Station water data intelligent collection terminal and system

Abstract

The present disclosure relates to a station water affairs data intelligent collection terminal and system, including a data collection unit, a network communication unit and a core control unit, wherein the core control unit is configured with an offline DQN module and a spare storage module, and when working, the offline DQN module obtains the collection parameter matrix of the data collection unit and the signal parameter matrix of the network communication unit in each working cycle, and obtains the energy consumption state of the data collection unit and the network communication unit, and constructs an energy consumption state matrix; the data is trained for a predetermined round with a preconfigured action space and a reward function; the action space includes: signal transmission offset time, signal collection offset time and average occupancy ratio of the reserve storage module; after a predetermined period of data training, the offline DQN sends the output data matrix to the core control unit, and the core control unit manages and controls the work of the collection terminal based on the output data matrix. The present disclosure greatly improves the communication efficiency and the quality of collected data.

Description

Technical Field

[0001] The present invention belongs to the field of environmental protection, and in particular to an intelligent water affairs data collection terminal. Background Art

[0002] With the continuous development of the environmental protection water industry, more and more information technology, artificial intelligence technology, cloud computing and other technologies are being applied in the environmental protection field, promoting the refined development of environmental protection water emission monitoring, pre-warning, process control and post-verification.

[0003] The station water data acquisition terminal in existing environmental water monitoring technology usually includes a power supply module, a control module, an acquisition module and a communication module. It is dispatched by the control module, the acquisition module realizes on-site monitoring data acquisition, and the communication module realizes remote uploading. Most of them can only realize data acquisition and processing through specified interfaces such as serial ports or analog quantities. The functions are relatively simple and the power consumption is relatively high.

[0004] Due to the wide variety of production companies in the environmental protection and water sector and the variety of on-site monitoring equipment, data collection methods and efficiency are greatly limited, making it difficult to meet the growing demand for real-time and comprehensive monitoring in the environmental protection industry. At the same time, in relatively remote areas, batteries need to be replaced frequently, and the maintenance workload is also relatively large. Summary of the Invention

[0005] Purpose of the invention: To provide an intelligent water data collection terminal for a station building to solve the above-mentioned problems existing in the prior art.

[0006] Technical solution: Station water data intelligent collection terminal, including:

[0007] The data acquisition unit is connected to at least two data acquisition devices via a data interface.

[0008] The network communication unit includes at least two types of communication modules.

[0009] The core control unit is connected to the data acquisition unit and the network communication unit, and is used to receive and process the collected data, and communicate with the host computer through the network communication unit;

[0010] The core control unit is equipped with an offline DQN module and a backup storage module.

[0011] During operation, the offline DQN module obtains the acquisition parameter matrix of the data acquisition unit and the signal parameter matrix of the network communication unit in each working cycle, obtains the energy consumption status of the data acquisition unit and the network communication unit, and constructs the energy consumption status matrix; it trains the data for a predetermined number of rounds using the preconfigured action space and reward function;

[0012] The action space includes: signal transmission offset time, signal collection offset time and average occupancy ratio of the standby storage module;

[0013] The reward function R is:

[0014] ;

[0015] Where, w i+1 、 w i are the energy consumption of the network communication unit in the i+1th and ith cycles respectively; w 0 is the minimum energy consumption of the network communication unit under theoretical conditions;

[0016] v i+1 、 v i are the energy consumption of the data acquisition unit in the i+1th and ith cycles respectively; v 0 is the minimum energy consumption of the data acquisition unit under theoretical conditions; a and b are set constants;

[0017] After a predetermined period of data training, the offline DQN sends the output data matrix to the core control unit, which controls the work of the acquisition terminal based on the output data matrix.

[0018] According to one aspect of the present application, the backup storage module is divided into a number of storage areas of predetermined capacity, and the number of storage areas is not less than the number of data acquisition units;

[0019] The core control unit sequentially collects the working data streams of each data acquisition unit in the first working cycle and establishes an acquisition time sequence set of the data acquisition unit;

[0020] Sequentially collect the signal transmission data streams of each network communication unit and establish a transmission timing set of the data collection unit;

[0021] Based on the collection timing set and the sending timing set, the working data flow of each data collection unit is constructed to the sending data flow of each network communication unit, and a data flow OD matrix is ​​established;

[0022] Calculate the proportion of data flow directions in the data flow OD matrix within each working cycle and arrange them in descending order;

[0023] Calculate the data transmission time ratio in the data flow OD matrix within each working cycle and arrange them in descending order;

[0024] Adjust the capacity ratio of each storage area in the backup storage module based on the data flow ratio and data transmission time ratio.

