Green building environment monitoring system and method and storage medium

Abstract

The invention relates to the technical field of green building environment monitoring, and discloses a green building environment monitoring system and method and a storage medium. The method comprises the steps that environmental parameters such as thermodynamic parameters, humidity parameters and gas concentration parameters are continuously collected through intelligent sensing nodes at multiple positions in a green building; the environment parameters are transmitted to a central processing unit, and the central processing unit combines a pre-stored environment standard value and a real-time data stream and generates an environment control command by adopting a self-adaptive learning algorithm; the environment execution device carries out an environment regulation activity according to the command; and monitoring and adjusting an activity execution sequence in real time, calculating a performance evaluation index, generating a normal or abnormal performance state signal, and automatically triggering an alarm process when the performance state signal is abnormal. According to the method, a complete environment management and control closed loop is constructed, all-directional dynamic monitoring and accurate adjustment of the environment of the green building are achieved, the core requirements for environment protection and comfort of the green building are met, and it is guaranteed that the internal environment of the building is in a proper state.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of green building environment monitoring, in particular to a green building environment monitoring system and method and a storage medium. BACKGROUND

[0002] With the transformation of the construction industry towards low-carbon and environmental protection, green building has become the mainstream trend of industry development, and its core is to achieve the dual goals of efficient resource utilization and living environment optimization through scientific design and control. As a key evaluation dimension of green building, environmental quality is directly related to the physical and mental health of residents and the environmental performance of buildings, so accurate and continuous monitoring and regulation of the internal environment of green buildings has become an important part of the operation and maintenance process of green buildings.

[0003] Currently, in the existing green building environment monitoring scheme, some use a single sensing node for data collection, which has the problems of limited monitoring range and incomplete data coverage, and it is difficult to reflect the environmental differences in different areas of the building, resulting in a lack of targetedness of subsequent control measures. Although some schemes set up multiple sensing nodes, the types of environmental parameters collected are relatively single, often only focusing on a few indicators such as temperature or humidity, ignoring parameters such as gas concentration that have important influence on human health and indoor air quality, making the monitoring results not comprehensive enough to provide complete data support for environmental regulation.

[0004] In the environmental control link, most existing technologies use fixed control logic to generate regulation commands, lacking dynamic adaptation capability to real-time environmental data. Such fixed logic is difficult to cope with the complex changes in the internal environment of the building, such as sudden changes in environmental parameters caused by factors such as personnel flow and external environmental fluctuations, which can easily lead to poor regulation results and fail to restore the environment to a suitable state in a timely manner. At the same time, the existing scheme lacks an effective monitoring mechanism for the execution process of environmental regulation activities, making it difficult to know the actual execution and effect of the regulation measures and to determine whether the regulation activities have achieved the expected goal.

[0005] The existing technology lacks a perfect evaluation system for environmental regulation performance, and often can only find problems when there are obvious environmental abnormalities, lacking the ability to predict and warn in advance. Once environmental parameters exceed the standard or regulation devices fail, it is difficult to trigger the alarm process quickly, which may result in the persistence of abnormal conditions, affecting the experience of residents and even endangering their health. The existence of these problems makes the existing green building environment monitoring and regulation scheme lack practicality and reliability, making it difficult to fully meet the high standards of green buildings for environmental quality and limiting the full play of the environmental protection and comfort value of green buildings. SUMMARY

[0006] The present application aims to provide a green building environment monitoring system, method and storage medium to solve the problems in the background art.

[0007] To achieve the above object, the present application provides a green building environment monitoring method, comprising: Intelligent sensing nodes distributed in multiple locations inside the green building continuously collect environmental parameters, including thermodynamic parameters, humidity parameters and gas concentration parameters; The environmental parameters are sent to a central processing unit, which generates environment control commands using an adaptive learning algorithm based on pre-stored environmental standard values and real-time data streams; An environment execution device adjusts the environment according to the environment control commands; The execution sequence of the environment adjustment activities is monitored in real time, performance evaluation indicators are calculated, and performance status signals are generated based on the performance evaluation indicators, including normal signals and abnormal signals; When the performance status signal is an abnormal signal, an alarm process is triggered.

[0008] Preferably, the process of calculating performance evaluation indicators and generating performance status signals based on the performance evaluation indicators comprises: The environment adjustment activities are divided into optimized activities or non-optimized activities using a hierarchical performance analysis technique; The ratio of the frequency of non-optimized activities to the total number of activities is calculated within a fixed monitoring period to obtain a non-optimized ratio; If the non-optimized ratio is greater than a pre-set non-optimized critical value, an abnormal signal is output; If the non-optimized ratio is not greater than the pre-set non-optimized critical value, the delay time and energy consumption offset of the environment adjustment activities are measured; The delay time and energy consumption offset are normalized and fused to obtain an overall performance score; If the overall performance score exceeds a pre-set performance limit, an abnormal signal is output, otherwise a normal signal is output.

[0009] Preferably, the process of applying a hierarchical performance analysis technique to divide the environment adjustment activities into optimized activities or non-optimized activities comprises: The time point at which the central processing unit issues an environment control command is recorded as the starting time, and the time point at which the environment execution device completes the environment adjustment activity is recorded as the end time, and the duration between the starting time and the end time is defined as the monitoring interval; The ratio of the environment control command to the monitoring interval is taken as the command efficiency index, and the activity fluctuation index is obtained through a stability test method; if the command efficiency index is not within a predefined efficiency range or the activity fluctuation index is higher than a predefined fluctuation upper limit, marking the environmental regulation activity as a non-optimized activity; if the command efficiency index is within a predefined efficiency range and the activity fluctuation index is not higher than a predefined fluctuation upper limit, marking the environmental regulation activity as an optimized activity.

[0010] Preferably, the process of the stability test method comprises: constructing a two-dimensional coordinate system in the time dimension and the environmental parameter value dimension, and obtaining the environmental parameter change trajectory during the environmental regulation activity; setting multiple sampling points on the environmental parameter change trajectory, recording the environmental parameter difference between adjacent sampling points as the fluctuation amplitude, calculating the standard deviation of all fluctuation amplitudes as the fluctuation coefficient, and counting the proportion of fluctuation amplitudes exceeding the predetermined fluctuation interval as the abnormal coefficient; deriving the activity fluctuation index through linear combination of the fluctuation coefficient and the abnormal coefficient.

