Intelligent fluid delivery control device

Abstract

The "Intelligent fluid delivery control DEVICE" is a two-way device, ie. one input and two outputs, equipped with a flow inhibitor capable of analyzing the fluid conducted in real time through indirect measurement sensors governed by a microcontroller that allow the detection and / or measurement of various physical quantities, as well as carrying out local processing of the detected data on a temporal basis in order to identify a sudden variation in the values of the observed quantities which could trigger semi-autonomous control actions

Description

[0001] Intelligent fluid delivery control DEVICE

[0002] DESCRIPTION

[0003] The "Intelligent fluid delivery control DEVICE" is a pass-through device based on the use of microcontrollers(l), direct and / or indirect detection sensors(2), diverters(3) or intelligent actuators and valves(4), which uses an internal conduit(5) - or a portion of it - composed of a suitable material and different configurations or combinations of parts in the assembled components specifically and present in the state of the art, but through a specific know-how and whose realization is classified and is carried out according to the claims listed below and intended for joint use with sensors for indirect detection of various substances and / or materials, variable depending on the specific use and the fluid carried therein, as well as for the detection and / or measurement of various physical quantities such as, for example, pressure, flow, temperature, quantity of fluid delivered, etc.

[0004] It is different and here is the "quid novi" in terms of utility model compared to the pre-existing "state of the art" and compared to the current purification model systems to eliminate the impurities of the fluids, because this system is not a simple filtering of the impurities of the fluids, but is able to interpret and analyze in real time (so-called RT ) and in an intelligent way, thanks to the use of edge Al techniques, i.e. Artificial Intelligence, implemented by existing components on the market and combinations of material parts in the right proportions and sizes , assembled in an original and innovative way creating a microcontroller(l), created in a highly effective way and easily usable and employable in every industrial sector, the physical characteristics of the fluid conducted there and eliminate them, where the simple filter or human conduct does not reach, as well as being able to detect in a semi-autonomous way, i.e. through ML / AI functionality of the microcontroller component, the presence of certain substances that vary according to the specific use and its programming, with the ability to activate autonomously both the flow delivery inhibition valves (4) and the intelligent diverter (3) to bypass the main delivery / inlet channel.

[0005] With the Italian ministerial Decree of 14 June 2017 , which modifies Annex II (Control) and Annex III (Specifications for the analysis of parameters) of the Legislative Decree of 2 February 2001, n. 31, the Water Safety Plan (WSP or PSA) was implemented, a global system, introduced in 2004 by the WHO, based on the evaluation and management of the risk associated with each phase that makes up the water supply chain, from collection to end user, to ensure the protection of water resources and the reduction of potential risks to human health in water intended for human consumption.

[0006] The decree requires the implementation of WSPs for all operators of drinking water systems, large and small, in order to strengthen the prevention and control of any potential hazardous events in the water supply chain (including the environment in which water is collected).

[0007] The WSP redefines the limits of quality control systems for water intended for human consumption, shifting focus from the current surveillance of limited segments of the water supply chain (collection, treatment) and from random monitoring of distributed water, towards the implementation of a holistic risk assessment and management system extending to the entire water supply chain, from collection to the user tap.

[0008] The quality of water intended for human consumption is guaranteed by applying the provisions contained in Legislative Decree 18 / 2023 IMPLEMENTING DIRECTIVE (EU) 2020 / 2184.

[0009] The new EU Directive introduces revised rules to protect human health from the adverse effects of contamination of water intended for human consumption, and also:

[0010] • establishes the hygiene requirements for materials, objects, chemical reagents and filtration and treatment media that come into contact with drinking water;

[0011] • improves access for all to water intended for human consumption, in particular by ensuring access for vulnerable and marginalized groups, improving access for those who already benefit from it and promoting its use at the tap;

[0012] • introduces the risk assessment and management of river basins for drinking water abstraction points and supply systems, as well as the risk assessment of domestic distribution systems;

[0013] • improves transparency on water issues and consumers' access to up-to- date information.

[0014] Drinking water, regardless of its category, must be wholesome and clean, meaning it must not contain microorganisms or other substances in concentrations that pose a potential threat to human health. The proposed system goes beyond TDS, which stands for Total Dissolved Solids.

