Community water purification equipment cluster management system based on LoRa networking

Abstract

The present application relates to water purification equipment management technical field, especially in community water purification equipment cluster management system based on LoRa networking, including: water quality data acquisition module, for obtaining the historical water quality data and real-time water quality data of water purification equipment in target area;Water quality situation prediction module, for predicting the water quality change trend of future period according to the historical water quality data and real-time water quality data;Filter core loss analysis module, for comparing real-time water quality data, historical water quality data and preset grading standard, determining the filter core loss coefficient corresponding to current water quality, combining the cumulative total water volume of filter core after installation to calculate the actual loss amount of filter core, and comparing with the rated total loss amount of filter core, completing the detection of actual use loss of filter core, the system can accurately detect filter core loss combined with actual water quality, avoid excessive waste or filter core failure problem caused by fixed period replacement, both reduce operation and maintenance cost, and ensure filter core filtering effect.

Description

Technical Field

[0001] This invention relates to the field of water purification equipment management technology, specifically a community water purification equipment cluster management system based on LoRa networking. Background Technology

[0002] The LoRa-based community water purification equipment cluster management system is an intelligent water purification management solution adapted to community scenarios. Relying on the low power consumption, wide coverage, and high reliability characteristics of LoRa networking, it connects multiple water purification devices deployed in various areas of the community, collects data such as water quality and equipment operation through various sensors, and combines relevant modules to realize functions such as water quality monitoring and filter management, providing safe and stable drinking water for community residents. It has been widely used in various residential communities.

[0003] Currently, in existing community water purification equipment cluster management technologies, filter replacement generally adopts a periodic replacement model. This model has obvious drawbacks: it does not take into account the actual water quality and equipment usage load to detect the actual wear and tear of the filter, but only forces replacement according to a fixed cycle. This easily leads to excessive waste of filter cartridges before they reach their actual service life, increasing operation and maintenance costs. At the same time, the filter cartridges may fail prematurely due to water quality deterioration, but they may not be replaced in time, resulting in a decrease in filtration effect and a lack of drinking water safety. It cannot adapt to the dynamic changes in different water quality scenarios and makes it difficult to achieve refined and intelligent management of filter cartridges. Summary of the Invention

[0004] The purpose of this invention is to provide a community water purification equipment cluster management system based on LoRa networking to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] The community water purification equipment cluster management system based on LoRa networking includes: a water quality data acquisition module, which is used to acquire historical and real-time water quality data of water purification equipment in the target area;

[0007] The water quality trend prediction module is used to predict the water quality change trend in the future based on the historical water quality data and real-time water quality data.

[0008] The filter cartridge wear analysis module compares real-time water quality data, historical water quality data, and preset grading standards to determine the filter cartridge wear coefficient corresponding to the current water quality. It calculates the actual wear of the filter cartridge based on the total cumulative water volume filtered since installation and compares it with the rated total wear of the filter cartridge to complete the detection of the actual wear of the filter cartridge. The filter cartridge life assessment module calculates and determines the actual remaining life of the filter cartridge based on the predicted future water quality change trend and the actual wear status of the filter cartridge.

[0009] Preferably, the water quality trend prediction module includes: a feature extraction unit, used to extract time-series features from historical and real-time water quality data using a long short-term memory network model, capturing the gradual change patterns and abrupt changes of water quality parameters; and a prediction output unit, used to analyze the extracted features in conjunction with a preset water quality warning threshold, output water quality prediction data and abnormal risk levels for future periods, and complete the prediction of water quality change trends for future periods.

[0010] Preferred parameters of the Long Short-Term Memory (LSTM) network model are set as follows: the input layer dimension is set to 4, corresponding to the four core water quality parameters: TDS value, turbidity, pH value, and residual chlorine content; the hidden layer is set to 2 layers, with 64 neurons in each layer; the forget gate threshold is set to 0.7, the input gate threshold is set to 0.3, and the output gate threshold is set to 0.5; the time step is set to 24, adapting to the low-frequency data transmission characteristics of LoRa networking; the training batch size is 32, and the number of iterations is 100 rounds; the Adam optimizer is used, the learning rate is set to 0.001, and the mean squared error loss function is selected.

[0011] Preferably, the feature extraction process of the feature extraction unit is as follows: first, the real-time data and historical data are standardized and preprocessed to automatically remove abnormal data; then, the core features of water quality parameters are extracted, and key features are screened in combination with related data to provide support for the prediction output unit.

[0012] Preferably, the filter cartridge wear analysis module includes: a wear coefficient determination unit, used to compare real-time water quality data, historical water quality data and preset grading standards to determine the filter cartridge wear coefficient corresponding to the current water quality, wherein the preset grading standards include water quality impurity content grading standards and water quality fluctuation amplitude grading standards, and the wear coefficient is positively correlated with impurity content and fluctuation amplitude; and a wear amount detection unit, used to collect the cumulative total water volume filtered by the filter cartridge since installation, calculate the actual wear amount of the filter cartridge in combination with the determined filter cartridge wear coefficient, and compare it with the rated total wear amount of the filter cartridge to complete the detection and feedback of the actual wear of the filter cartridge.

[0013] Preferably, the filter life assessment module includes: a water consumption prediction unit, used to determine the future average daily water consumption, wherein the future average daily water consumption is determined by taking the average value after removing abnormal days, based on the recent actual daily water consumption of each water purification device;

[0014] The lifespan calculation unit combines future water quality forecast data, actual filter cartridge wear status, and rated filter cartridge capacity with the average daily water consumption to calculate the actual remaining lifespan of the filter cartridge. The calculation results are accurate to the day and can be dynamically adjusted according to future water quality trends.