[0011] Preferably, the method further comprises: performing auxiliary abnormality detection when a normal signal is generated, and the process of the auxiliary abnormality detection comprises: collecting vibration amplitude data and sound pressure level data during the operation of the environmental execution device, and if the vibration amplitude data or the sound pressure level data exceeds the respective preset safety limit, determining that the environmental execution device has a potential fault; accumulating the duration of the potential fault within the monitoring period, and calculating the proportion of the duration of the potential fault to the total operation time of the environmental execution device to obtain the fault time proportion; counting the number of cases in which the number of consecutive occurrences of the potential fault within the monitoring period exceeds a predetermined number as the high-frequency fault count; recording the longest duration of a single potential fault within the monitoring period as the maximum fault duration; combining the fault time proportion, the high-frequency fault count, and the maximum fault duration to obtain a fault evaluation value through weighted summation; if the fault evaluation value is greater than a preset fault threshold, generating an auxiliary abnormality indication; and if the fault evaluation value is not greater than the preset fault threshold, generating an auxiliary normal indication.

[0012] Preferably, the process of the auxiliary abnormality detection further comprises: when the auxiliary abnormality indication is generated, further performing device health diagnosis, and the process of the device health diagnosis comprises: obtaining the manufacturing date of the environmental execution device, calculating the difference between the current date and the manufacturing date to obtain the device service age; extracting the cumulative operation time of the environmental execution device from the historical database to obtain the total operation duration of the device; obtaining the high-risk exposure time of the environmental execution device through environmental risk analysis; query the number of times that the maintenance interval of the environmental execution device exceeds the standard maintenance interval from the maintenance record as a non-standard maintenance number; The equipment health index is calculated by a multi-layer perceptron model based on the equipment service age, the total running time of the equipment, the high-risk exposure time and the non-standard maintenance number; If the equipment health index is lower than the preset health standard, a device replacement suggestion is generated.

[0013] Preferably, the process of the environmental risk analysis comprises: The surrounding temperature value and the surrounding humidity value at the installation position of the environmental execution device are collected, the absolute error between the surrounding temperature value and the ideal temperature value is calculated as the temperature deviation amount, and the absolute error between the surrounding humidity value and the ideal humidity value is calculated as the humidity deviation amount; The concentration of suspended particulate matter at the installation position of the environmental execution device is collected as the dust concentration value; The temperature deviation amount, the humidity deviation amount and the dust concentration value are input into a risk assessment function, and an environmental risk coefficient is output; If the environmental risk coefficient is greater than the preset risk standard, it is determined that the environmental execution device is in a high-risk environment, and the total time of the high-risk environment is accumulated as the high-risk exposure time.

[0014] Preferably, the process in which the central processing unit generates the environmental control command using an adaptive learning algorithm comprises: The difference between the environmental parameters and the pre-stored environmental standard values is input into a neural network model, the neural network model generates an environmental control command based on training data and real-time trends, and the environmental control command includes an adjustment type and an adjustment intensity; The adjustment type corresponds to a temperature raising operation, a temperature lowering operation, a humidity increasing operation, a humidity decreasing operation or an air purification operation, and the adjustment intensity is represented by a percentage scale.

[0015] Preferably, the present application further comprises a green building environment monitoring system, which comprises a memory, a processor and a computer program stored in the memory and running on the processor, and the processor implements the steps of the above-mentioned green building environment monitoring method when executing the computer program.

[0016] Preferably, the present application further comprises a computer readable storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the steps of the above-mentioned green building environment monitoring method.

[0017] Compared with the prior art, the present application has the following advantages: By setting intelligent sensing nodes at multiple locations inside the green building, comprehensive collection of environmental parameters is achieved. Multiple distributed sensing nodes can cover different functional areas and different spatial positions inside the building, avoiding the limitations of single node monitoring and ensuring that the collected data can truly reflect the overall and local environmental conditions inside the building. At the same time, the collected parameters cover thermodynamic parameters, humidity parameters and gas concentration parameters, comprehensively covering the key indicators affecting indoor environmental quality and residential comfort, providing a rich and complete data basis for subsequent environmental regulation, making environmental regulation more scientific.

[0018] The central processing unit generates environmental control commands using an adaptive learning algorithm, breaking the limitations of traditional fixed control logic. This algorithm can combine pre-stored environmental standard values with real-time collected data streams to continuously optimize control strategies, and can flexibly respond to dynamic changes in the building's internal environment, whether it is personnel flow, equipment operation or parameter changes caused by external environmental fluctuations. It can quickly respond to generate adjustment commands that adapt to the current state, making environmental regulation more targeted and flexible, effectively improving the accuracy of environmental regulation.

[0019] The environmental execution device carries out adjustment activities according to the control commands, while cooperating with the design of real-time monitoring execution sequence. Through real-time tracking of the execution sequence of adjustment activities, the implementation process of adjustment measures can be clearly understood, and the running state of the execution device can be understood, avoiding adjustment failure caused by execution device failure or inadequate execution. The calculation of performance evaluation indicators can objectively reflect the actual effect of adjustment activities, timely find problems in the adjustment process, provide direction for subsequent optimization of adjustment strategy and improvement of execution device running state, and ensure that environmental regulation activities are in efficient operation state.

[0020] The setting of performance state signals and the triggering mechanism of alarm procedures realize the rapid response to environmental abnormalities. When the performance evaluation index reflects poor adjustment effect or environmental parameters appear abnormal, an abnormal signal can be generated in time and an alarm can be triggered, so that relevant personnel can know the problem at the first time and take measures to troubleshoot and handle the problem quickly, avoiding the continuous spread or extension of the abnormal state. This timely early warning and alarm mechanism can minimize the impact of environmental abnormalities on the health of residents and ensure that the indoor environment is always in a safe and suitable state. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 A working principle diagram of the green building environmental monitoring method described in the present application; Figure 2 A flowchart for environmental regulation activities; Figure 3 A flowchart for stability testing method; Figure 4A bar chart for equipment failure evaluation index comparison; Figure 5 A heat map for equipment health influencing factors. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work are within the protection scope of the present application.

