vector for this control cycle.

[0071] The default values ​​for control cycle length, control time window length, upper limit of candidate combination number, control improvement threshold, control degradation threshold, safety boundary parameters, power consumption cost weight, noise cost weight, start-stop cost weight, and score similar threshold are uniformly configured in the parameter table, and the parameter table version number is recorded.

[0072] While generating the target equipment adjustment vector, the control center writes the target equipment adjustment vector, along with the version number of the five constant templates used in this cycle, the version number of the equipment impact response table, the version number of the feasible adjustment space, the version number of the parameter table, and the current disturbance type label, into the control decision record and appends it to the control decision archive. This provides a traceable version basis for issuing control commands in the subsequent hierarchical execution phase and revising the equipment impact response table and feasible adjustment space in the self-learning phase.

[0073] In implementations with limited computing power or requiring strict real-time performance, a discrete candidate adjustment combination library can be pre-built during the system design phase. Representative equipment level combinations and their estimated effects are stored in the control center in the form of records. During operation, the control center selects a set of combinations from the candidate combination library within the feasible adjustment space, up to the upper limit of the number of candidate combinations. By querying the equipment impact response table and applying the aforementioned cost and scoring rules, these combinations are subjected to the same prediction and comprehensive scoring, and the optimal combination is selected. As long as the selection of candidate combinations is still constrained by the feasible adjustment space and the calculation of the comprehensive score is still based on the equipment impact response table and cost configuration rules disclosed in this invention, it can be regarded as an equivalent implementation of this step.

[0074] S5. Decompose the target adjustment vector into building-level and room-level control commands, and issue them to centralized equipment and room terminals. Call the preset control strategy template according to the disturbance type and adjust the feasible adjustment space according to the state deviation. The specific implementation is as follows: After generating target equipment adjustment vectors for each room, the control center splits these vectors into building-level control commands and room-level control commands according to equipment hierarchy, to distinguish the execution paths of centralized equipment and room-level terminal equipment. Building-level control commands are used to adjust centralized fresh air units, centralized cooling sources, and main exhaust systems, while room-level control commands are used to adjust indoor air conditioning units, dehumidifiers, local exhaust systems, and terminal fans in each room. The two types of commands are distinguished within the control center by the command type field.

[0075] The system pre-establishes a control strategy template library and associates the disturbance type labels determined during the disturbance type identification stage with the control strategy templates. Each control strategy template is indexed by the template identifier and strategy version number and includes at least the priority order of building-level and room-level adjustments, the activation or restriction rules of various devices under specific disturbance types, and the allowed adjustment range or power range. In the window opening disturbance scenario, the power increase of the centralized cold source can be limited and the upper limit of the room-level dehumidification level can be restricted. In the bathing disturbance scenario, the priority of the local exhaust ventilation device in the bathroom room is increased and the deviation of the temperature setting of other rooms is restricted. In the party disturbance scenario, the temperature control precision of the bedroom or living room is appropriately relaxed to ensure that the fresh air volume and air cleanliness indicators meet the corresponding target range.

[0076] When generating each specific control instruction, the control center uses the target equipment adjustment vector of the current control cycle as a basis and, according to the priority order and restriction rules defined in the control strategy template, generates building-level control instruction records for centralized fresh air units, centralized cold sources, and main exhaust systems, as well as room-level control instruction records for indoor air conditioning units, dehumidifiers, local exhaust systems, and terminal fans in each room. Each instruction record includes at least the building identifier, room identifier (a specific building identifier may be used for only building-level equipment), equipment identifier, time synchronization, target gear or power level, strategy version number, idempotency identifier, and sequence number.

[0077] The idempotency identifier is an identifier that uniquely identifies the same control intent within the idempotency window length. The idempotency window length is used to limit the effective time interval of the same control intent and is configured in the parameter table in the form of time length or number of control cycles. The sequence number is used to establish the order of successive control commands of the same device. The generation rules and encoding format of the idempotency identifier and sequence number are registered in the parameter dictionary.