[0015] Preferably, it also includes a water quality anomaly handling module, which is used to generate corresponding water quality anomaly early warning information and handling methods based on the water quality change trend and anomaly risk level output by the water quality situation prediction module. The handling methods include pushing early warning prompts to the administrator terminal, sending shutdown commands or automatic flushing commands to the corresponding water purification equipment.

[0016] Preferred: It also includes a regional anomaly collaborative judgment module, which is used to summarize the water quality prediction data of all water purification equipment and identify equipment clusters in the same area based on the deployment location information of water purification equipment; when multiple water purification equipment in the same area are predicted to have water quality anomalies, it is determined to be a regional water supply problem, and regional location, number of abnormal equipment and summary information of abnormal parameters are added to the early warning prompt to support the overall scheduling of operation and maintenance resources.

[0017] Compared with existing technologies, the beneficial effects of this invention are as follows: This system is based on LoRa networking and includes four major modules: water quality data acquisition, water quality trend prediction, filter element wear analysis, and filter element life assessment. By collecting historical and real-time water quality data, it detects the actual wear of the filter element, predicts future water quality, and then calculates the remaining lifespan of the filter element, realizing intelligent management of community water purification equipment clusters. Compared with the traditional method of periodically replacing filter elements, this system can accurately detect filter element wear based on actual water quality conditions, avoiding excessive waste or filter element failure caused by fixed-cycle replacement. This reduces maintenance costs, ensures filter element filtration effect, protects drinking water safety, and adapts to the usage needs of different water quality scenarios.

[0018] Meanwhile, this system can predict future water quality changes based on historical and real-time water quality data, anticipate water quality anomalies in advance, solve the pain point of untimely water quality monitoring in traditional technologies, facilitate administrators to prepare in advance, prevent unqualified drinking water from flowing out, and enhance the initiative in managing community water purification equipment.

[0019] In addition, this system calculates the remaining service life by combining future water quality trends with the actual wear and tear of the filter cartridges. It can dynamically correct the service life data, accurately control the timing of filter cartridge replacement, avoid water quality impact from untimely replacement or waste caused by excessive replacement, realize intelligent management of filter cartridges, and improve operation and maintenance efficiency. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the first part of the water purification equipment cluster management system in an embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of the water quality status prediction module in an embodiment of the present invention.

[0022] Figure 3 This is a schematic diagram of the filter element loss analysis module in an embodiment of the present invention.

[0023] Figure 4This is a schematic diagram of the water quality status prediction module in an embodiment of the present invention.

[0024] Figure 5 This is a schematic diagram of the second part of the water purification equipment cluster management system in an embodiment of the present invention.

[0025] Figure 6 This is a schematic diagram of the third part of the water purification equipment cluster management system in an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0027] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0028] In one embodiment, see Figure 1 A community water purification equipment cluster management system based on LoRa networking includes: a water quality data acquisition module 1, used to acquire historical water quality data and real-time water quality data of water purification equipment in the target area;

[0029] Water quality trend prediction module 2 is used to predict the water quality change trend in the future period based on the historical water quality data and real-time water quality data.

[0030] The filter cartridge wear analysis module 3 is used to compare real-time water quality data, historical water quality data and preset grading standards to determine the filter cartridge wear coefficient corresponding to the current water quality. It also calculates the actual wear of the filter cartridge by combining the total amount of water filtered since the filter cartridge was installed, and compares it with the rated total wear of the filter cartridge to complete the detection of the actual wear of the filter cartridge. The filter cartridge life assessment module 4 is used to calculate and determine the actual remaining life of the filter cartridge based on the predicted future water quality change trend and the actual wear status of the filter cartridge.