[0015] This item measures minerals, salts and metals present in water, through the DEVICE equipped with artificial intelligence, innovative and original which has its synthesis in the claims that this new system intends to resolve.

[0016] The filters currently on the market are not able to completely eliminate fluid impurities, because a residual part of them always remains inside them, regardless of the filter installed. However, the proposed system, adapted to all fluid delivery systems and applied in any industrial phase, and through a series of microcontrollers, created with detailed and secret know-how and the use of a combination of elements and raw materials, different but originally assembled and existing in the state of the art, completely eliminates these impurities which in nature remain inside the fluid, regardless of the filtering systems. This intelligent diversion method overcomes the settings of simple filtering systems currently on the market which do not absolutely eliminate the problem.

[0017] A) The sensors perform indirect measurements without intervening on the fluid carried there (unlike filters)

[0018] B) They perform a measurement of the flow and the quantity of fluid delivered by taking a pressure / temperature measurement upstream and downstream of the device: any thermodynamic phenomenon that can influence, directly or indirectly, the change in concentration of the conducted fluid can be reconstructed starting from the aforementioned detections / measurements through the realizable device that analyses them and identifies a greater accuracy on the quantity of fluid delivered (and consequent billing by the service manager to the user)

[0019] C) The deployment of several fluid-control devices, well-assembled and with the right combination of secret materials, within the fluid distribution infrastructure allows, in real time, the acquisition of any information on the sections of the system in question, also for the purpose of identifying pipeline sections where there are leaks or the quantities of fluid conducted with the related physical quantities (i.e. temperature, pressure, volume, etc.): this is an excellent use for the distribution service manager who can thus monitor the status of its infrastructure / distribution network in real time.

[0020] DEVICE FUNCTIONALITIES AND FEATURES

[0021] The "Intelligent fluid delivery control DEVICE" is a pass-through device, with a combination of specific materials, specifically designed for use in both the civil (i.e. in the terminal phases of the network such as homes, offices, micro and smallmedium enterprises or end users in general) and industrial (e.g. laboratories, plants, premises, etc.) sectors through the application of artificial intelligence and the assembly of united and specific materials aimed at creating a microcontroller designed to create fluid control and the interception, diversion and elimination of all elements and foreign bodies existing within the fluid.

[0022] This DEVICE is capable of managing and interpreting a vast quantity of fluid substances, among which the following can be listed, by way of example, in its various industrial applications, but not limited to: water, fuels in general, including those of vegetal origin and / or eco-sustainable, methane and / or natural gas including mixtures in general, vegetable and / or synthetic oil or fluids produced from the transformation of vegetal and non-vegetal raw materials (including seeds and / or mixtures thereof), wines and beverages in general, including those produced from the transformation and / or distillation of alcoholic and non-alcoholic vegetal materials, milk and fluids used / produced in the transformation of milk or vegetal- origin beverages including their mixtures, etc.

[0023] The "Intelligent fluid delivery control DEVICE" is a pass-through device created through the assembly of a series of components of materials and / or technological artefacts existing in the state of the art whose assembly is absolutely secret and falls within the technical KNOW-HOW of the materials used for this purpose: a system based on a microcontroller (1) called MCI), which implements all the detection and control functions of the device, communication and real-time processing of data relating to the operational functioning detected by the sensors (2) and, through the use of edge-AI and machine-learning, implement semi-autonomous (supervised and non-supervised) operating and / or "command and control" models by acting intelligently on the diverters (3) and on the actuators and / or valves (4) in order to interrupt or divert the fluid passing through the duct (5) which connects respectively the inlet valve (6) of the device with the diverter (3) to two outlets (7), one towards the main duct by means of the intelligent valve (8), the other towards the secondary line (9) controlled by the intelligent valve (10).

[0024] The internal duct (5) - or at least a portion of it - is made of a suitable material, the result of a combination of elements, not disclosed nor public and attributable to the secret know-how for its construction, for joint use with sensors (2) for direct and / or indirect detection of different substances and / or materials, which vary depending on the specific context of use and the fluid carried there.