[0078] The aforementioned control commands are transmitted to the corresponding devices or device controllers via fieldbus or local area network. The communication method is uniformly agreed upon during the system design phase. The control center configures the command round-trip delay limit and retransmission limit for each type of communication link in the parameter table. If no execution result summary and status code are received from the device within the command round-trip delay limit after the command is issued, the command is marked as timed out and retransmitted a limited number of times according to the retransmission limit. If no feedback with a valid status code is received after the specified number of retransmissions, the device is degraded according to the degradation control rules in the parameter table, limiting the current operating state of the device to a predefined safety level or safe power range. The safety level and safe power range are identified by the safety flag field in the feasible adjustment space. The control center records this degradation control behavior, along with the device identifier, occurrence time, communication status, and error code, in the operation and maintenance log and evidence chain log for subsequent investigation and tracing.

[0079] Since control commands from different rooms may be issued in parallel within the same control cycle, the control center maintains a device command management table locally. Each issued command is registered according to the device identifier, idempotency identifier, and sequence number. When the execution result summary and status code of the device feedback are received, the control center first determines whether the feedback corresponds to the current valid control intent based on the idempotency identifier. Duplicate feedback with the same idempotency identifier and sequence number is deduplicated. Feedback with the same idempotency identifier but a sequence number lower than the current sequence number registered in the device command management table is identified as expired feedback according to the principle of "higher sequence number overwrites lower sequence number". Only the current setting status of the device corresponding to the highest sequence number is retained in the device command management table.

[0080] The status codes are defined in the parameter dictionary and have a range of values. Preferably, they include at least the following types: execution successful, parameter invalid, device self-test failed, device in protection state, communication abnormality, and partial execution successful. The control center can add different levels of prompt information to the operation and maintenance log based on different status codes.

[0081] To suppress on-site oscillations and excessive overshoot during execution, the control center continuously monitors the deviations of the five constant indicators in each room based on the environmental state vector. Parameters such as oscillation observation window length, oscillation amplitude threshold, overshoot amplitude threshold, gear switching frequency threshold, and boundary crossing number threshold are pre-configured in the parameter table. The oscillation observation window length is used to limit the time range for identifying oscillation behavior; the oscillation amplitude threshold is used to limit the minimum amplitude of the indicator's fluctuation around the target interval boundary within the window; the boundary crossing number threshold is used to limit the minimum number of round trips the indicator makes across the target interval boundary within the oscillation observation window; the overshoot amplitude threshold is used to limit the maximum allowable deviation of the indicator's peak value relative to the target interval or safety boundary; and the gear switching frequency threshold is used to limit the upper limit of the allowed number of gear or power level changes for equipment in the same room within the oscillation observation window.

[0082] If the control center detects within the oscillation observation window that a certain constant index of a certain room crosses the boundary of the target interval a number of times and the fluctuation amplitude exceeds the oscillation amplitude threshold, or if the peak value of the index exceeds the overshoot amplitude threshold of the target interval or safety boundary, or if the number of times the equipment switches gears in the room exceeds the gear switching frequency threshold, the control center will temporarily tighten the feasible adjustment space corresponding to the room according to the field stability constraint rules in the parameter table.

[0083] To standardize the definitions of "high-level gears," "large-amplitude adjustments," and "small-step adjustments," a gear step size threshold is pre-configured in the parameter table. In this scheme, the gear change step size is defined as the difference between the target gear or target power level and the current gear or current power level on a unified scale. Adjustments with a gear change step size greater than the gear step size threshold are considered large-amplitude adjustments or high-level gear transitions, while adjustments with a gear change step size less than or equal to the gear step size threshold are considered small-step adjustments or gradual adjustments.

[0084] When tightening the feasible adjustment space, the control center marks all candidate adjustment actions in the room with a gear change step size greater than the gear step size threshold as temporarily unavailable, retaining only candidate adjustment combinations with a gear change step size less than or equal to the gear step size threshold, and registering the room's current adjustment convergence protection state in the room file. The minimum holding time of the adjustment convergence protection state is configured in the parameter table in the form of time length or number of control cycles. Within this holding time interval, the control center prioritizes candidate adjustment combinations whose adjustment amplitude meets the gear step size threshold constraint and whose comprehensive score still meets the control improvement threshold requirements, avoiding instability in field operation caused by frequent switching and repeated overshoot.

[0085] The aforementioned processes for generating, issuing, processing feedback, degrading control, identifying oscillations, and adjusting convergence protection of building-level and room-level control commands, together with the target equipment adjustment vector, control strategy template version number, equipment impact response table version number, and feasible adjustment space version number in the aforementioned control decision record, form a complete execution chain. This enables those skilled in the art to implement the hierarchical control and on-site steady-state protection of this invention under different building scales and equipment topologies, provided that the communication links, parameter table configurations, and equipment capability constraints are known.