[0023] Please refer to Figure 1 The present application provides a green building environment monitoring method, which comprises: continuously collecting environment parameters by intelligent sensing nodes distributed in multiple positions inside the green building, the environment parameters including thermodynamic parameters, humidity parameters and gas concentration parameters. The intelligent sensing nodes send the collected environment parameters to a central processing unit, and the central processing unit generates environment control commands by using an adaptive learning algorithm according to pre-stored environment standard values and real-time data streams. An environment execution device performs environment adjustment activities according to the received environment control commands. The system monitors the execution sequence of the environment adjustment activities in real time, calculates performance evaluation indexes, and generates performance state signals based on the performance evaluation indexes, the performance state signals including normal signals and abnormal signals. When the performance state signal is an abnormal signal, the system triggers an alarm process to notify relevant personnel to take action.

[0024] Embodiment 1: Please refer to Figure 2 In a specific implementation, the process of generating the performance state signal starts with the classification of a series of environment adjustment activities. The system applies a hierarchical performance analysis technique to divide each complete environment adjustment activity into an optimized activity or a non-optimized activity. The specific operation of the hierarchical performance analysis technique is to record the precise time point at which the central processing unit issues an environment control command to the environment execution device, which is defined as the starting time. The system also records the time point at which the environment execution device completely completes the environment adjustment activity required by the environment control command, which is defined as the ending time. The time length between the starting time and the ending time is clearly defined as the monitoring interval, which represents the total time consumption of a single environment adjustment activity from instruction to complete completion.

[0025] In specific implementations, the classification process requires the calculation of two key indicators. The system defines a command efficiency index as the ratio of the adjustment intensity implied by the environmental control command itself to the monitoring interval, which reflects the performance of the environmental adjustment activities per unit of time. The system obtains another indicator, the activity fluctuation index, in parallel through a stability test method. The specific operation process of the stability test method is to construct a two-dimensional coordinate system in the time dimension and the environmental parameter value dimension, and then obtain the actual change trajectory of the relevant environmental parameters during the current environmental adjustment activity from start to end. On the environmental parameter change trajectory, the system sets multiple sampling points according to fixed time intervals or adaptive intervals, calculates the difference in environmental parameter values between adjacent sampling points, and records this difference as the fluctuation amplitude. The system calculates the standard deviation of all fluctuation amplitudes and defines this standard deviation as the fluctuation coefficient. The system also calculates the proportion of the number of fluctuation amplitudes whose absolute values exceed the predetermined fluctuation interval to the total number of fluctuation amplitudes, and defines this proportion as the abnormal coefficient. Finally, the activity fluctuation index is derived through the linear combination of the fluctuation coefficient and the abnormal coefficient, and the weight coefficient of the linear combination is a pre-set fixed value.

[0026] In specific implementations, after the calculation of the command efficiency index and the activity fluctuation index is completed, the system executes the classification judgment logic. The system internally pre-stores a predefined efficiency range and a predefined fluctuation upper limit. If the calculated command efficiency index value is not within the predefined efficiency range, or the calculated activity fluctuation index value is higher than the predefined fluctuation upper limit, the system will mark the current environmental adjustment activity as a non-optimized activity. Conversely, if the command efficiency index value is within the predefined efficiency range and the activity fluctuation index value is not higher than the predefined fluctuation upper limit, the system will mark the current environmental adjustment activity as an optimized activity. Each environmental adjustment activity will obtain a clear classification label.

[0027] In some embodiments, the system performs overall evaluation of performance within a fixed monitoring period. The fixed monitoring period can be set to twenty-four hours or one week. Within the fixed monitoring period, the system counts the frequency of occurrence of all environmental adjustment activities marked as non-optimized activities, and calculates the ratio of the frequency of occurrence of non-optimized activities to the total number of environmental adjustment activities within the fixed monitoring period, which is defined as the non-optimized ratio. The non-optimized ratio quantitatively describes the proportion of activities with poor system environmental adjustment performance within the fixed monitoring period.

[0028] In some embodiments, the system compares the calculated non-optimization ratio with a preset non-optimization threshold. The preset non-optimization threshold is a threshold value preset according to the reliability requirement of the system. If the non-optimization ratio is greater than the preset non-optimization threshold, the system directly determines that the overall performance is poor, and immediately outputs an abnormal signal. If the non-optimization ratio is not greater than the preset non-optimization threshold, it indicates that most of the activities are normal, but the system needs to further analyze the quality of the qualified activities, at which time the performance evaluation enters a more detailed stage.

[0029] Optionally, in the case of a qualified non-optimization ratio, the system measures two more subtle performance indicators for all the environmental adjustment activities marked as optimization activities within the fixed monitoring period. The first indicator is the delay time, which is the time interval from when the environmental execution device receives the environmental control command from the central processing unit to when it actually starts to execute the action. The second indicator is the energy consumption offset, which is the absolute difference between the actual energy consumption of the environmental execution device in completing a single environmental adjustment activity and the expected energy consumption calculated according to historical data or a model.

[0030] Optionally, the system needs to standardize the measured delay time and energy consumption offset to eliminate the dimensional effect. The system uses a normalization method to map the delay time and energy consumption offset into the value range of zero to one, respectively. After normalization, the system combines the normalized delay time value and energy consumption offset value into a single numerical indicator through a weighted fusion algorithm. This single numerical indicator is referred to as the overall performance score. The weights assigned to the delay time and energy consumption offset in the weighted fusion algorithm are preset according to the importance of their impact on the overall performance of the system.

[0031] It can be understood that after obtaining the overall performance score, the system compares it with a preset performance limit. The preset performance limit is a threshold value used to distinguish between good performance and critical performance. If the overall performance score exceeds the preset performance limit, it indicates that there is observable degradation in the execution efficiency or energy consumption control even in the environment classified as optimization activities, so the system outputs an abnormal signal. If the overall performance score does not exceed the preset performance limit, it indicates that the system is currently in good running condition, and the system outputs a normal signal.