[0025] These intelligent sensors (2) allow the detection and / or measurement of various physical quantities such as, for example, the pressure upstream and downstream of the device in real time, the flow, the temperature, the quantity of fluid supplied, etc., as well as carrying out local processing of the detected data on a temporal basis in order to identify a sudden variation in the values of the observed quantities which could trigger semi-autonomous control actions such as, for example, the signalling or the deviation or inhibition of the flow by closing the intelligent valves (4).

[0026] The material composing the duct (5) can be used through this secret combination of elements in nature which constitute their originality and the specificity of the purposes they intend to achieve. For electromagnetic compatibility reasons, also in order to minimize any external interference on the operating parameters and on the components of the device , a different shelter can be used depending on the fluid conducted there and the specific use .

[0027] The MCI) interfaces with the sensors via a cabling / BUS system (e.g., one-wire, i2c, etc.) or wirelessly (e.g., Bluetooth, Wi-Fi, etc.), with the option of communicating in secure and / or encrypted mode, including using standard and proprietary encryption / coding systems integrated directly into the MCI) firmware. These LAN communication capabilities can also be used for connectivity to extended WAN data networks, for example, to expose the device's operational services or for integration with back-end / remote systems for system control, including via a cellular / LTE network (e.g., GSM, 2G / 3G, 4G, 5G, 6G, etc.).

[0028] This device has the ability, thanks to the basic and / or application functions of the MCU, to perform the "discovery" of similar and / or compatible devices present in the LAN (both wired and wireless). This allows it, by exploiting specific "probing" or "signaling" functions, to be able to configure itself in semi-automatic mode as a single infrastructure without the need for further back-end functions, provided that the devices involved can share some specific parameters known a priori - which will uniquely identify this infrastructure - preset during the initial configuration phase of the systems involved.

[0029] The identified infrastructure will be controlled by a master MCU, according to criteria set during the configuration phase of the participating devices; the various MCUs communicate locally with each other in a secure manner using the aforementioned functions. The master MCU is responsible for controlling the identified local infrastructure in a semi-autonomous, intelligent manner via the identified local LAN network, or even via the extended WAN or cell ular / LTE network, including through interaction with a remote back-end system securely connected to the network in question. However, the components, only in the appropriate proportions and quantities needed to achieve the desired result, are known only to the owners implementing the prototypes.

[0030] The above-mentioned features are accessible remotely via API, or directly through a dedicated APP and in any case capable of implementing a communication protocol for loT (so-called Internet of Things) - whether standard or non-standard - based on messaging in order to make the system interoperable with any remote device / system: such a standard lightweight messaging protocol for the loT world could be, for example but not limited to, the MQTT protocol which is a standard ISO push-subscribe protocol.

[0031] DEVICE Features:

[0032] 1. Initial basic configuration: defines the basic parameters relating to the usage characteristics and any topological parameters for the implementation of infrastructures ("cluster" mode)

[0033] 2. Configuration of fluid type, delivery and quality parameters, including types of detections and substances, physical quantities and units of measurement

[0034] 3. Configuring operating parameters and thresholds for trigger events

[0035] 4. Dataset configuration for training and testing of neural network operation, configuration of machine learning parameters for operational control with dynamic real-time determination of critical situations with alarm / alert generation and configuration of RPA processes and interoperability and communication protocols

[0036] 5. Configuration of operating parameters for standard real-time operation, definition of minimum and maximum threshold values for the selected fluid, configuration of physical parameters (e.g. temperature, pressure, volume, flow, etc.) and normalization and local processing criteria with historicization for quality purposes

[0037] 6. Configuration of deviation conditions parameters, i.e. inhibition and / or reduction of fluid delivery in supervised, semi-supervised and autonomous modes

[0038] 7. Dynamic / real-time configuration of the operating parameters of intelligent indirect sensors for the detection of the presence of anomalous substances and the management of related exceptions

[0039] 8. Configuration of communication functions, parameters relating to the LAN / WAN / LTE cellular network, including the setting of encryption systems, i.e. symmetric / asymmetric encoding / decoding, both standard and proprietary

[0040] 9. Manage security and authorization policies based on the User / Role / Privilege scheme