[0086] S6. After the control is completed, record the five constant recovery characteristics, energy consumption, and number of start-stop cycles. Record user manual intervention as negative feedback. Periodically adjust the demand template, cost weight, and impact response table parameters based on the operation records and negative feedback. The specific implementation is as follows: After each control time window corresponding to the target equipment adjustment vector ends, the control center records the key operational performance of the adjustment process in a unified manner into the historical record for subsequent self-learning and version revision. In this scheme, for each room and each control cycle, the control center calculates and records the five constant recovery characteristics, power consumption statistics, and equipment start-up and shutdown count statistics for this adjustment based on the environmental state vector sequence continuously stored at the time of the time signal within the control time window, combined with the five constant target intervals and safety boundaries effective in that cycle.

[0087] The five constant indicators recovery characteristics include three categories of indicators: the time required to recover to the target range, the maximum overshoot during the recovery process, and the fluctuation range after stabilization. The time required to recover to the target range is defined as the time interval from the moment the control command is issued until all the five constant indicators specified by the current target equipment adjustment vector in the room have entered their respective target ranges and have not exceeded the boundary of the target range again within a preset duration. This preset duration is configured in the parameter table in the form of time length or number of control cycles. The maximum overshoot is defined as the maximum absolute value of the measured value of any five constant indicator that exceeds the upper limit of the target range or falls below the lower limit of the target range during the above recovery process. The fluctuation range after stabilization is defined as the fluctuation range of the measured values ​​of each five constant indicator within the target range during the period of recovery to the target range and maintaining stability. The fluctuation statistical window length and the statistical methods of extreme value difference, standard deviation, or quantile bandwidth are registered in the parameter table and parameter dictionary.

[0088] The energy consumption record includes the total building energy consumption within the control time window, as well as the energy consumption statistics by equipment and by room. The total energy consumption can be provided by the difference in the building's total electricity meter reading or by the energy consumption metering system. The itemized statistics are provided by the energy metering devices corresponding to each controlled equipment or power distribution circuit. The control center registers the mapping relationship between the metering device identifier and the equipment identifier in the parameter dictionary to ensure consistency in the itemized statistics.

[0089] The equipment start-stop count record includes the start-stop counts of various controlled equipment within the control time window. A start-stop event is defined as a state change where the equipment switches from a stopped state to a running state or from a running state to a stopped state. Each start-stop event is marked with a timestamp and event type in the equipment status record. At the end of the control time window, the control center summarizes the counts according to the equipment identifier and associates them with the start-stop count limit and start-stop penalty factor registered in the parameter table to support subsequent cost assessment and equipment life analysis.

[0090] To collect user-side subjective dissatisfaction with the control strategy, the system configures a negative feedback observation window length and a negative feedback trigger threshold system in the parameter table. The negative feedback observation window length is used to limit the time range for judging negative feedback behavior. The setpoint change number threshold is used to limit the minimum count of setpoint changes for temperature, humidity, or scene-related settings in the same room within this time range. The device shutdown number threshold is used to limit the minimum count of user-initiated shutdown of a certain type of controlled device in the same room within this window. The scene switching number threshold is used to limit the minimum count of scene mode switching. When, within the time range corresponding to the negative feedback observation window length, the number of setpoint changes in a room reaches or exceeds the setpoint change number threshold, the number of device shutdowns reaches or exceeds the device shutdown number threshold, or the number of scene switching reaches or exceeds the scene switching number threshold, the control center registers this behavior as a negative feedback event, writes it into the negative feedback record using room identifier, disturbance type label, and scene mode identifier as indexes, and archives it according to disturbance type and scene label.

[0091] The control center periodically summarizes and analyzes historical records and negative feedback records according to the analysis cycle length configured in the parameter table. The analysis cycle length can be configured in daily, weekly, or monthly units. During the summary analysis, the average recovery time, recovery time distribution, maximum overshoot distribution, stability fluctuation range statistics, power consumption statistics, and negative feedback event frequency are calculated for each room within the analysis cycle and compared with the recovery time upper limit threshold, overshoot upper limit threshold, fluctuation range upper limit threshold, and negative feedback frequency threshold configured in the parameter table.