[0025] According to one aspect of the present application, in the next working cycle, an OSD matrix is ​​constructed for each port of the data acquisition unit, each storage area of ​​the spare storage module, and each communication module of the network communication unit, and data streams with a data flow ratio higher than a preset value and a data transmission time ratio higher than a threshold are stored in the spare storage module in proportion, and the time series of the data stream is adjusted. After the time series is adjusted, the power consumption reduction ratio of the data acquisition unit and the network communication unit is calculated; the above process of adjusting the data stream time series is repeated until the power consumption reduction ratio reaches the preset value.

[0026] According to one aspect of the present application, based on the data flow ratio and the data transmission time ratio, each data acquisition device and communication module is graded, and a weight matrix of the data acquisition device and the communication module is constructed respectively; based on the weight matrix, the data flow time series in the subsequent working cycle is adjusted.

[0027] According to one aspect of the present application, when the power consumption reduction ratio does not reach a preset value during a predetermined training period, the acquisition timing set and the transmission timing set are compressed, and the similarity between the compressed local data stream and the uncompressed complete data stream is calculated. Based on the similarity between the local data stream and the complete data stream within the predetermined period and the data evaluation effect of the local data stream, a similarity threshold is set;

[0028] During use, if the power consumption reduction ratio does not reach the preset value, the data resolution of the corresponding data acquisition device is adjusted according to the data flow ratio, so that the power consumption reaches the lowest value within the training cycle while being higher than the similarity threshold.

[0029] According to one aspect of the present application, the data acquisition device includes a temperature and humidity sensor, a flue gas monitor, a COD monitor, and a flow meter. The communication module includes a LoRa module, a 4G module, and an Ethernet module. The data interface includes RS232, RS485, ADC, and a switch.

[0030] Further provided is a station water data intelligent collection system, comprising the above-mentioned intelligent collection terminal, and a host computer connected to the intelligent collection terminal via a communication network;

[0031] The host computer includes an intelligent acquisition terminal management module, which receives data from each intelligent acquisition terminal in each working cycle; the intelligent acquisition terminal management module is used to:

[0032] Collect the location information of each intelligent collection terminal and mark them one by one in the GIS system of the predetermined area; build the geographic spatial topology map of the water network in the area where the intelligent collection terminal is located;

[0033] Based on the flow direction of the water network, the distance between adjacent smart terminals and the regional flow rate, and based on the position of the smart collection terminal on the water network, a terminal directed graph is constructed; based on the distance between adjacent smart terminals and the regional flow rate, a time offset matrix of the smart collection terminal is constructed;

[0034] Based on the terminal directed graph, the coupling relationship of the intelligent collection terminal is mined, and the difference between the theoretical calculation value and the actual measurement value of the downstream intelligent collection terminal is compared; if the difference is within a predetermined range, the intelligent collection terminal is marked as a non-independent terminal.

[0035] Beneficial effects: Greatly improved communication efficiency and collected data quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 FIG. 4 is a schematic diagram of the topological structure of this embodiment.

[0037] Figure 2 Schematic diagram of the topological structure of the acquisition terminal in this embodiment.

[0038] Figure 3 This is a simplified diagram of the data flow of this embodiment.

[0039] Figure 4 This is a data flow diagram of this embodiment. DETAILED DESCRIPTION

[0040] The technical background and technical overview of this application are described as follows: First, the hardware structure of this application is described.

[0041] like Figure 1 and Figure 2 As shown in the figure, the hardware backup of this terminal is divided into components such as power supply unit, core control unit, data acquisition unit, LoRa networking unit and public network communication unit.

[0042] The power supply unit, which includes components such as a switching power supply and a backup battery, manages terminal power usage. Its power output is connected to the power inputs of the core control unit, data acquisition unit, LoRa networking unit, and public network communication unit. If the switching power supply fails, the backup battery can provide continuous operation for over six hours. This can be achieved using a UPS, storage battery, or related products.