[0031] In this embodiment, the system is applied to a community water purification equipment cluster scenario. The target area is various residential communities, including high-rise buildings, multi-story residences, and public areas of parks. Multiple water purification devices are deployed in the above-mentioned areas. Each water purification device serves as a terminal sensing node of the system, equipped with a water quality sensor, a flow sensor, and a LoRa communication module. Data interaction with various functional modules is achieved through LoRa networking, ensuring low power consumption, wide coverage, and high reliability of data transmission. This is suitable for scenarios where community equipment is dispersed and wireless signals are weak in some areas. The LoRa networking adopts a star topology. The LoRa communication module of each water purification device establishes a communication connection with the LoRa gateway deployed in the community. After the LoRa gateway aggregates the data from all terminal devices, it transmits the data to various functional modules of the system through a broadband network, realizing a closed-loop data process of "terminal acquisition - network transmission - module processing". The core function of the water quality data acquisition module 1 is to acquire historical and real-time water quality data of all water purification devices within the target area, providing accurate data support for subsequent water quality prediction, filter wear analysis, and lifespan assessment. The data acquisition process relies on LoRa networking: the water quality sensor of each water purification device collects core water quality parameters in real time, while the flow sensor simultaneously collects water consumption data during device operation. This real-time data is transmitted to the LoRa gateway via the LoRa communication module on the water purification device in a low-frequency, low-power manner. The LoRa gateway then pushes the aggregated real-time data to the water quality data acquisition module 1. Simultaneously, this module has a built-in data storage unit that pre-stores historical water quality data for at least one complete four-season cycle, device operation data (such as filter replacement records, device start-up / shutdown times, and water production duration), and environmental data (such as community temperature and rainfall). This historical data can be retrieved from each water purification device terminal via LoRa networking or manually entered by the administrator, ensuring data integrity and continuity to meet the needs of subsequent time-series analysis and prediction. The water quality trend prediction module 2, based on historical and real-time water quality data acquired by the water quality data acquisition module 1, predicts future water quality trends, providing a basis for subsequent filter cartridge life assessment and water quality anomaly handling. Its data input relies entirely on the water quality data acquisition module 1, combining real-time data transmitted via LoRa networking with pre-stored historical data. This data is then analyzed using a built-in algorithm model to capture the changing patterns of water quality parameters, ultimately outputting the water quality trend for future periods (preferably 1-7 days), including the fluctuation range of water quality parameters and the presence of any abnormal risks. This allows for advance prediction of water quality, preventing the outflow of substandard drinking water. The filter cartridge wear analysis module 3's core function is to combine water quality data and preset grading standards to detect the actual wear and tear of the filter cartridges, providing wear data support for filter cartridge life assessment.Its working process is as follows: First, it receives real-time water quality data and historical water quality data transmitted by the water quality data acquisition module 1, compares the above data with the preset grading standards, and determines the filter element wear coefficient corresponding to the current water quality; Second, it obtains the cumulative total water volume filtered by each water purifier filter element since installation through LoRa networking (collected in real time by the flow sensor of the water purifier and transmitted to this module via the LoRa communication module and LoRa gateway); Finally, it calculates the actual wear of the filter element by combining the filter element wear coefficient and the cumulative total water volume filtered, and compares the actual wear with the rated total wear of the filter element to accurately detect the actual wear status of the filter element. For example, when the actual wear of the filter element reaches 80% of the rated total wear, it is determined to be severely worn and needs to be promptly reported to the filter element life assessment module 4. The core function of the filter cartridge life assessment module 4 is to calculate and determine the actual remaining lifespan of the filter cartridge based on the future water quality change trend output by the water quality trend prediction module 2 and the actual usage and wear status of the filter cartridge detected by the filter cartridge wear analysis module 3. This enables intelligent management of the filter cartridge and avoids untimely or excessive replacement. Its operation relies on data support from the first three modules: receiving the future water quality change trend output by the water quality trend prediction module 2 to determine the impact of future water quality on filter cartridge wear; receiving the actual usage and wear status of the filter cartridge output by the filter cartridge wear analysis module 3 to obtain the current wear level of the filter cartridge; and combining parameters such as the rated filtration capacity of the filter cartridge and the average daily water consumption in the future, calculating the remaining lifespan of the filter cartridge through a preset algorithm to ensure that the calculation results accurately match the actual usage scenario. At the same time, the lifespan data can be dynamically corrected according to future water quality change trends to improve the accuracy of lifespan prediction. This system, based on LoRa networking, comprises four main modules: water quality data acquisition, water quality trend prediction, filter cartridge wear analysis, and filter cartridge life assessment. By collecting historical and real-time water quality data, it detects the actual wear of the filter cartridges, predicts future water quality, and calculates the remaining lifespan of the filter cartridges. This enables intelligent management of community water purification equipment clusters. Compared to the traditional method of periodically replacing filter cartridges, this system can accurately detect filter cartridge wear based on actual water quality conditions, avoiding excessive waste or filter cartridge failure caused by fixed-cycle replacement. This reduces maintenance costs while ensuring filter cartridge filtration effectiveness, protecting drinking water safety, and adapting to the usage needs of different water quality scenarios.

[0032] Meanwhile, this system can predict future water quality changes based on historical and real-time water quality data, anticipate water quality anomalies in advance, solve the pain point of untimely water quality monitoring in traditional technologies, facilitate administrators to prepare in advance, prevent unqualified drinking water from flowing out, and enhance the initiative in managing community water purification equipment.

[0033] In addition, this system calculates the remaining service life by combining future water quality trends with the actual wear and tear of the filter cartridges. It can dynamically correct the service life data, accurately control the timing of filter cartridge replacement, avoid water quality impact from untimely replacement or waste caused by excessive replacement, realize intelligent management of filter cartridges, and improve operation and maintenance efficiency.

[0034] In another embodiment, please refer to Figure 2 The water quality trend prediction module 2 includes: a feature extraction unit 21, used to extract time-series features from historical water quality data and real-time water quality data through a long short-term memory network model, capturing the gradual change pattern and abrupt change characteristics of water quality parameters; and a prediction output unit 22, used to analyze the extracted features in combination with a preset water quality warning threshold, output water quality prediction data and abnormal risk level for future periods, and complete the prediction of water quality change trends for future periods.

[0035] In this embodiment, the module includes two core units: a feature extraction unit 21 and a prediction output unit 22. The two units work together to predict the trend of water quality changes in the future. The data transmission of the two units is achieved through LoRa networking to ensure the real-time performance and reliability of the data.

[0036] The core function of the feature extraction unit 21 is to extract time-series features from historical and real-time water quality data acquired by the water quality data acquisition module 1 using a Long Short-Term Memory (LSTM) network model, capturing the gradual change patterns and abrupt changes in water quality parameters. The specific implementation process is as follows: First, it receives real-time water quality data (such as TDS values ​​and turbidity) transmitted via LoRa networking and retrieves historical water quality data, performing preliminary processing on the data. Second, it activates the built-in LSTM network model, which possesses powerful time-series data mining capabilities and can accurately process massive amounts of historical and real-time water quality time-series data. Through the model's hidden layers, it extracts the core features of water quality parameters, including the rate of change (such as the daily increase / decrease in TDS values), fluctuation range (such as the daily difference between the maximum and minimum deviations in turbidity), and peak features (such as the daily maximum and minimum values ​​of water quality parameters). Simultaneously, it captures the gradual change patterns (such as slow fluctuations in water quality due to seasonal changes) and abrupt changes (such as sudden deterioration of water quality due to pipeline construction or municipal water supply adjustments), providing accurate feature support for subsequent prediction output.