[0092] To standardize the criteria for determining "energy consumption within an acceptable range," the parameter table sets upper limit thresholds or energy consumption deviation thresholds for different room uses and operating scenarios. The upper limit threshold limits the maximum absolute level of energy consumption per unit time within the analysis period, while the energy consumption deviation threshold limits the allowable deviation range relative to the baseline energy consumption. In this scheme, the baseline energy consumption is defined as the benchmark unit time energy consumption registered in the version index table for room use and operating scenario. It can be generated based on the statistical average of the previous analysis period or the reference value given in the design phase, and the source type and update cycle are recorded in the parameter table. When the unit time energy consumption of a room does not exceed the upper limit threshold and the absolute value of the energy consumption deviation does not exceed the energy consumption deviation threshold, it is considered that the energy consumption is within an acceptable range.

[0093] During the above statistical comparison process, when it is found that a certain room exceeds the upper limit threshold for recovery time, the upper limit threshold for maximum overshoot, the upper limit threshold for fluctuation range after stabilization, or the corresponding negative feedback frequency threshold within the analysis period, the control center marks the room as a room that needs strategy optimization.

[0094] For rooms marked as requiring strategy optimization, the control center selects the corresponding template record from the demand template library based on the room's purpose, disturbance type label, and the currently effective demand template version. Combining the room's recovery time, overshoot characteristics, fluctuation characteristics, and energy consumption performance within the analysis period, the control center adjusts the weights of the five constant target ranges and cost weights in the demand template. For example, in cases of large overshoot, the target range is narrowed or the corresponding indicator weights are increased. In cases of excessively long recovery time but acceptable energy consumption, the energy consumption weights are appropriately relaxed to improve response speed. The control center also corrects the typical change range in the relevant equipment impact response table to make the model more closely resemble actual operating performance.

[0095] The revised requirement templates and parameter tables are written to the template library and parameter table storage area with new version numbers. The new version number, effective time, previous version number, and corresponding evidence chain log position are registered in the version index table. The evidence chain log contains the historical statistical summary, triggering conditions, operator identification, and change content summary on which this version adjustment is based. In subsequent control cycles, the currently effective version is queried in the version index table. The latest version of the requirement templates and parameter configurations is used first, while the old versions are retained to support the traceability of the operation process.

[0096] To ensure the system's safety and compliance in scenarios involving data anomalies or unreliable models, the parameter table pre-configures a set of key five constant indicators and a threshold for consecutive missing periods. The set of key five constant indicators is used to identify a subset of the five constant indicators that must be monitored in terms of safety and health, and the threshold for consecutive missing periods is used to limit the upper limit of the number of control periods that can be consecutively missing. At the same time, the parameter table configures minimum sample size requirements according to equipment type and operating condition category, which is used to limit the minimum number of valid samples required for each equipment and operating condition combination in the equipment impact response table.

[0097] When the control center detects that any indicator belonging to the set of five key constant indicators cannot be obtained or is marked as unqualified within the number of control cycles corresponding to the consecutive missing cycle threshold, or detects that the number of valid samples for the working condition combination corresponding to a certain type of key equipment in the equipment impact response table is lower than the minimum sample number requirement, or the requirement template record is marked as pending review in the version management interface, the control center automatically switches the corresponding room or equipment group to safe mode operation.

[0098] In this solution, the safety mode is defined as only allowing the use of safety gear combinations identified by the safety marker field within the feasible adjustment space and the corresponding target equipment adjustment vectors. It does not actively trigger high load gear or high noise gear combinations, and prioritizes conservative adjustment strategies that can maintain the five constant indicators within the safety boundary. The basic acceptable range of the environment is considered to be consistent with the aforementioned safety boundary in this solution. During the safety mode, the control center displays the safety mode status on the maintenance interface with room and equipment identifiers, and prompts maintenance personnel to manually check and repair the on-site sensors, equipment operating status, and template configuration according to the prompt information.

[0099] The conditions for releasing the safety mode are configured in the parameter table. These conditions may include the following: the data quality of the five key constant indicators returns to normal within a number of consecutive control cycles and no longer triggers the missing data judgment; the number of valid samples for the corresponding equipment operating condition combination reaches or exceeds the minimum sample number requirement; and the requirement template is marked as approved in the review process. Once the conditions for releasing the safety mode are met, the control center will restore the requirement template and parameter table configuration under the normal control mode in the version index table.