[0032] It can be understood that the entire performance evaluation process is a hierarchical judgment structure, which first performs preliminary screening through the non-optimization ratio and then performs in-depth evaluation of optimization activities. This hierarchical performance analysis technique ensures that the evaluation of the system performance is both comprehensive and efficient, and can timely detect slight performance degradation trends, providing a data basis for preventive maintenance. The entire process is fully automated and does not require human intervention, and the evaluation results are output in the form of performance status signals, directly determining whether to trigger the subsequent alarm process.

[0033] Embodiment 2: refer to Figure 3 The stability test method is executed by constructing a data space, in which the system constructs a two-dimensional coordinate system with time dimension and environmental parameter value dimension. The time dimension is expressed in standard time units, covering the entire duration from the start to the end of the environmental conditioning activity; the environmental parameter value dimension is determined according to the specific monitoring object, which can be temperature value, humidity value, or specific gas concentration value, etc. In this two-dimensional coordinate system, the system plots the trajectory of the continuously changing environmental parameter from the start time to the end time of the environmental conditioning activity. This environmental parameter change trajectory is composed of a series of time-ordered data points, each containing a time stamp and a corresponding environmental parameter sampling value.

[0034] In specific implementation, after obtaining the environmental parameter change trajectory, the system sets multiple sampling points on the environmental parameter change trajectory along the time axis. The sampling point setting strategy can adopt equal time interval method, or adaptive interval method according to the severity of environmental parameter change, setting relatively sparse sampling points in the flat change area and relatively dense sampling points in the severe change area. The system calculates the environmental parameter value difference between adjacent sampling points and defines this difference as the fluctuation amplitude. The fluctuation amplitude is a positive and negative value, with positive value indicating environmental parameter rising and negative value indicating environmental parameter falling, and its absolute value size reflects the severity of environmental parameter change between adjacent sampling time.

[0035] In specific implementation, the system performs two independent statistical analyses based on the calculated sequence of all fluctuation amplitude values. The first analysis is to calculate the standard deviation of all fluctuation amplitudes, which is defined as the fluctuation coefficient. The fluctuation coefficient is a dimensionless statistical quantity that quantifies the dispersion degree of environmental parameter change rate during the entire environmental conditioning activity; a higher fluctuation coefficient means that the environmental parameter changes very unstably, sometimes severely and sometimes slowly. The second analysis is to set a predetermined fluctuation interval, which is a value range pre-set based on historical normal data and environmental control accuracy requirements. The system counts the proportion of the number of fluctuation amplitudes whose absolute values exceed the predetermined fluctuation interval to the total number of fluctuation amplitudes, which is defined as the abnormal coefficient. The abnormal coefficient directly reflects the frequency of abnormal severe fluctuations in the environmental parameter change process.

[0036] In some embodiments, the setting of the predetermined fluctuation range is based on multiple aspects. The lower and upper limit values of the predetermined fluctuation range are derived from the analysis of long-term historical data of green buildings in stable operation state, usually taking the average value of the normal fluctuation range plus several times the standard deviation. For different environmental parameters, such as temperature, humidity and carbon dioxide concentration, the numerical range of the predetermined fluctuation range is different. The system maintains a separate predetermined fluctuation range parameter table for each monitored environmental parameter, and calls the corresponding parameter value for calculation when performing stability test.

[0037] In some embodiments, after the calculation of the fluctuation coefficient and the abnormal coefficient, the system fuses the fluctuation coefficient and the abnormal coefficient into a single index, i.e. the activity fluctuation index, through a linear combination formula. The linear combination formula is that the activity fluctuation index is equal to the weight A multiplied by the fluctuation coefficient, plus the weight B multiplied by the abnormal coefficient. The weight A and the weight B are pre-set fixed coefficients, and the values of the weight A and the weight B are determined by analyzing the influence degree of the fluctuation coefficient and the abnormal coefficient on the stability of the system in the historical data. The sum of the weight A and the weight B is not necessarily 1, and the focus is that the weight A and the weight B can reasonably reflect the relative importance of the fluctuation coefficient and the abnormal coefficient in the overall instability evaluation.

[0038] Optionally, the calculation process of the activity fluctuation index also includes a standardization step. Since the fluctuation coefficient and the abnormal coefficient may have different orders of magnitude, direct linear combination may make one side dominant. Therefore, before combination, the system will normalize the fluctuation coefficient and the abnormal coefficient respectively, mapping them to a comparable numerical scale between zero and one, and then apply the weight A and the weight B for linear combination, to ensure the fairness and accuracy of the activity fluctuation index.

[0039] Optionally, after the activity fluctuation index is calculated, it is mainly used for subsequent classification judgment of environmental regulation activities. The activity fluctuation index, as a quantitative index to measure the stability of a single environmental regulation process, is used together with the command efficiency index to judge whether the activity should be marked as an optimization activity or a non-optimization activity. A higher activity fluctuation index indicates that the regulation process of the environmental parameter is accompanied by significant fluctuations or abnormal jumps, which usually means that there is interference or device performance degradation in the execution process.

[0040] It can be understood that the core of the stability test method is to convert the continuous environmental parameter change curve into a quantifiable stability index. By calculating the fluctuation coefficient and the abnormal coefficient, the method evaluates the stationarity of the regulation process from two dimensions of consistency of change and probability of abnormal value occurrence. This method does not rely on a single statistical quantity, and can more comprehensively capture the instability characteristics in the change trajectory, providing a key basis for accurately evaluating the performance state of the environmental execution device.

[0041] It can be understood that the stability test method is an important part of the whole hierarchical performance analysis technology. The active fluctuation index and the command efficiency index complement each other, one focuses on the efficiency of the adjustment process, and the other focuses on the stability of the adjustment process, together forming the basis for comprehensive evaluation of the quality of environmental regulation activities. This data-driven method can effectively identify non-optimal activities that have problems in the process although the final adjustment target is achieved, and improve the sensitivity of system fault early warning.