[0037] The core function of the prediction output unit 22 is to combine the preset water quality early warning threshold, analyze the features extracted by the feature extraction unit 21, output the water quality prediction data and abnormal risk level for the future period, and complete the prediction of the water quality change trend for the future period. The specific implementation process is as follows: First, the filtered water quality feature data transmitted by the feature extraction unit 21 is received; second, the preset water quality warning threshold is called (this threshold can be set differently according to the community water quality standards, seasonal changes, and the type of water purification equipment, and is entered by the administrator through the backend, for example, the turbidity warning threshold is set to 5 NTU in summer and 3 NTU in winter); then, the feature data is compared and analyzed with the water quality warning threshold to determine the trend of water quality parameters in the future period and predict the water quality data (such as TDS value, turbidity, etc.) for the future days; finally, the abnormal risk level is output according to the comparison results (divided into two levels: general risk and emergency risk). General risk refers to slight fluctuations in water quality that will not affect drinking safety in the short term, while emergency risk refers to serious exceedances of water quality standards that may endanger the drinking safety of residents; the prediction results are transmitted to the filter life assessment module 4 and the subsequent water quality anomaly handling module 5 through the system, and can also be fed back to the administrator through LoRa networking, so that the administrator can grasp the water quality situation in advance.

[0038] In another embodiment, the parameters of the Long Short-Term Memory (LSTM) network model are set as follows: the input layer dimension is set to 4, corresponding to the four core water quality parameters of TDS, turbidity, pH, and residual chlorine content; the hidden layer is set to 2 layers, with 64 neurons in each layer; the forget gate threshold is set to 0.7, the input gate threshold is set to 0.3, and the output gate threshold is set to 0.5; the time step is set to 24 to adapt to the low-frequency data transmission characteristics of LoRa networking; the training batch size is 32, and the number of iterations is 100 rounds; the Adam optimizer is used, the learning rate is set to 0.001, and the loss function is the mean squared error.

[0039] In this embodiment, the Long Short-Term Memory (LSTM) network model used by the feature extraction unit 21 has had its parameters optimized through actual testing to adapt to the low-frequency data transmission characteristics of LoRa networking (LoRa networking transmits real-time data once per hour, and 24 sets of data per day), while also taking into account prediction accuracy and system computing cost. The specific parameter settings and implementation logic are as follows:

[0040] 1. Input layer dimension: set to 4, corresponding to the four core water quality parameters: TDS value, turbidity, pH value, and residual chlorine content. This setting matches the core parameters of subsequent water quality data, which avoids redundant data from increasing the computational power consumption of the model and ensures that the core water quality characteristics are not omitted, so as to comprehensively reflect the water quality status.

[0041] 2. Hidden layer setting: Set to 2 layers, with 64 neurons in each layer. This setting balances prediction accuracy and computing cost, adapts to the conventional computing power of cloud systems, avoids prediction delays caused by overly complex model structures, and can accurately capture the gradual change patterns and abrupt changes of water quality parameters in the next 1-7 days, meeting the water quality prediction needs of this system.

[0042] 3. Gating threshold settings: The forget threshold is set to 0.7 to prioritize the retention of effective trend features in historical water quality data (such as water quality fluctuation patterns caused by seasonal changes) and filter out short-term irrelevant fluctuations (such as accidental data transmission deviations); the input threshold is set to 0.3 to filter and retain effective features in real-time data and avoid invalid data interfering with model predictions; the output threshold is set to 0.5 to accurately output prediction results and ensure the reliability of prediction data.

[0043] 4. Time step: Set to 24 to adapt to the low-frequency data transmission characteristics of LoRa networking - LoRa networking transmits 24 sets of real-time data per day (once per hour). With the time step set to 24, it can fully capture the daily water quality change pattern, provide sufficient time-series data support for the model, and ensure that the model can accurately mine the time-series change characteristics of water quality parameters.

[0044] 5. Training Parameter Settings: The batch size is 32, the number of iterations is 100, the Adam optimizer is used (learning rate set to 0.001), and the mean squared error (MSE) loss function is selected. Actual testing shows that with these parameter settings, the model can stably converge within 100 iterations, with the prediction error controlled within 5% and the prediction accuracy not less than 96%, fully meeting the accuracy requirements for community water quality prediction. Furthermore, these parameter settings are compatible with the computing power of the cloud system, avoiding prediction delays caused by insufficient computing power and ensuring that prediction results can be output to subsequent modules in a timely manner.

[0045] In addition, before deployment, the Long Short-Term Memory Network model needs to be trained using historical water quality data (at least one complete four-season cycle) obtained by the water quality data acquisition module 1. During the training process, the parameters are continuously optimized to ensure that the model can adapt to the prediction needs of different seasons and different water quality states. At the same time, it can automatically adapt to the data characteristics of LoRa network transmission to avoid the impact of data transmission delay and data deviation on the model's prediction accuracy.

[0046] In another embodiment, the feature extraction process of the feature extraction unit 21 is as follows: first, the real-time data and historical data are standardized and preprocessed to automatically remove abnormal data; then, the core features of water quality parameters are extracted, and key features are screened in combination with related data to provide support for the prediction output unit 22.