[0100] In some implementations, the initial version of the equipment impact response table can be manually compiled by technicians based on performance curves, experimental test data, and historical operation records of similar buildings provided by the equipment manufacturer, according to the equipment type field, gear or power level field, five constant indicators change field, and response time window field specified in this invention. This serves as the default version for the initial stage of system operation. As the building gradually accumulates local operation samples under different seasons and operating scenarios, the control center writes the local statistical results into a new impact response table version according to the aforementioned model correction and version update mechanism, gradually replacing the initial version. As long as the five constant indicators set, environmental state vector field caliber, and target equipment adjustment vector definition range remain consistent, it is considered an equivalent implementation of this step, thereby ensuring that the self-learning process is carried out within the existing model framework and does not change the overall control logic and hierarchical decision-making approach of this invention.

[0101] Example 2: Figure 2 A schematic diagram of the five-constant environment adaptive control system based on the Internet of Things (IoT) of this invention is provided. The five-constant environment adaptive control system based on the IoT includes: The environmental status construction module is used to collect five constant indicators, room occupancy information and equipment operating status. Based on the time anchor, the collected data is time-aligned, an environmental status vector of each room is constructed, and the room use and building characteristics are registered. The demand and disturbance identification module is used to obtain the corresponding five constant target intervals and weight parameters from the demand template based on the room use, occupancy information and current time, analyze the changing characteristics of the environmental state vector, and identify and mark the current disturbance type. The equipment impact modeling module is used to establish an impact response table of various terminal equipment on the five constant indicators based on debugging data and operation data, determine the feasible adjustment space of each equipment under the current operating conditions, and introduce electricity price constraints, noise constraints and start-stop constraints in the modeling process. The adjustment combination solution module is used to generate candidate adjustment combinations based on environmental state deviation, five constant target ranges and feasible adjustment space within the control cycle. It uses the influence response table to predict the changes in five constant indicators and control costs caused by each candidate combination, and selects the target equipment adjustment vector based on the control costs. The hierarchical execution module is used to split the target device adjustment vector into building-level control commands and room-level control commands, which are then sent to the central device and the room terminal device, respectively. It also calls the preset control strategy template according to the disturbance type and adjusts the feasible adjustment space according to the environmental state deviation during the execution process. The self-learning correction module is used to record the recovery characteristics of the five constant indicators, energy consumption data, and the number of equipment start-ups and shutdowns after the control ends. It registers user manual interventions as negative feedback information and periodically corrects the demand template, control cost weight, and impact response table parameters based on the operation records and negative feedback information.

[0102] The technical connections and implementation logic of the six modules are as follows: During the overall system operation, the environmental state construction module first collects five constant indicators, room occupancy information, and equipment operating status based on the time anchor. It then performs time alignment on the multi-source data to form an environmental state vector labeled with room usage and building characteristics, providing a unified state baseline for subsequent judgments. Based on this, the demand and disturbance identification module reads the corresponding five constant target intervals and weight parameters from the demand template according to room usage, occupancy information, and the current time. It then identifies and labels the current disturbance type based on the time change characteristics of the environmental state vector, transforming the original state information into control demands with target and disturbance labels. The equipment impact modeling module uses debugging and operational data to construct impact response tables for various types of end-point equipment and determine feasible adjustment spaces. It also introduces electricity price constraints, noise constraints, and start-stop constraints to establish a correspondence between equipment adjustment actions and changes in the five constant indicators and control costs, providing a model foundation for subsequent optimization. The adjustment combination solution module calls the aforementioned environmental state vector and five constant target intervals within each control cycle. The system generates candidate control combinations based on weight parameters and feasible adjustment space. It predicts the changes in the five constant indicators and corresponding control costs of each candidate combination through the influence response table, and selects the target equipment adjustment vector under constraints. The hierarchical execution module decomposes the target equipment adjustment vector into building-level control commands and room-level control commands, and issues them to the central equipment and room terminal equipment. It also calls the corresponding control strategy template based on the disturbance type. During the execution process, it dynamically tightens or loosens the feasible adjustment space according to the real-time environmental state deviation. The self-learning correction module records the recovery characteristics of the five constant indicators, energy consumption data, and the number of equipment start-ups and shutdowns after each control operation. It registers user manual intervention as negative feedback information and periodically corrects the demand template, control cost weights, and influence response table parameters based on the accumulated operation records and negative feedback information. It also updates the templates and models used by the aforementioned modules in reverse, thus forming an adaptive control logic with environmental state construction, demand identification, equipment modeling, optimization solution, hierarchical execution, and self-learning correction in a closed loop.