[0043] The hardware part of the core control unit includes a high-performance ARM9 embedded real-time operating system and auxiliary circuits, and has functions such as data acquisition control, data call control, data storage and reporting control, event alarm control, LoRa networking control and wireless public network networking control.

[0044] In a certain engineering case, the data acquisition unit includes 8-channel AD sampling, 1-channel Bluetooth, 2-channel RS485, 3-channel RS232, and 4-channel switch quantity modules. It can collect on-site monitoring data in a variety of ways and support reading various water online automatic monitoring instruments, such as flow meters, COD monitors, flue gas monitors, temperature and humidity sensors, etc.

[0045] The network communication unit includes modules such as LoRa and Ethernet. The LoRa networking unit contains two LoRa modules for on-site local area networking, collecting monitoring data from various cloud modules and supporting the reading of water and electricity meters using standard national grid protocols. The public network communication unit includes two 4G and one 100M Ethernet modules, enabling the establishment of two primary and backup wireless public network communication links and one 100M Ethernet link. This allows interaction with multiple host computer master stations, enabling on-demand, real-time reporting and scientific management of environmental monitoring data. In other embodiments, 2G, 3G, 5G, and 6G modules can also be used. Zigbee and other modules can also be used.

[0046] For example, in one embodiment, a water utility site has a temperature acquisition device with an RS485 communication interface, a pressure acquisition device with an RS232 interface, and a flow meter with a 4-20mA analog interface. These interfaces can be connected to on-site monitoring instruments via the reserved RS485, RS232, and analog interfaces. Temperature, pressure, and flow data can be queried using the communication protocols of each of these three devices. The responses from the on-site devices are then repackaged and uploaded to a host computer (either the Environmental Protection Bureau's water monitoring platform or the enterprise's self-testing platform) via a public network communication unit.

[0047] For example, if there are multiple State Grid smart meters on site and real-time power consumption needs to be collected to monitor their operating status, this terminal can connect to each State Grid smart meter through the reserved LoRa signal. This terminal can read the data from each smart meter wirelessly via LoRa, repackage it, and upload it to the host computer (environmental protection bureau, enterprise) through the public network interaction unit.

[0048] The core control unit interacts with the data acquisition unit and LoRa networking unit in the downstream direction, and communicates with the public network communication unit in the upstream direction, implementing functions such as data acquisition control, data call control, data storage and reporting control, event alarm control, LoRa networking control, and public network control. For example, when the host computer master station calls for data, the core control unit receives the call command after the public network is established, begins downstream interaction with the data acquisition unit, connects to the specified on-site monitoring instrument through the designated acquisition interface, obtains data, and performs data storage and data reporting based on the current data reporting status.

[0049] Next, we will mainly describe the software part of the intelligent data collection terminal. Since the operation and implementation of the software requires hardware support, we will briefly describe the topology of the hardware part again.

[0050] In this embodiment, the hardware topology is simplified and summarized.

[0051] In this embodiment, the station water data intelligent collection terminal mainly includes modules such as a data collection unit, a network communication unit, and a core control unit.

[0052] The data acquisition unit connects to at least two data acquisition devices via a data interface. These devices can be any of the devices described above, or devices with similar functionality. Devices that achieve the same or similar results and substantially the same or similar functions, including temperature and humidity sensors, flue gas monitors, COD monitors, and flow meters, are all acceptable. Due to space limitations, a comprehensive list is not provided. Data interfaces include RS232, RS485, ADC, and switching.

[0053] The network communication unit includes at least two types of communication modules. The communication modules can adopt the various modules mentioned above, including LoRa modules, 4G modules and Ethernet modules. They can also adopt other industrial or civilian modules with similar functions, which are not described one by one here.

[0054] The core control unit is connected to the data acquisition unit and the network communication unit, and is used for receiving and processing the collected data, and communicating with the host computer through the network communication unit.

[0055] In other words, the hardware mainly consists of a data acquisition unit, a network communication unit and a core control unit, which can complete the entire process of data acquisition and data transmission.

[0056] The software module mainly includes the intelligent management and control unit, or the neural network processing unit. The following describes the relevant functions and implementation details of the software module.