[0047] In this embodiment, the feature extraction process of the feature extraction unit 21 is the core step in achieving accurate water quality prediction. Its complete process is divided into three stages, each of which is optimized based on the data source characteristics of this system (real-time data and historical data transmitted via LoRa networking), as follows:

[0048] 1. Data Standardization Preprocessing Stage: First, the system receives real-time data and retrieves historical data transmitted via LoRa networking from water quality data acquisition module 1. Due to the differences in the magnitude of different water quality parameters (e.g., TDS values ​​of 0-500 ppm and turbidity of 0-10 NTU), all data needs to be standardized and preprocessed to convert parameters of different magnitudes into a unified magnitude (e.g., converting them into values ​​between 0 and 1), eliminating the influence of dimensions and ensuring that the model can accurately analyze the changing patterns of various parameters. Second, abnormal data is automatically removed. Abnormal data mainly includes three categories: first, abnormal data caused by equipment failure (e.g., a sudden spike in TDS values ​​to over 1000 ppm due to sensor failure in water purification equipment); second, abnormal data caused by abnormal data transmission (e.g., data loss or data deviation during LoRa networking transmission); and third, abnormal data caused by sudden external factors (e.g., short-term water quality changes caused by municipal water supply improvements or pipeline construction; this type of data does not belong to the normal water quality change trend and needs to be removed). By removing abnormal data, the accuracy of feature extraction is avoided.

[0049] 2. Core Feature Extraction Stage: After data preprocessing, the core features of the preprocessed time-series data are extracted through the hidden layer of the Long Short-Term Memory (LSTM) network model. The focus is on extracting three core features of water quality parameters: first, the rate of change, i.e., the daily increase or decrease in various water quality parameters, reflecting the trend of water quality change; second, the fluctuation range, i.e., the difference between the daily maximum and minimum deviations of various water quality parameters, reflecting the stability of water quality; and third, peak features, i.e., the daily maximum and minimum values ​​of various water quality parameters, reflecting the extreme states of water quality. Simultaneously, correlation features between various water quality parameters (such as the positive correlation between TDS values ​​and turbidity) are extracted to support subsequent key feature selection.

[0050] 3. Key Feature Screening Stage: After the core features are extracted, key features affecting water quality in the future are screened by combining related data (including equipment operation data and environmental data), and irrelevant features are eliminated. For example, based on rainfall data in the environmental data, the key feature "turbidity fluctuation amplitude after rainfall" is screened (rainfall will cause turbidity to increase, affecting future water quality); based on water production time in the equipment operation data, irrelevant features "water production time is not related to water quality parameters" are eliminated; after screening, the key features are transmitted to the prediction output unit 22 to provide accurate support for water quality prediction in the future and ensure the accuracy of the prediction results.

[0051] In another embodiment, please refer to Figure 3 The filter cartridge wear analysis module 3 includes: a wear coefficient determination unit 31, used to compare real-time water quality data, historical water quality data and preset grading standards to determine the filter cartridge wear coefficient corresponding to the current water quality. The preset grading standards include water quality impurity content grading standards and water quality fluctuation amplitude grading standards. The wear coefficient is positively correlated with impurity content and fluctuation amplitude; and a wear amount detection unit 32, used to collect the cumulative total water volume filtered by the filter cartridge since installation, calculate the actual wear amount of the filter cartridge based on the determined filter cartridge wear coefficient, and compare it with the rated total wear amount of the filter cartridge to complete the detection and feedback of the actual wear of the filter cartridge.

[0052] In this embodiment, the filter cartridge wear analysis module 3 is internally divided into a wear coefficient determination unit 31 and a wear amount detection unit 32. The two units work together to detect the actual wear of the filter cartridge. The data input comes from the water quality data acquisition module 1 (transmitted via LoRa network). The specific implementation details are as follows:

[0053] 1. Loss Coefficient Determination Unit 31: The core function of this unit is to compare real-time water quality data, historical water quality data, and preset grading standards to determine the filter cartridge loss coefficient corresponding to the current water quality, providing core parameters for calculating the actual loss of the filter cartridge. The specific implementation process is as follows: First, it receives real-time water quality data (TDS value, turbidity, pH value, residual chlorine content, etc.) transmitted by the water quality data acquisition module 1 through LoRa networking, as well as retrieved historical water quality data; second, it calls the preset grading standards, which include water quality impurity content grading standards and water quality fluctuation amplitude grading standards. Among them, the water quality impurity content grading standards are divided into 3 levels (pure, slightly polluted, heavily polluted) based on TDS value and turbidity, and the water quality fluctuation amplitude grading standards are divided into 2 levels (small fluctuation, large fluctuation) based on the deviation of daily water quality parameters from historical benchmark values; then, it compares the real-time and historical data... Historical water quality data is compared with the aforementioned grading standards to determine the current water quality's impurity content level and fluctuation range level. Finally, the corresponding filter cartridge wear coefficient is matched according to the level, following the principle that "the higher the impurity content and the greater the fluctuation range, the greater the filter cartridge wear coefficient." For example: pure level + small fluctuation, the wear coefficient is 1.0; slightly polluted level + small fluctuation, the wear coefficient is 1.3; slightly polluted level + large fluctuation, the wear coefficient is 1.5; heavily polluted level + large fluctuation, the wear coefficient is 2.0. At the same time, the wear coefficient is dynamically adjusted based on the changing trends of historical water quality data to ensure that the wear coefficient accurately matches the actual water quality conditions.