[0103] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0104] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0105] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0106] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0107] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0108] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0109] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0110] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0111] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0112] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A five-constant environment adaptive control method based on the Internet of Things, characterized in that, include: S1. Collect the five constant indicators, occupancy and equipment status, construct the environmental status vector of each room according to the time anchor, and register the room use and building characteristics; S2. Based on the room's purpose, occupancy, and time, obtain the five constant target intervals and weight parameters from the demand template, analyze the state vector changes, and identify and label the disturbance types. S3. Based on debugging and operation data, establish equipment impact response tables and feasible adjustment space for each controlled device, taking into account electricity price, noise and start-stop constraints; S4. During the control cycle, a set of candidate control combinations is generated based on the state deviation, target range and feasible adjustment space. The five constant changes and costs of each combination are predicted using the response table. The equipment control vector that satisfies the constraints and has the lowest cost is selected. S5. Decompose the target adjustment vector into building-level and room-level control commands, send them to centralized equipment and room terminals, call the preset control strategy template according to the disturbance type, and adjust the feasible adjustment space according to the state deviation. S6. After the control ends, record the five constant recovery characteristics, energy consumption and number of start-stop cycles. Record user manual intervention as negative feedback. Periodically adjust the demand template, cost weight and impact response table parameters based on the operation records and negative feedback.

2. The adaptive control method for five constant environments based on the Internet of Things according to claim 1, characterized in that, S1 includes: The control center follows the time synchronization provided by the unified time synchronization device and the acquisition cycle, observation window length, deviation threshold and continuous missing report duration in the parameter table; The five constant indicators, occupancy information and equipment operating status of each room are collected, time-aligned and quality-corrected, and the collection channels that meet the continuous missing reporting time are marked as invalid. At each time synchronization, the corrected data above is combined into an environmental state vector, which is then associated with the room use and building characteristics and stored in the room file. If the key five constant indicators and key equipment status of the room are not recorded within the predetermined time limit, the room will be marked as a monitoring downgraded state, and subsequent adjustments will be limited to a conservative adjustment strategy that does not increase the cooling level, fresh air level, and fan level.

3. The five constant environment adaptive control method based on the Internet of Things according to claim 1, characterized in that, S2 include: In each control cycle, the control center reads the room's purpose, building characteristics, current time, and the environmental state vector sequence arranged in chronological order within the observation window from the room archives. Match a template record in the demand template library according to the room use, building characteristics and time period to obtain the five constant target intervals and weight parameters corresponding to temperature, relative humidity, air renewal index, air cleanliness index and noise sound pressure level index. The five constant target intervals, weight parameters, template identifiers, and template version numbers are written into the status record at that time.

4. The adaptive control method for five constant environments based on the Internet of Things according to claim 3, characterized in that: After obtaining the five constant target intervals and weight parameters, the control center uses the observation window length, deviation threshold, number of triggers, disturbance type label retention time, slow change threshold, sudden change time threshold, small change threshold and hold time threshold configured in the parameter table. The environmental state vector inside the observation window is jointly judged with the five constant target intervals, outdoor meteorological records, and door and window opening and closing status records. States that meet the criteria for weather disturbance, window opening disturbance, bathing disturbance, and gathering disturbance are labeled with the corresponding disturbance type tags; States that do not meet the four disturbance determination conditions are marked as unclassified states, and disturbance type labels are associated with and stored with the corresponding state records.

5. The five-constant environment adaptive control method based on the Internet of Things according to claim 1, characterized in that, S3 includes: Based on the debugging and operation records, the control center establishes an equipment impact response table and feasible adjustment space for each controlled device; The Equipment Impact Response Table records the changes in the five constant indicators and response delay of each piece of equipment within the response time window at different gears and power levels. The feasible adjustment space links each device's gear and power level with energy consumption estimates, noise contribution values, and start-stop penalty factors; The parameter table configures the correction observation window length, deviation revision threshold, and revision trigger count; Compare the predicted changes with the actual changes within the calibration observation window; When the deviation exceeds the deviation revision threshold and the number of times reaches the revision trigger count, update the device impact response table and feasible adjustment space, and register the updated version number in the version index table.