[0042] In specific implementation, the auxiliary abnormality detection process is started after the system generates a normal signal, serving as a redundant safety check mechanism. The core of the auxiliary abnormality detection is to monitor the running state of the environmental execution device body rather than the adjustment effect of the environmental parameters. The system continuously collects vibration amplitude data and sound pressure level data through special sensors installed on or near the environmental execution device. The vibration amplitude data is obtained by an acceleration sensor, representing the vibration intensity of the mechanical parts of the environmental execution device in operation; the sound pressure level data is obtained by a microphone sensor, quantifying the sound pressure level of the noise generated by the environmental execution device in operation. These data are collected in real time at a high frequency and transmitted to the processing unit.

[0043] In specific implementation, the system compares the real-time collected vibration amplitude data and sound pressure level data with the preset safety limit value. The preset safety limit value is a threshold value preset according to the model, specification and baseline data collected under the long-term normal running state of the environmental execution device. For the vibration amplitude data, there is a preset vibration safety limit value; for the sound pressure level data, there is a preset sound pressure safety limit value. During the entire operation of the environmental execution device, the system continuously monitors: if the collected vibration amplitude data exceeds the preset vibration safety limit value, or the collected sound pressure level data exceeds the preset sound pressure safety limit value, the system immediately determines that the environmental execution device has a potential fault at the current time. The potential fault is a state flag indicating that the device may be in an unhealthy running state.

[0044] In some embodiments, the system continuously tracks the potential fault within a complete monitoring period. The monitoring period is a fixed time period, for example, twenty-four hours. The system accumulates the total duration of all potential faults marked in the monitoring period. At the same time, the system records the total running time of the environmental execution device in the same monitoring period. The system calculates the ratio of the cumulative duration of the potential fault to the total running time of the environmental execution device, and defines this ratio as the fault time proportion. The fault time proportion reflects the proportion of time that the device is in an abnormal running state.

[0045] In some embodiments, the system also counts the number of cases where the potential faults occur in clusters within the monitoring period. The system sets a predetermined number of consecutive sampling periods, for example, three. The system scans the time series data of the entire monitoring period and counts the number of cases where the potential fault state occurs consecutively more than the predetermined number of times. This number of cases is defined as the high-frequency fault count. The high-frequency fault count reveals that the faults are not randomly isolated but occur in a continuous and clustered manner, which can indicate a persistent device defect or external interference.

[0046] In some embodiments, the system also records the longest duration of a single potential fault event within the monitoring period. The specific method is to identify the start and end points of each potential fault event from the time series data, calculate the duration of each potential fault event, and then find the value of the longest duration among all potential fault events. This value is defined as the maximum fault duration. The maximum fault duration reflects the duration of the most severe abnormal state.

[0047] Optionally, after obtaining the fault time ratio, high-frequency fault count, and maximum fault duration, the system calculates a comprehensive fault evaluation value through a weighted summation model. The weighted summation model is represented as: In this model, represents the calculated fault evaluation value. represents the fault time ratio, is the weight assigned to the fault time ratio. represents the high-frequency fault count, is the weight assigned to the high-frequency fault count. represents the maximum fault duration, is the weight assigned to the maximum fault duration. The weights , and are constants pre-set based on historical fault data analysis and expert experience. The values of the weights , and reflect the importance differences of the three indicators on the overall fault evaluation.

[0048] In a specific implementation, after the fault evaluation value is calculated, the system compares it with a preset fault threshold. The preset fault threshold is a key criterion for distinguishing normal redundant signals from abnormal redundant signals. If the fault evaluation value is greater than the preset fault threshold, it indicates that although the main performance evaluation index is normal, the device body state has a high risk, and the system generates an auxiliary abnormal indication. If the fault evaluation value is not greater than the preset fault threshold, it indicates that the device body state is basically normal, and the system generates an auxiliary normal indication. The auxiliary abnormal indication and the auxiliary normal indication constitute the output of the auxiliary abnormal detection.

[0049] It can be understood that the auxiliary abnormal detection process is independent of the main evaluation process based on the environmental adjustment activity performance, and it provides a second guarantee from the perspective of the device physical state. This redundant design enhances the reliability of the system, and can discover potential problems through abnormal characteristics of the device itself when the main process fails to timely alarm due to some reasons, thereby realizing earlier warning and intervention.

[0050] It can be understood that the monitoring of vibration and noise in the auxiliary abnormal detection process directly reflects the mechanical health state of the environmental execution device. The three indexes of fault time proportion, high-frequency fault count, and maximum fault duration respectively depict the total proportion, the aggregation of occurrence, and the severity of a single occurrence of the abnormal state from three dimensions, and then a comprehensive evaluation value is obtained through weighted fusion. This method can more comprehensively and sensitively capture early signs of degradation of the device.

[0051] Referring to Figure 4 This figure quantitatively presents the fault characteristics of different environmental execution devices from the dimensions of fault time proportion, high-frequency fault count, and maximum fault duration, with device number as the horizontal axis and standardized numerical value as the vertical axis, around the auxiliary abnormal detection needs of the green building environmental monitoring system. This figure is a visualization of the auxiliary abnormal detection process, which intuitively distinguishes the severity and characteristics of device faults through the columnar distribution of multiple indexes, provides data support for the decision of generating an auxiliary abnormal indication when the fault evaluation value is greater than the preset threshold, helps operation and maintenance personnel quickly identify high-fault-risk devices, realizes efficient conversion from fault data to abnormal warning, and is a direct representation of the multi-dimensional quantitative fault characteristic technical solution in the auxiliary abnormal detection.

[0052] In a specific implementation, when the auxiliary anomaly detection process generates an auxiliary anomaly indication, the system automatically triggers a deeper device health diagnosis process, which aims to assess the health condition of the environmental execution device from multiple dimensions and generate maintenance decisions. The device health diagnosis process first needs to collect multiple basic data of the environmental execution device. The system obtains the manufacturing date of the environmental execution device, which is usually stored in the device archives or electronic tags. The system calculates the difference between the current date and the manufacturing date of the environmental execution device, and the obtained difference is defined as the device service age, which is quantified in days or years, quantifying the length of time since the environmental execution device was put into use.