[0054] 2. Loss Detection Unit 32: The core function of this unit is to collect the cumulative total water volume filtered since the filter element was installed, combine it with the filter element loss coefficient determined by the loss coefficient determination unit 31, calculate the actual loss of the filter element, and compare it with the rated total loss of the filter element to complete the detection and feedback of the actual loss of the filter element. The specific implementation process is as follows: First, the total cumulative water volume filtered by each water purifier filter cartridge since installation is obtained through LoRa networking. The flow sensor of each water purifier collects water flow data in real time, which is transmitted to the unit via LoRa communication module and LoRa gateway. The unit automatically accumulates and records the total water volume filtered by the filter cartridge since installation, with calculation accuracy down to the liter. Second, the filter cartridge loss coefficient output by unit 31 is determined by combining the loss coefficient. The actual loss of the filter cartridge is calculated using the formula "actual loss of filter cartridge = total cumulative water volume filtered × filter cartridge loss coefficient". Finally, the actual loss is compared with the rated total loss of the filter cartridge (determined by the filter cartridge factory standard and entered into the system by the administrator in the background) to obtain the loss ratio of the filter cartridge, thus completing the detection of the actual use loss of the filter cartridge. The detection results are fed back to the filter cartridge life assessment module 4 in real time, providing accurate data support for the calculation of the remaining service life of the filter cartridge. At the same time, it can be pushed to the administrator through LoRa networking, so that the administrator can keep track of the filter cartridge loss status.

[0055] In another embodiment, please refer to Figure 4 The filter life assessment module 4 includes: a water consumption prediction unit, used to determine the future average daily water consumption, which is determined by taking the average value after removing abnormal days, based on the recent actual daily water consumption of each water purification device.

[0056] The lifespan calculation unit combines future water quality forecast data, actual filter cartridge wear status, and rated filter cartridge capacity with the average daily water consumption to calculate the actual remaining lifespan of the filter cartridge. The calculation results are accurate to the day and can be dynamically adjusted according to future water quality trends.

[0057] In this embodiment, the filter cartridge life assessment module 4 is internally divided into a water consumption prediction unit and a life calculation unit. The two units work together to calculate the actual remaining lifespan of the filter cartridge. Its data input comes from the water quality status prediction module 2 and the filter cartridge wear analysis module 3 (both transmit relevant data via LoRa networking). The specific implementation details are as follows:

[0058] 1. Water Consumption Prediction Unit: The core function of this unit is to determine the average daily water consumption in the future, providing a load basis for calculating the remaining lifespan of the filter cartridge and ensuring that the calculation results closely match the actual usage scenario. The specific implementation process is as follows: First, the recent daily actual water consumption data of each water purifier is obtained through LoRa networking—the flow sensor of each water purifier collects the daily water consumption in real time, which is transmitted to this unit via the LoRa communication module and LoRa gateway; Second, the collected recent daily actual water consumption data is processed, and abnormal days are removed (such as abnormally low or high water consumption due to equipment failure and maintenance, and zero water consumption due to water outage); Finally, the average of the daily water consumption data for the remaining valid days is taken to determine the future average daily water consumption. This average value can accurately reflect the actual usage load of each water purifier, providing reliable support for subsequent lifespan calculations.

[0059] 2. Lifespan Calculation Unit: The core function of this unit is to combine water quality forecast data for future periods, the actual wear and tear of the filter cartridge, and the rated filtration capacity of the filter cartridge, and take into account the average daily water consumption in the future to calculate the actual remaining lifespan of the filter cartridge. The lifespan data can also be dynamically corrected according to future water quality change trends. The specific implementation process is as follows: First, the water quality trend forecast module 2 receives the future water quality change trend (data processed by feature extraction unit 21 and prediction output unit 22) to determine the filter cartridge wear coefficient change trend corresponding to the future water quality. Second, the filter cartridge wear analysis module 3 receives the actual wear status of the filter cartridge (actual wear amount and wear ratio) and obtains the rated filtration capacity of the filter cartridge (the standard value of the filter cartridge entered by the administrator through the background). Then, the average daily water consumption in the future determined by the water consumption forecast unit is incorporated, and the remaining service life of the filter cartridge is calculated through a preset algorithm—first, the remaining filterable capacity of the filter cartridge is calculated (remaining filterable capacity = rated filtration capacity - actual wear amount), then, combined with the future daily water consumption and the wear coefficient corresponding to the future water quality, the expected daily wear amount of the filter cartridge is calculated. Finally, the remaining service life of the filter cartridge is obtained by "remaining service life = remaining filterable capacity ÷ average expected daily wear amount in the future", and the calculation result is accurate to the day. Finally, if the water quality changes in the future (such as predicted water quality deterioration and increased wear coefficient), the daily expected wear amount is dynamically adjusted, and the remaining service life calculation result is corrected synchronously to ensure the accuracy and adaptability of the service life prediction.

[0060] In another embodiment, please refer to Figure 5 It also includes a water quality anomaly handling module 5, which is used to generate corresponding water quality anomaly early warning information and handling methods based on the water quality change trend and anomaly risk level output by the water quality situation prediction module 2. The handling methods include pushing early warning prompts to the administrator terminal, sending shutdown instructions or automatic flushing instructions to the corresponding water purification equipment.

[0061] In this embodiment, the water quality anomaly handling module 5 is an important extension module of this system. Its core function is to generate corresponding water quality anomaly early warning information and handling methods based on the water quality change trend and anomaly risk level output by the water quality situation prediction module 2, so as to realize rapid response and handling of water quality anomalies and ensure the safety of residents' drinking water. Its implementation process relies entirely on the data support of the modules mentioned above and the transmission function of LoRa networking, as detailed below:

[0062] First, this module establishes a real-time data connection with the water quality situation prediction module 2, and receives the future water quality change trend and abnormal risk level (general risk, emergency risk) output by the prediction output unit 22. At the same time, it obtains the real-time operating status data of each water purification device through LoRa networking to ensure that the specific situation of water quality abnormality (such as abnormal device location, abnormal parameters, abnormal time period) can be accurately grasped.