[0057] The core control unit is equipped with an offline DQN module and a backup storage module.

[0058] During operation, the offline DQN module obtains the acquisition parameter matrix of the data acquisition unit and the signal parameter matrix of the network communication unit in each working cycle, obtains the energy consumption status of the data acquisition unit and the network communication unit, and constructs the energy consumption status matrix; it trains the data for a predetermined number of rounds using the preconfigured action space and reward function;

[0059] The action space includes: signal sending offset time, signal collection offset time and average occupancy ratio of the preparation storage module;

[0060] The reward function R is:

[0061] ;

[0062] Where, w i+1 、 w i are the energy consumption of the network communication unit in the i+1th and ith cycles respectively; w 0 is the minimum energy consumption of the network communication unit under theoretical conditions;

[0063] v i+1 、 v i are the energy consumption of the data acquisition unit in the i+1th and ith cycles respectively; v 0 is the minimum energy consumption of the data acquisition unit under theoretical conditions; a and b are set constants, which are positive numbers;

[0064] After a predetermined period of data training, the offline DQN sends the output data matrix to the core control unit, which controls the work of the acquisition terminal based on the output data matrix.

[0065] In this embodiment, during development, a neural network module, namely a DQN module, is first constructed on a background host computer. The module is then embedded in each intelligent data acquisition terminal. During operation, data from the intelligent terminal is collected and trained at predetermined intervals. In other words, at the beginning, the intelligent data acquisition terminal performs optimization training while operating, achieving continuous optimization and iteration, thereby achieving better and more power-saving results. Alternatively, historical data can be used as a training set before shipment, and then the intelligent data acquisition terminal is trained. Then, when in use, the trained model can be directly used for processing, improving the efficiency of data collection and transmission and reducing the terminal's power consumption.

[0066] In this embodiment, the model is trained by setting the action space, state space, and reward function required for DQN training, mainly to improve the timeliness of signal transmission and reduce overall power consumption.

[0067] The reward function is explained as follows, for example, w 0The minimum power of the communication unit as a whole, for example, the silent power, is 1.0. In a certain operating cycle, the current power consumption is 1.5. After adjusting the signal transmission offset, signal transmission becomes more compact, the overall operating time is reduced, and thus the energy consumption is reduced. The energy consumption reaches 1.2, and the reward is -a(-0.3 / 0.2) = a*1.5, indicating that the direction is correct. If the energy consumption reaches 1.1, then -a(-0.4 / 0.1) = a*4, indicating a better direction and a higher reward. However, due to adjusting the signal transmission offset, the overall power consumption increases to 1.8. In this case, the reward of the communication unit is -a(0.3 / 0.8) = -a*0.375. The reward is negative, indicating that the direction is reversed. For the data collection unit, the reward function principle is similar and will not be detailed here.

[0068] a and b are weight coefficients. By adjusting the data of a and b, the proportion of information collection and communication modules can be allocated to achieve overall optimization. By training multiple rounds, for example, 20,000 rounds, energy consumption can be continuously optimized and reduced, which also shows that the signal processing method is more reasonable.

[0069] In order to adjust the timing of collection and transmission, cache queuing is required under certain working conditions. The existing core control unit is generally implemented using an integrated MCU, and the storage module is small. It is necessary to expand or add additional storage modules to maximize efficiency. In other words, the preparatory storage module can be part of the MCU or an independent new storage module. At the same time, if the process of information collection to transmission is analyzed in terms of timing, it can be seen that there is a certain overlap in the time of collection, storage, and transmission. Therefore, it is necessary to analyze the timing of the data. To this end, the following solution is provided:

[0070] According to one aspect of the present application, the backup storage module is divided into a number of storage areas of predetermined capacity, and the number of storage areas is not less than the number of data acquisition units;

[0071] The core control unit sequentially collects the working data streams of each data acquisition unit in the first working cycle and establishes an acquisition time sequence set of the data acquisition unit;

[0072] Sequentially collect the signal transmission data streams of each network communication unit and establish a transmission timing set of the data collection unit;

[0073] Based on the collection timing set and the sending timing set, the working data flow of each data collection unit is constructed to the sending data flow of each network communication unit, and a data flow OD matrix is ​​established;

[0074] Calculate the proportion of data flow directions in the data flow OD matrix within each working cycle and arrange them in descending order;

[0075] Calculate the data transmission time ratio in the data flow OD matrix within each working cycle and arrange them in descending order;

[0076] Adjust the capacity ratio of each storage area in the backup storage module based on the data flow ratio and data transmission time ratio.