[0053] In a specific implementation, the system extracts the cumulative running time of the environmental execution device from a specially designed historical database, which continuously records the start and stop time of each environmental execution device. The cumulative running time is the sum of all running periods, and the obtained cumulative running time is defined as the total running time of the device, which is quantified in hours, reflecting the actual mechanical wear time of the environmental execution device. The system obtains the high-risk exposure time of the environmental execution device through the environmental risk analysis process, which includes collecting the surrounding temperature value and the surrounding humidity value of the installation location of the environmental execution device. The surrounding temperature value and the surrounding humidity value are continuously monitored by intelligent sensing nodes deployed near the environmental execution device. The system calculates the absolute error between the surrounding temperature value and the preset ideal temperature value, which is defined as the temperature deviation amount; the system calculates the absolute error between the surrounding humidity value and the preset ideal humidity value, which is defined as the humidity deviation amount. The system simultaneously collects the concentration of suspended particulate matter at the installation location of the environmental execution device, which is defined as the dust concentration value. The system inputs the temperature deviation amount, humidity deviation amount and dust concentration value into a preset risk assessment function, and the risk assessment function outputs a quantitative environmental risk coefficient. Referring to Table 1, the corresponding relationship between the environmental risk coefficient and the risk level.

[0054] Table 1: Corresponding relationship table between environmental risk coefficient and risk level Range of environmental risk factor Risk level Whether high-risk exposure time is counted 0.0-0.3 Low risk No 0.3-0.7 Medium risk No 0.7-1.0 High risk Yes If the environmental risk coefficient is greater than 0.7, the system determines that the environmental execution device is currently in a high-risk environment, and starts to accumulate the total time when the environmental execution device is in a high-risk environment. This cumulative total time is defined as the high-risk exposure time, which reflects the history of the environmental execution device running in harsh working conditions.

[0055] In a specific implementation, the system also needs to query the historical maintenance information of the environmental execution device from the maintenance record database, which records the execution date of each maintenance. The system compares the actual time interval of the adjacent two maintenances with the standard maintenance interval recommended by the manufacturer of the environmental execution device. If the actual maintenance interval exceeds the standard maintenance interval, it is recorded as a non-standard maintenance event. The system counts the total number of non-standard maintenance events of the environmental execution device in the entire life cycle, and defines this total number as the number of non-standard maintenances, which reflects the timeliness and standardization of preventive maintenance work.

[0056] In some embodiments, the core of the device health diagnosis process is to calculate a comprehensive device health index. The system takes the device service age, total device runtime, high-risk exposure time, and non-standard maintenance number as input features, and inputs them into a pre-trained multi-layer perceptron model. The multi-layer perceptron model is an artificial neural network, which includes an input layer, at least one hidden layer, and an output layer. The input layer contains four neurons corresponding to the four input features: device service age, total device runtime, high-risk exposure time, and non-standard maintenance number. The hidden layer contains multiple neurons, each of which performs weighted summation on the input values and applies a nonlinear activation function. The output layer contains one neuron, whose output value is mapped to a value between 0 and 1 after Sigmoid function activation. This value is the device health index. The closer the device health index is to 1, the better the health status of the environmental execution device; the closer the device health index is to 0, the worse the health status of the environmental execution device.

[0057] Optionally, the last step of the device health diagnosis process is to make a decision based on the device health index. The system has a pre-set device health standard value, which is a threshold value between 0 and 1. The system compares the calculated device health index with the pre-set health standard value. If the device health index is lower than the pre-set health standard value, it indicates that the health status of the environmental execution device has deteriorated to the extent that it needs to be paid attention to, and the system automatically generates a device replacement recommendation. The recommendation can include device identification, current health index, evaluation time, etc., and is displayed through the human-machine interface or sent to the maintenance personnel. If the device health index is not lower than the pre-set health standard value, the process ends and no recommendation is generated.

[0058] Optionally, the training process of the multi-layer perceptron model is offline. The training process requires the use of a large amount of historical data, which contains feature data such as the service age, total runtime, high-risk exposure time, and number of non-standard maintenance of the environmental execution device, and label data indicating whether the device is finally normally scrapped or replaced due to failure. By using optimization methods such as backpropagation algorithm, the weights and bias parameters in the multi-layer perceptron model are continuously adjusted, so that the predicted output of the multi-layer perceptron model is as close to the true label as possible, thereby enabling the multi-layer perceptron model to learn the ability to predict the device health status from the input features.

[0059] It can be understood that the device health diagnosis process comprehensively considers the information of the use time, workload, harshness of the environment, and maintenance history of the environmental execution device in multiple dimensions, and performs nonlinear fusion through the neural network model, thereby being able to more comprehensively and accurately evaluate the potential remaining life and health status of the device. This data-based predictive diagnosis helps to shift the maintenance strategy from post-maintenance to predictive maintenance, avoiding losses caused by sudden failures. The generated device replacement recommendations provide quantitative data support for maintenance decisions.

[0060] Referring to Figure 5 The heat map presents the health impact degree of the six devices from four dimensions of service age, total runtime, high-risk exposure time, and number of non-standard maintenance, around the health diagnosis needs of the environmental execution device in the green building environment monitoring system. The vertical axis lists devices 1 to 6, and the horizontal axis corresponds to the four core influencing factors. The diagram provides an intuitive and quantitative analysis tool for device health diagnosis through multi-dimensional heat distribution: it can quickly identify high-risk devices and low-risk devices, and locate the key influencing factors of each device, helping maintenance personnel to develop targeted maintenance or replacement strategies. It is a visualization of the technical scheme of comprehensive device health diagnosis considering service age, total runtime, high-risk exposure time, and number of non-standard maintenance, and realizes efficient conversion from multi-dimensional data to decision basis.

[0061] Embodiment 5: The process of generating environment control commands by the central processing unit using adaptive learning algorithm is a closed-loop control flow that makes decisions based on real-time data and predetermined criteria. The central processing unit continuously receives environmental parameters uploaded from intelligent sensor nodes distributed in multiple locations inside the green building, including thermodynamic parameters, humidity parameters, and gas concentration parameters. The central processing unit has pre-stored environmental standard values corresponding to different areas and different time periods, such as comfortable room temperature range, ideal humidity level, and safe carbon dioxide concentration upper limit. The first step in generating environment control commands is to calculate the difference, the system compares the real-time collected environmental parameters with the pre-stored environmental standard values, and calculates the difference between the environmental parameters and the pre-stored environmental standard values; the difference can be a simple arithmetic difference, such as the difference between the current temperature and the target temperature, or a relative difference or standardized deviation value.