[0063] Secondly, based on the level of abnormal risk, corresponding water quality anomaly warning information and handling methods are generated. Specific handling methods include pushing warning prompts to the administrator, sending shutdown commands or automatic flushing commands to the corresponding water purification equipment, and different handling logics corresponding to different risk levels:

[0064] 1. General Risk: When the water quality situation prediction module 2 outputs a general risk (slight fluctuations in water quality that will not affect drinking safety in the short term), the module only generates an early warning message, which is pushed to the administrator (including mobile app and web backend) via LoRa networking. The early warning message includes detailed information such as the abnormal prediction time, abnormal parameters, corresponding device location, and abnormal risk level, reminding the administrator to pay attention to the water purification equipment in the corresponding area and prepare for inspection. There is no need to shut down or flush the water purification equipment, ensuring the normal operation of the equipment.

[0065] 2. Emergency Risk: When the water quality situation prediction module 2 outputs an emergency risk (the water quality will seriously exceed the standard, which may endanger the drinking safety of residents), the module will perform two actions simultaneously: First, it will push an emergency warning to the administrator, reminding the administrator to immediately arrange maintenance personnel to go to the corresponding area for handling. The warning information will highlight the emergency risk level and the handling priority. Second, it will send a shutdown command to the water purification equipment in the corresponding area through LoRa networking, control the equipment to stop water production, and prevent the outflow of substandard water. After the maintenance personnel have handled the situation and the water quality has returned to normal, the administrator will send a power-on command through the background, and the equipment will resume operation.

[0066] 3. Recoverable Anomalies: When the water quality situation prediction module 2 predicts that the water quality anomaly is due to gradual deterioration caused by slight clogging of the filter cartridge (which falls under the category of general risk but requires timely handling), the module sends an automatic flushing command to the corresponding water purification equipment through LoRa networking. This controls the equipment to automatically flush the filter cartridge, alleviate filter cartridge wear, temporarily restore water quality, and buy time for maintenance personnel to conduct subsequent inspections and filter cartridge replacements. After flushing is completed, the equipment automatically resumes normal operation, and the module simultaneously pushes a flushing completion notification to the administrator, facilitating the administrator's subsequent tracking of the filter cartridge status.

[0067] In another embodiment, please refer to Figure 6 It also includes a regional anomaly collaborative judgment module 6, which is used to summarize the water quality prediction data of all water purification equipment and identify equipment clusters in the same area based on the deployment location information of the water purification equipment; when multiple water purification equipment in the same area are predicted to have water quality anomalies, it is determined to be a regional water supply problem, and regional location, number of abnormal equipment and summary information of abnormal parameters are added to the early warning prompt to support the overall scheduling of operation and maintenance resources.

[0068] In this embodiment, the regional anomaly collaborative judgment module 6 is an extension module of the water quality anomaly handling module 5. Its core function is to aggregate the water quality prediction data of all water purification equipment, identify equipment clusters in the same area, and determine whether the water quality anomaly is a regional water supply problem or a single equipment failure. This provides support for the overall scheduling of operation and maintenance resources and improves the efficiency of anomaly handling. Its implementation process relies entirely on the transmission function of LoRa networking and the data support of the modules mentioned above, as detailed below:

[0069] 1. Data Aggregation and Regional Classification: This module receives water quality prediction data from all water purification devices via LoRa networking (output by prediction output unit 22 of water quality situation prediction module 2, aggregated and transmitted after LoRa gateway). At the same time, it retrieves the deployment location information of each water purification device (such as building number and park area division) pre-stored in the system, performs regional classification on all water quality prediction data, divides water purification devices in the same building or the same park area into a regional device cluster, establishes the association relationship of "region-device-water quality prediction data", and ensures that all water purification devices in the same area can be quickly identified.

[0070] 2. Regional Anomaly Identification and Judgment: This module summarizes and analyzes the water quality prediction data of each regional equipment cluster in real time, focusing on detecting water quality anomaly predictions of water purification equipment within the region. The "preset number of multiple water purification devices" can be flexibly set according to the actual deployment scale of the community (e.g., if more than 3 out of 8 water purification devices in the same building are predicted to have water quality anomalies, a regional anomaly judgment is triggered). This preset number can be adjusted by the administrator through the cloud backend to adapt to the equipment density requirements of different communities. When multiple water purification devices in the same region that reach the preset number are all predicted to have water quality anomalies, this module combines historical data from the water quality data acquisition module 1 with equipment operation data to rule out the possibility of single-device malfunctions (such as sensor failures or filter blockages causing anomaly predictions). It automatically determines that the water quality anomaly is a regional water supply problem (such as aging of the municipal water supply network, abnormalities in regional secondary water supply facilities, or overall water quality fluctuations caused by municipal water supply improvements), rather than a single device malfunction.

[0071] 3. Early Warning Optimization and Operation and Maintenance Support: When this module determines that there is a regional water supply problem, it immediately sends a regional anomaly signal to the Water Quality Anomaly Handling Module 5. When generating early warning information, the Water Quality Anomaly Handling Module 5 adds an additional regional location identifier (such as "Building 1" or "East Zone of the Park"), the specific number of abnormal devices in the area, the location of all abnormal devices, and a summary of abnormal parameters (such as "The TDS value of all 3 devices in Building 1 is predicted to exceed the standard, and the current predicted maximum value is 160ppm"). At the same time, the optimized early warning information is pushed to the administrator and operation and maintenance personnel through LoRa networking, enabling the administrator to quickly grasp the overall situation of the regional anomaly and coordinate the operation and maintenance resources in the area (such as arranging operation and maintenance personnel to go to the area first to conduct a comprehensive inspection of all abnormal devices and surrounding water supply facilities). This avoids the dispersion of operation and maintenance resources due to misjudging a single device failure, greatly improves the efficiency of handling regional water quality anomalies, and reduces the risk of community residents drinking substandard water.