[0077] like Figure 3 and Figure 4 As shown in Figure 2, the general process of data collection and transmission is given.

[0078] In a preferred embodiment, the number of storage areas is generally set to be greater than the number of acquisition units; that is, each acquisition unit is assigned at least one storage area. In other embodiments, acquisition units with smaller data volumes may not be assigned a storage area, or may share the storage areas of other acquisition units. Within a cycle, at times t0, t1, t2, ..., tn, a certain acquisition port a may send information n times, with k times to communication port a, m times to communication port b, and nmk times to communication port c. Similarly, acquisition ports b, c, and d may also experience similar situations.

[0079] Therefore, when acquisition port a sends information to communication port b, the information sent by acquisition port c to communication port b is occupying communication port b and requires time ∆t, so it needs to wait. At this time, acquisition port a needs to be in working state and cannot be in standby state to save energy.

[0080] Some information may be sent very frequently but with relatively small amounts of data, resulting in a shorter transmission time. Other information may be sent less frequently but with relatively large amounts of data. To improve system efficiency and reduce ineffective processing time, we can analyze the number of transmissions (frequency) and total transmission time for each acquisition port. We can then optimize high-frequency and time-consuming information transmission paths, improving the overall efficiency of acquisition and communication ports and reducing energy consumption.

[0081] For example, within a certain working cycle, the data flow ratio of a certain direction in the statistical data flow OD matrix is ​​counted, and the data flow ratios are sorted to filter out several groups with high data flow frequencies. At the same time, according to a similar method, the data transmission time ratio is counted and several groups are filtered out.

[0082] Given weight coefficients for the data flow ratio and data transmission time ratio, a total statistical analysis is performed to obtain an overall ranking. The data flow OD ranking is then given, and based on this ranking, a correction coefficient is used to determine the storage area capacity ratio. Storage capacity is then allocated based on this capacity ratio.

[0083] In other words, some data streams may appear relatively high in frequency statistics, but relatively low in time. Therefore, a comprehensive approach is taken to assign a relatively balanced ranking, similar to a weighted score. Storage capacity is then allocated based on the normalized weighted score. This approach balances the operating time of each data stream, improving the system's collection and distribution efficiency. Ideally, after data streams are collected, they can be sent out promptly within a predetermined timeframe, reducing queuing and communication congestion. After sending data, the collection unit goes into hibernation, saving energy. The communication unit also goes into hibernation after sending data, reducing energy consumption. This avoids the existing scenario where all collection and communication units are on standby, spending the majority of their time inactive. This new solution allows for the intensification of the information collection and transmission process. Information is collected and transmitted during a specific timeframe, while all or some units go into hibernation during other timeframes. This allows the collection and transmission process to be compressed to a specific timeframe, minimizing energy waste.

[0084] After pre-queuing using the spare storage module, compared with the existing method, the density of signal collection and transmission per unit time is greatly improved, and the sleep time of each unit is extended.

[0085] However, subsequent optimization efforts revealed that the above embodiment only considered the signal path from the acquisition port to the transmission port, but did not consider the occupancy time and occupancy rate of the backup storage module. Analysis and calculation revealed that incorporating this into the evaluation criteria could further optimize overall resource utilization, extend the sleep time of the core processing unit, and thus further improve system efficiency. Therefore, further optimized embodiments are provided below:

[0086] According to one aspect of the present application, in the next working cycle, an OSD matrix is ​​constructed for each port of the data acquisition unit, each storage area of ​​the spare storage module, and each communication module of the network communication unit, and data streams with a data flow ratio higher than a preset value and a data transmission time ratio higher than a threshold are stored in the spare storage module in proportion, and the time series of the data stream is adjusted. After the time series is adjusted, the power consumption reduction ratio of the data acquisition unit and the network communication unit is calculated; the above process of adjusting the data stream time series is repeated until the power consumption reduction ratio reaches the preset value.