[0062] In specific implementation, the calculated difference is taken as the main input feature into a pre-trained neural network model. The neural network model is a mathematical model that simulates the connection of human brain neurons for calculation. This specific neural network model operates based on two key pieces of information, one is the training data from the historical data set, which contains a large number of past successful environmental adjustment history cases, which record various environmental parameter deviations and the corresponding verified effective environment control commands; the other is the real-time trend, which is obtained by real-time analysis of the environmental parameter data in the recent time series, such as whether the environmental parameter is in an upward channel, a downward channel, or a stable state. The neural network model performs nonlinear transformation and feature extraction on the input difference and real-time trend through its internal multi-layer network structure, and finally generates specific environment control command parameters in the output layer.

[0063] In a specific implementation, the environment control command output by the neural network model contains two core elements, namely the adjustment type and the adjustment strength. The adjustment type indicates the essential category of the environment adjustment operation that needs to be performed, and the adjustment type corresponds to basic operations such as temperature raising operation, temperature lowering operation, humidity increasing operation, humidity decreasing operation, or air purification operation. The selection of the adjustment type is directly determined by the positive or negative sign of the input difference and the type of the environment parameter to which it corresponds; for example, if the temperature difference is negative (i.e., the current temperature is lower than the standard temperature), the adjustment type is the temperature raising operation; if the humidity difference is positive (i.e., the current humidity is higher than the standard humidity), the adjustment type is the humidity decreasing operation. The adjustment strength quantifies the strength of the environment adjustment operation, and the adjustment strength is represented by a percentage scale, which varies continuously from 0% to 100%, with 0% indicating no adjustment and 100% indicating the maximum capacity of the environment execution device. The specific value of the adjustment strength is determined by factors such as the absolute value of the difference and the slope of the real-time trend, and the larger the absolute value of the difference, the higher the percentage of the adjustment strength required.

[0064] In some embodiments, the structure of the neural network model is designed to have one input layer, several hidden layers, and one output layer. The number of input layer neurons matches the number of input features, which at least includes the difference of each environment parameter. The hidden layer can be one or more layers, each hidden layer contains multiple neurons, each neuron performs weighted summation on all its inputs and applies a nonlinear activation function such as ReLU function or Sigmoid function. The number of output layer neurons is consistent with the number of control parameters that need to be generated, for example, the output layer can have two neurons, one neuron's output value is processed by the Softmax function and mapped to a discrete category representing different adjustment types, and the other neuron's output value is mapped to the interval of 0% to 100% to represent the adjustment strength.

[0065] In some embodiments, the training process of the neural network model is crucial. The training process is performed using a historical data set containing a large number of labeled samples, each sample includes a set of input conditions (such as historical environment parameter difference, historical trend data) and a set of expected output results (i.e., the optimal environment control command taken in the historical record under the input condition, including the correct adjustment type and adjustment strength). The training algorithm (such as the backpropagation algorithm) continuously adjusts the connection weights and bias terms between neurons in the neural network model through iterative methods, so that the error (loss function) between the predicted output of the neural network model and the expected output for the training sample is minimized. Through sufficient training, the neural network model can learn the mapping relationship from the complex environment state to the optimal control action.

[0066] Optionally, an important feature of the adaptive learning algorithm is that the model can be fine-tuned online. In addition to offline training with historical data, the system can also make small adjustments to the parameters of the neural network model based on real-time control effect feedback. For example, if a certain environmental control command causes the environmental parameters to approach the standard value too quickly and excessively (overshoot), the system may record this suboptimal regulation and use it as a new training sample to optimize the weights during the next model parameter update, resulting in smoother subsequent commands, which reflects the adaptive ability of the algorithm.

[0067] Optionally, the generation frequency of the environmental control command can be set as needed. The system can be set to generate at regular intervals, such as calculating the difference amount and generating a command every five minutes; or it can be set to be event-driven, that is, when the difference amount of a certain environmental parameter exceeds a certain threshold, the neural network model is triggered to calculate and generate a new environmental control command immediately. This flexible triggering mechanism ensures the timeliness of control.

[0068] Optionally, for complex situations where multiple environmental parameters need to be adjusted, the neural network model needs to have multi-objective decision-making ability. For example, when the temperature and humidity deviate from the standard value at the same time, the neural network model needs to consider comprehensively, and its output may be a combined command that contains both the judgment of adjustment type (such as prioritizing temperature reduction or dehumidification) and the adjustment intensity allocation for the coordination of multiple environmental execution devices. This requires training data to include cases of such complex scenarios, and the neural network model to have a complex enough structure to learn these advanced strategies.

[0069] It can be understood that the method of generating environmental control commands by the central processing unit using an adaptive learning algorithm upgrades the traditional fixed rule-based controller to an intelligent decision center with learning and adaptive capabilities. By processing the difference between the environmental parameters and the standard value through a neural network model and considering real-time trends, the system can generate more accurate and efficient environmental control commands, thereby ensuring environmental comfort while optimizing energy consumption and improving the overall intelligent operation level of green buildings. This method enables the environmental control system to better cope with changes in building usage patterns, external weather disturbances, and other uncertain factors.

[0070] Although embodiments of the present application have been shown and described, it will be understood by those having ordinary skill in the art that various changes, modifications, replacements, and variations can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A green building environment monitoring method, characterized in that, The method comprises: continuously collecting environmental parameters including thermodynamic parameters, humidity parameters and gas concentration parameters through intelligent sensing nodes distributed in multiple locations inside a green building; sending the environmental parameters to a central processing unit, which generates environmental control commands using an adaptive learning algorithm based on pre-stored environmental standard values and real-time data streams; environmental execution devices perform environmental adjustment activities according to the environmental control commands; monitoring the execution sequence of the environmental adjustment activities in real time, calculating performance evaluation indicators, and generating performance status signals based on the performance evaluation indicators, the performance status signals including normal signals and abnormal signals; triggering an alarm process when the performance status signal is an abnormal signal.