[0072] In addition, if only a single device in a certain area is predicted to have water quality anomalies, the module will determine it as a single device malfunction and only trigger the water quality anomaly handling module 5 to push a warning prompt for that single device, reminding the administrator to conduct a separate inspection of the device (such as checking whether the sensor is faulty or the filter is clogged), avoiding false triggering of area anomaly judgments and further improving the system's intelligence level and operation and maintenance accuracy.

[0073] 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, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A community water purification equipment cluster management system based on LoRa networking, characterized in that, include: The water quality data acquisition module is used to acquire historical and real-time water quality data of water purification equipment within the target area; The water quality trend prediction module is used to predict the water quality change trend in the future based on the historical water quality data and real-time water quality data. The filter cartridge wear analysis module is used to compare real-time water quality data, historical water quality data and preset grading standards to determine the filter cartridge wear coefficient corresponding to the current water quality. It also calculates the actual wear of the filter cartridge by combining the total amount of water filtered since the filter cartridge was installed with the total rated wear of the filter cartridge, and compares it with the rated wear of the filter cartridge to complete the detection of the actual wear of the filter cartridge. The filter cartridge life assessment module is used to calculate and determine the actual remaining service life of the filter cartridge based on the predicted future water quality change trend and the actual wear and tear of the filter cartridge.

2. The community water purification equipment cluster management system based on LoRa networking according to claim 1, characterized in that; The water quality trend prediction module includes: a feature extraction unit, used to extract time-series features from historical and real-time water quality data using a long short-term memory network model, capturing the gradual change patterns and abrupt changes in water quality parameters; and a prediction output unit, used to analyze the extracted features in conjunction with preset water quality warning thresholds, output water quality prediction data and abnormal risk levels for future periods, and complete the prediction of water quality change trends for future periods.

3. The community water purification equipment cluster management system based on LoRa networking according to claim 2, characterized in that; The parameters of the Long Short-Term Memory (LSTM) network model are set as follows: the input layer dimension is set to 4, corresponding to the four core water quality parameters of TDS, turbidity, pH, and residual chlorine content; the hidden layer is set to 2 layers, with 64 neurons in each layer; the forget gate threshold is set to 0.7, the input gate threshold is set to 0.3, and the output gate threshold is set to 0.5; the time step is set to 24 to adapt to the low-frequency data transmission characteristics of LoRa networking; the training batch size is 32, and the number of iterations is 100 rounds; the Adam optimizer is used, the learning rate is set to 0.001, and the loss function is the mean squared error.

4. The community water purification equipment cluster management system based on LoRa networking according to claim 2, characterized in that; The feature extraction process of the feature extraction unit is as follows: first, the real-time data and historical data are standardized and preprocessed to automatically remove abnormal data; then, the core features of water quality parameters are extracted, and key features are screened in combination with related data to provide support for the prediction output unit.

5. The community water purification equipment cluster management system based on LoRa networking according to claim 1, characterized in that; The filter cartridge wear analysis module includes: a wear coefficient determination unit, used to compare real-time water quality data, historical water quality data and preset grading standards to determine the filter cartridge wear coefficient corresponding to the current water quality. The preset grading standards include water quality impurity content grading standards and water quality fluctuation amplitude grading standards. The wear coefficient is positively correlated with impurity content and fluctuation amplitude; and a wear amount detection unit, used to collect the cumulative total water volume filtered by the filter cartridge since installation, calculate the actual wear amount of the filter cartridge based on the determined filter cartridge wear coefficient, and compare it with the rated total wear amount of the filter cartridge to complete the detection and feedback of the actual wear of the filter cartridge.

6. The community water purification equipment cluster management system based on LoRa networking according to claim 1, characterized in that, The filter life assessment module includes: a water consumption prediction unit, used to determine the future average daily water consumption, which is determined by taking the average value after removing abnormal days, based on the recent actual daily water consumption of each water purification device. The lifespan calculation unit combines future water quality forecast data, actual filter cartridge wear status, and rated filter cartridge capacity with the average daily water consumption to calculate the actual remaining lifespan of the filter cartridge. The calculation results are accurate to the day and can be dynamically adjusted according to future water quality trends.

7. The community water purification equipment cluster management system based on LoRa networking according to any one of claims 1-6, characterized in that, It also includes a water quality anomaly handling module, which is used to generate corresponding water quality anomaly early warning information and handling methods based on the water quality change trend and anomaly risk level output by the water quality situation prediction module. The handling methods include pushing early warning prompts to the administrator terminal, sending shutdown commands or automatic flushing commands to the corresponding water purification equipment.

8. The community water purification equipment cluster management system based on LoRa networking according to claim 7, characterized in that, It also includes a regional anomaly collaborative judgment module, which is used to summarize the water quality prediction data of all water purification equipment and identify equipment clusters in the same area based on the deployment location information of the water purification equipment; when multiple water purification equipment in the same area are predicted to have water quality anomalies, it is determined to be a regional water supply problem, and regional location, number of abnormal equipment and summary information of abnormal parameters are added to the early warning prompt to support the overall scheduling of operation and maintenance resources.

Images