[0087] By constructing an OSD matrix, the working status of the storage area is included in the optimization parameters. After placing the data in the preliminary storage area, the time sequence of the signal from acquisition to transmission is optimized and adjusted. The overall time and power consumption are optimized in accordance with the above embodiment, and the energy consumption reduction ratio is adjusted to the expected target.

[0088] For example, during the signal transmission process, if only the timing of signal acquisition and transmission is adjusted, it is possible that only some of the signal acquisition units and signal transmission units are working. However, by storing part of the data in the reserve storage area, more signal acquisition and transmission units can work in a shorter period of time, increasing the proportion of working time. Then, the port that has collected data can put the data into the reserve storage area and then go to sleep. After the sending port is queued, the data can be directly retrieved from the reserve storage area and sent. Therefore, the collaborative efficiency of each port is improved, the problem of misplaced work and low work efficiency is greatly reduced, which further increases energy loss and extends the service life of the intelligent acquisition terminal.

[0089] Because some data are relatively important, the immediacy requirement is higher, while the immediacy requirement of some data is not that high. According to the above embodiment, simply adjusting the working sequence according to the data flow frequency and time proportion may cause some data to be placed at the back. On the basis of the above embodiment, in order to further improve the efficiency of data acquisition and use, according to one aspect of the present application, based on the data flow proportion and data transmission time ratio, each data acquisition device and communication module are graded, and the weight matrix of the data acquisition device and communication module is constructed respectively; based on the weight matrix, the data flow time series in the subsequent working cycle is adjusted.

[0090] By hierarchically processing the collection devices and communication modules, the issues of information priority and immediacy can be resolved.

[0091] For example, if a data flow has a high frequency and takes a long time to transmit, but is of low importance, then while the above optimization scheme can improve overall efficiency and reduce energy consumption, it may cause important information to be ranked lower. To address this issue, ports are weighted and their sequence numbers adjusted.

[0092] In other embodiments, the weights can be combined into a weight matrix, for example, by matrix multiplication to integrate the individual weight matrices into a single weight matrix, thereby accelerating the training speed and improving the convergence speed of the neural network.

[0093] In practice, it has been found that due to the large number of devices or ports on some information collection terminals, the workload of data collection and transmission is heavy, making it difficult to reduce energy consumption to an ideal level. In remote areas or certain specific situations, electricity is inconvenient, requiring batteries that can last for a long time. Therefore, a preferred technical solution is provided, as detailed below.

[0094] According to one aspect of the present application, when the power consumption reduction ratio does not reach a preset value during a predetermined training period, the acquisition timing set and the transmission timing set are compressed, and the similarity between the compressed local data stream and the uncompressed complete data stream is calculated. Based on the similarity between the local data stream and the complete data stream within the predetermined period and the data evaluation effect of the local data stream, a similarity threshold is set;

[0095] During use, if the power consumption reduction ratio does not reach the preset value, the data resolution of the corresponding data acquisition device is adjusted according to the data flow ratio, so that the power consumption reaches the lowest value within the training cycle while being higher than the similarity threshold.

[0096] In other words, by optimizing the resolution (volume or frequency) of collected data, a reasonable amount of data is collected to reflect the actual situation. Some data can be discarded or the resolution reduced, allowing the partial data to reflect the overall situation. In other words, data sparsification reduces the proportion of non-critical data and increases the effectiveness of data, allowing less data to reflect the overall situation. This, in turn, reduces the workload of intelligent data collection terminals and extends their service life.

[0097] During training, historical data can be divided into training sets and test sets. The ratio of reducing data resolution can be evaluated through partial data, and then a similarity threshold is obtained. This similarity threshold is then used as a constraint condition for training.

[0098] It should be noted that the similarity threshold can be set to a range or multiple point values ​​and adjusted according to the power consumption of the intelligent collection terminal. For example, different similarity thresholds can be used when connecting to the power grid and when using a backup battery to achieve better results.