2. A green building environment monitoring method as claimed in claim 1, wherein, The process of calculating performance evaluation indicators and generating performance status signals based on the performance evaluation indicators comprises: applying a hierarchical performance analysis technique to divide the environmental adjustment activities into optimized activities or non-optimized activities; calculating the ratio of the frequency of non-optimized activities to the total number of activities within a fixed monitoring period to obtain a non-optimized ratio; if the non-optimized ratio is greater than a pre-set non-optimized critical value, output an abnormal signal; if the non-optimized ratio is not greater than the pre-set non-optimized critical value, measure the delay time and energy consumption offset of the environmental adjustment activities; normalizing and fusing the delay time and energy consumption offset to obtain an overall performance score; if the overall performance score exceeds a pre-set performance limit, output an abnormal signal, otherwise output a normal signal.

3. A green building environment monitoring method as claimed in claim 2, wherein, The process of applying a hierarchical performance analysis technique to divide the environmental adjustment activities into optimized activities or non-optimized activities comprises: record the time point when the central processing unit sends the environmental control command as the starting time, and record the time point when the environmental execution device completes the environmental adjustment activity as the ending time, the time interval between the starting time and the ending time is defined as the monitoring interval; the ratio of the environmental control command to the monitoring interval is taken as the command efficiency index, and the activity fluctuation index is obtained through a stability test method; if the command efficiency index is not within the pre-defined efficiency range or the activity fluctuation index is higher than the pre-defined fluctuation upper limit, the environmental adjustment activity is marked as a non-optimized activity; if the command efficiency index is within the pre-defined efficiency range and the activity fluctuation index is not higher than the pre-defined fluctuation upper limit, the environmental adjustment activity is marked as an optimized activity.

4. A green building environment monitoring method as claimed in claim 3, wherein, The process of the stability test method comprises: constructing a two-dimensional coordinate system in the time dimension and the environmental parameter value dimension to obtain the environmental parameter change trajectory during the environmental adjustment activity; setting multiple sampling points on the environmental parameter change trajectory, the environmental parameter difference between adjacent sampling points is recorded as the fluctuation amplitude, the standard deviation of all fluctuation amplitudes is calculated as the fluctuation coefficient, and the proportion of fluctuation amplitudes exceeding the pre-defined fluctuation interval is calculated as the abnormal coefficient; the activity fluctuation index is derived through linear combination of the fluctuation coefficient and the abnormal coefficient.

5. The green building environment monitoring method of claim 1, wherein, The method further comprises performing auxiliary abnormal detection when a normal signal is generated, the process of auxiliary abnormal detection comprises: The vibration amplitude data and the sound pressure level data are collected during the operation of the environmental execution device, and if the vibration amplitude data or the sound pressure level data exceeds the respective preset safety limit, it is determined that the environmental execution device has a potential fault; The duration of the potential fault is accumulated in a monitoring period, and the ratio of the duration to the total operation time of the environmental execution device is calculated to obtain a fault time ratio; The number of cases in which the number of consecutive occurrences of the potential fault exceeds a predetermined number in the monitoring period is counted as a high-frequency fault count; The longest duration of a single potential fault in the monitoring period is recorded as a maximum fault duration; The fault evaluation value is obtained by weighted summation of the fault time ratio, the high-frequency fault count, and the maximum fault duration; If the fault evaluation value is greater than a preset fault threshold, an auxiliary abnormal indication is generated; if the fault evaluation value is not greater than the preset fault threshold, an auxiliary normal indication is generated.

6. A green building environment monitoring method as claimed in claim 5, wherein, The process of the auxiliary abnormality detection further includes: When the auxiliary abnormal indication is generated, further device health diagnosis is performed, and the process of the device health diagnosis includes: Obtaining the manufacturing date of the environmental execution device, calculating the difference between the current date and the manufacturing date to obtain the service age of the device; Extracting the cumulative operation time of the environmental execution device from the historical database to obtain the total operation time of the device; Obtaining the high-risk exposure time of the environmental execution device through environmental risk analysis; Querying the number of times that the maintenance interval of the environmental execution device exceeds the standard maintenance interval from the maintenance record as the number of non-standard maintenance; Comprehensively calculating the device health index through the multilayer perceptron model based on the service age of the device, the total operation time of the device, the high-risk exposure time, and the number of non-standard maintenance; If the device health index is lower than a preset health standard, a device replacement suggestion is generated.

7. A green building environment monitoring method as claimed in claim 6, wherein, The process of the environmental risk analysis includes: Collecting the ambient temperature value and the ambient humidity value of the installation position of the environmental execution device, calculating the absolute error between the ambient temperature value and the ideal temperature value as the temperature deviation, and calculating the absolute error between the ambient humidity value and the ideal humidity value as the humidity deviation; Collecting the concentration of suspended particulate matter at the installation position of the environmental execution device as the dust concentration value; Inputting the temperature deviation, the humidity deviation, and the dust concentration value into a risk assessment function to output an environmental risk coefficient; If the environmental risk coefficient is greater than a preset risk standard, it is determined that the environmental execution device is in a high-risk environment, and the total time of the high-risk environment is accumulated as the high-risk exposure time.

8. A green building environment monitoring method as claimed in claim 1, wherein, The process of generating the environmental control command by the central processing unit using the adaptive learning algorithm includes: Inputting the difference between the environmental parameters and the pre-stored environmental standard values into a neural network model, and the neural network model generates an environmental control command based on training data and real-time trends, the environmental control command including an adjustment type and an adjustment intensity; The adjustment type corresponds to a temperature raising operation, a temperature lowering operation, a humidity increasing operation, a humidity decreasing operation, or an air purification operation, and the adjustment intensity is represented by a percentage scale.

9. A green building environment monitoring system comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor, when executing the computer program, implements the steps of the green building environment monitoring method of any one of claims 1 to 8.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the green building environment monitoring method of any one of claims 1 to 8.

Images