6. The five constant environment adaptive control method based on the Internet of Things according to claim 1, characterized in that, S4 include: During the control cycle, the control center reads the environmental state vector, the five constant target intervals, and the weight parameters from the room files. Based on the feasible adjustment space and equipment impact response table, candidate adjustment combinations are generated, including the operating settings of air conditioning indoor units, fresh air terminals, dehumidifiers, local exhaust systems, and terminal fans. Within the control time window, predict the changes in the five constant indicators for each candidate control combination based on the equipment impact response table; The control center eliminates candidate control combinations that do not meet the constraints based on the safety boundary, control improvement threshold, control degradation threshold, power consumption cost weight, noise cost weight, and start-up / shutdown cost weight. Calculate the comprehensive score for the remaining candidate control combinations and select the candidate control combination with the smallest comprehensive score as the target equipment control vector.

7. The five constant environment adaptive control method based on the Internet of Things according to claim 1, characterized in that, S5 include: After generating the target equipment adjustment vector, the control center breaks down the target equipment adjustment vector into building-level and room-level control commands according to the control strategy template; Each control command is assigned a device identifier, timing time, policy version number, and idempotency identifier, and registered in the device command management table. The control command is then sent through the communication link. If a valid status code is not obtained within the command round-trip delay limit, the operating status of the corresponding controlled device will be limited to a safe combination of levels within the feasible adjustment space. The control center identifies oscillations based on the oscillation observation window length, oscillation amplitude threshold, overshoot amplitude threshold, gear switching frequency threshold, and boundary crossing frequency threshold, and tightens the feasible adjustment space corresponding to the room according to the gear step size threshold. Only retain adjustment combinations with gear change step sizes not exceeding the gear step size threshold, mark the remaining adjustment actions as unavailable, and register the adjustment convergence protection status in the room file.

8. The five constant environment adaptive control method based on the Internet of Things according to claim 1, characterized in that, S6 include: At the end of the control time window, the control center generates a historical record containing the recovery characteristics of the five constants, power consumption statistics, and equipment start-up and shutdown statistics based on the environmental state vector and the five constant target intervals. Within the time range corresponding to the length of the negative feedback observation window, temperature setting changes, humidity setting changes, and equipment shutdown behaviors are compared with the threshold values ​​for the number of setting changes and the threshold values ​​for the number of equipment shutdowns to generate negative feedback records. The control center compares historical records and negative feedback records with the recovery time upper limit threshold, overshoot amplitude threshold, and negative feedback frequency threshold; In the room archive, mark the rooms that meet the judgment criteria as rooms that require strategy optimization; Adjust the five constant target ranges and cost weights in the room demand template; Update the typical change records in the Equipment Impact Response Table.

9. An Internet of Things-based adaptive control system for five constant environments, used to implement the Internet of Things-based adaptive control method for five constant environments as described in any one of claims 1-8, characterized in that, include: The environmental status construction module is used to collect five constant indicators, room occupancy information and equipment operating status. Based on the time anchor, the collected data is time-aligned, an environmental status vector of each room is constructed, and the room use and building characteristics are registered. The demand and disturbance identification module is used to obtain the corresponding five constant target intervals and weight parameters from the demand template based on the room use, occupancy information and current time, analyze the changing characteristics of the environmental state vector, and identify and mark the current disturbance type. The equipment impact modeling module is used to establish an impact response table of various terminal equipment on the five constant indicators based on debugging data and operation data, determine the feasible adjustment space of each equipment under the current operating conditions, and introduce electricity price constraints, noise constraints and start-stop constraints in the modeling process. The adjustment combination solution module is used to generate candidate adjustment combinations based on environmental state deviation, five constant target ranges and feasible adjustment space within the control cycle. It uses the influence response table to predict the changes in five constant indicators and control costs caused by each candidate combination, and selects the target equipment adjustment vector based on the control costs. The hierarchical execution module is used to split the target device adjustment vector into building-level control commands and room-level control commands, which are then sent to the central device and the room terminal device, respectively. It also calls the preset control strategy template according to the disturbance type and adjusts the feasible adjustment space according to the environmental state deviation during the execution process. The self-learning correction module is used to record the recovery characteristics of the five constant indicators, energy consumption data, and the number of equipment start-ups and shutdowns after the control ends. It registers user manual interventions as negative feedback information and periodically corrects the demand template, control cost weight, and impact response table parameters based on the operation records and negative feedback information.

Images