[0099] Further provided is a station water data intelligent collection system, comprising the above-mentioned intelligent collection terminal, and a host computer connected to the intelligent collection terminal via a communication network;

[0100] The host computer includes an intelligent acquisition terminal management module, which receives data from each intelligent acquisition terminal in each working cycle; the intelligent acquisition terminal management module is used to:

[0101] Collect the location information of each intelligent collection terminal and mark them one by one in the GIS system of the predetermined area; build the geographic spatial topology map of the water network in the area where the intelligent collection terminal is located;

[0102] Based on the flow direction of the water network, the distance between adjacent smart terminals and the regional flow rate, and based on the position of the smart collection terminal on the water network, a terminal directed graph is constructed; based on the distance between adjacent smart terminals and the regional flow rate, a time offset matrix of the smart collection terminal is constructed;

[0103] Based on the terminal directed graph, the coupling relationship of the intelligent collection terminal is mined, and the difference between the theoretical calculation value and the actual measurement value of the downstream intelligent collection terminal is compared; if the difference is within a predetermined range, the intelligent collection terminal is marked as a non-independent terminal.

[0104] To further improve the overall efficiency of the system, the intelligent data collection terminals are comprehensively planned and adjusted through the host computer system. Since the pipe network or water network can be viewed as a connected topological structure, or a directed graph, some nodes are similar to circuit structures and have certain topological constraints. Due to this non-ideal structure, information dissipation may occur. For example, the information θ measured at upstream node a may be less than or greater than θ at downstream node b, but overall, there is correlation. By analyzing data drift or dissipation, relevant data can be obtained, thereby analyzing and verifying the data's authenticity. Furthermore, the necessity of a particular intelligent data collection terminal can be determined based on this analysis. For example, at a confluence node, if data collected by intelligent data collection terminals at locations a and b can be used to obtain data from a downstream intelligent data collection terminal at location c, in some cases, measurements at location c can be eliminated or reduced. For example, data at location c can be measured regularly and analyzed in conjunction with calculated data, eliminating the need for continuous measurement and data collection at location c.

[0105] In other words, system redundancy can be reduced based on system analysis, thereby reducing the workload of some intelligent data collection terminals, improving overall data collection efficiency, and reducing terminal energy consumption. It is important to note that the operating status of each intelligent data collection terminal can be controlled by the host computer to adjust the system. For example, non-independent nodes can be set as Class I control nodes, and independent nodes can be set as Class II control nodes. Different methods of regulation can be used for different node conditions.

[0106] In summary, this application supports various modes of station water data collection to achieve interface compatibility and unification; adopts LoRa (Long Range) communication networking technology to achieve low-power, long-distance station water monitoring data collection; supports various modes of data communication to achieve long-term data monitoring without blind spots; adopts gateway intelligent management technology to achieve scientific management of station water data information, and optimizes the system through the DQN module to reduce energy consumption.

[0107] It should be noted that the various specific technical features described in the above specific embodiments can be combined in any appropriate manner without contradiction. To avoid unnecessary repetition, the present invention will not further describe various possible combinations.

Claims

1. The station water data intelligent collection terminal is characterized by: include: The data acquisition unit is connected to at least two data acquisition devices via a data interface. The network communication unit includes at least two types of communication modules. The core control unit is connected with the data acquisition unit and the network communication unit, and is used for receiving and processing the collected data, and communicating with the host computer through the network communication unit; The core control unit is configured with an offline DQN module and a backup storage module. When working, the offline DQN module obtains the acquisition parameter matrix of the data acquisition unit and the signal parameter matrix of the network communication unit in each working cycle, and obtains the energy consumption state of the data acquisition unit and the network communication unit to construct the energy consumption state matrix; Train the data for a predetermined number of epochs with a preconfigured action space and reward function. The action space includes: signal transmission offset time, signal collection offset time and average occupancy ratio of the standby storage module; The reward function R is: R=-[a(( w i+1 -w i ) / ( w i+1 -w 0 ))+b(( v i+1 -v i ) / (v i+1 -v 0 ))]; In the formula, w i+1 , w i Respectively i+1 , No. i Energy consumption of the network communication unit during the cycle; w 0 is the minimum energy consumption of the network communication unit under theoretical conditions; v i+1 , v i are the energy consumption of the data acquisition unit in the i+1th and ith cycles respectively; v 0 is the minimum energy consumption of the data acquisition unit under theoretical conditions; a and b are set constants; After a predetermined period of data training, the offline DQN module sends the output data matrix to the core control unit, and the core control unit controls the work of the acquisition terminal based on the output data matrix.

2. The intelligent water affairs data collection terminal for a station house as claimed in claim 1, characterized in that: The backup storage module is divided into a number of storage areas of predetermined capacity, the number of which is not less than the number of data acquisition units; The core control unit collects the working data streams of each data acquisition unit in sequence during the first working cycle, and establishes a collection time sequence set of the data acquisition unit; Sequentially collect the signal transmission data streams of each network communication unit and establish a transmission timing set of the data collection unit; Based on the collection timing set and the sending timing set, a working data flow of each data collection unit is constructed to a sending data flow of each network communication unit, and a data flow OD matrix is ​​established; Calculate the proportion of data flow directions in the data flow OD matrix in each working cycle and arrange them in descending order; Calculate the data transmission time ratio in the data flow OD matrix in each working cycle and arrange them in descending order; Based on the data flow ratio and data transmission time ratio, adjust the capacity ratio of each storage area in the backup storage module.

3. The intelligent water data collection terminal for a station house as claimed in claim 2, characterized in that: In the next working cycle, an OSD matrix of each port of the data acquisition unit, each storage area of ​​the spare storage module and each communication module of the network communication unit is constructed, and data streams with a proportion higher than a preset value and a data transmission time ratio higher than a threshold are stored in the spare storage module in proportion, and the time series of the data stream is adjusted. After the time series is adjusted, the power consumption reduction ratio of the data acquisition unit and the network communication unit is calculated; the above process of adjusting the time series of the data stream is repeated until the power consumption reduction ratio reaches the preset value.

4. The intelligent water data collection terminal for a station house as claimed in claim 2, characterized in that: Based on the proportion of data flow and the ratio of data transmission time, each data acquisition device and communication module are graded, and weight matrices of the data acquisition device and communication module are constructed respectively; based on the weight matrix, the data flow time series in the subsequent working cycle is adjusted.

5. The intelligent water affairs data collection terminal for a station house as claimed in claim 3, characterized in that: When the power consumption reduction ratio does not reach the preset value during the training scheduled period, the collection timing set and the transmission timing set are compressed, and the similarity between the compressed local data stream and the uncompressed complete data stream is calculated. According to the similarity between the local data stream and the complete data stream within the scheduled period, combined with the data evaluation effect of the local data stream, the similarity threshold is set; During use, if the power consumption reduction ratio does not reach the preset value, the data resolution of the corresponding data acquisition device is adjusted according to the data flow ratio, so that the power consumption reaches the lowest value within the training cycle while being higher than the similarity threshold.

6. The intelligent water affairs data collection terminal for a station house as claimed in claim 5, characterized in that: The data acquisition device includes a temperature and humidity sensor, a flue gas monitor, a COD monitor and a flow meter.

7. The intelligent water data collection terminal for a station house as claimed in claim 1, characterized in that: The communication module includes a LoRa module, a 4G module and an Ethernet module.

8. The intelligent terminal for collecting water affairs data of a station house according to claim 1, characterized in that: The data interface includes RS232, RS485, ADC and switch quantity.

9. The station water data intelligent collection system is characterized by: The intelligent acquisition terminal comprises any one of claims 1 to 8, and a host computer connected to the intelligent acquisition terminal via a communication network; The host computer includes an intelligent acquisition terminal management module, which receives data from each intelligent acquisition terminal in each working cycle; the intelligent acquisition terminal management module is used to: Collect the location information of each intelligent collection terminal and mark them one by one in the GIS system of the predetermined area; construct the geographic spatial topology map of the water network in the area where the intelligent collection terminal is located; Based on the flow direction of the water network, the distance between adjacent intelligent terminals and the regional flow rate, and based on the position of the intelligent collection terminal on the water network, a terminal directed graph is constructed; The time offset matrix of the intelligent collection terminal is constructed based on the distance between adjacent intelligent terminals and the regional flow rate; Based on the terminal directed graph, the coupling relationship of the intelligent collection terminal is mined, and the difference between the theoretical calculation value and the actual measurement value of the downstream intelligent collection terminal is compared; If the difference is within a predetermined range, the intelligent acquisition terminal is marked as a non-independent terminal.

Images