Claims

CLAIMS1. The device is composed of a transit tube made of different compatible materials and using a series of materials and / or technological artefacts combined with each other, a synthesis of the secret technological knowhow that characterizes the originality of the same, and with sensors for indirect detection, including optoelectronic ones and / or those integrating technologies based on lasers (coherent sources) and / or reflection / refraction of photons, spectrum analyzers, as well as mixers of the sampled fluid with contrast media, solutions and / or solvents; it is also composed of an electro-mechanical deviation structure controlled by the microcontroller, moreover it constitutes the synthesis of components, technological artefacts and assembled materials that can be conveniently and effectively applied in various industrial production processes, including via an actuator located downstream of the device for the possible diversion of the outgoing fluid or for the interruption of the supply; it is also equipped with a microcontroller system with computing capabilities, edge-AI, data management and storage, including in secure mode through the use of standard or proprietary encryption and / or encoding-decoding systems, as well as wireless communication capabilities; it is also equipped with intelligent devices / sensors for measuring physical quantities such as, but not limited to, pressure, flow, and temperature, located upstream and downstream of the device;2. The device referred to in the previous point is also governed by a basic and application system, capable of interfacing with mobile apps and back-end systems, which implements indirect measurement techniques for physical parameters such as temperature, pressure, and the quantity of fluid dispensed, using Al / machine learning techniques following neural network training on an appropriate dataset. The sensors perform indirect measurements without intervening on the fluid being conveyed (unlike filters) based on a combination of innovative assembled materials that are combined and adapted to achieve the goal of diverting impurities contained in liquids.

3. This set of sensors / microcontrollers performs - based on a combination of innovative assembled materials that are combined and adapted - a measurement of the flow and quantity of fluid delivered by carrying out a pressure / temperature measurement upstream and downstream of the device: any thermodynamic phenomenon that may influence, directly or indirectly, the change in concentration of the conducted fluid can be reconstructed starting from the aforementioned detections / measurements through a registration of the combined materials for greater accuracy on the quantity of fluid delivered (and consequent correct invoicing by the service manager to the user and of the actual quantity of fluid, free from impurities and foreign bodies,administered according to the different industrial applications) and on the relative degree of quality objectively detected in real-time.

4. Furthermore, the deployment of various fluid-control devices within the fluid distribution infrastructure allows for real-time acquisition, thanks to the communication and application cooperation capabilities of the individual devices and based on a combination of innovative assembled materials that are combined and adapted, of any information on the sections of the system in question. This can also be used to identify pipeline sections with leaks or the quantities of fluid conducted with the related physical quantities (i.e., temperature, pressure, volume, etc.). This is an excellent use for the distribution service manager, who can thus monitor the status of its distribution infrastructure / network in real time, also through the objective and qualitatively valid acquisition of the related operational information.

5. Distinctive elements compared to existing solutions1. Tightly coupled hardware-software integration for fluid control: The system combines dedicated Al accelerators (embedded NPU / TPU) and optimized runtime directly with the measurement chain: raw data from plastic fiber optic SPR sensors (which measure refractive index changes induced by dissolved substances) are processed at low latency (<10 ms) to produce near-real-time concentration estimates. The tight co-design between optical acquisition hardware, pre-processing FPGA, and Al inference minimizes jitter and ensures reliable responses even in the presence of mechanical vibrations or thermal fluctuations. ii. Adaptive Multi-Range Quantization: To optimize accuracy and throughput, the model supports variable 4— 16-bit quantization based on the test type (e.g., ppb vs. vol. percent detection), automatically activated based on the estimated contamination profile. Compared to platforms using static 8-bit quantization, this approach reduces measurement errors by up to 30% without impacting performance. iii. Run-time profiling and auto-tuning : An integrated diagnostic module monitors latency, estimation accuracy and environmental conditions (sensor temperature, fluid pressure) in the background, dynamically adapting sampling rates and preprocessing parameters (digital filters, normalization) to prevent measurement drift and thermal throttling.

6. Customizable training i. On-device and on-edge retraining for specific water types: Using an SDK compatible with TensorFlow Lite and PyTorch Mobile, you can collect water samples in the field (e.g., industrial wastewater, regional drinking water) and run fine-tuning sessions directly on the microcontroller or edge gateway. This happens without transferring sensitive data to the cloud, preserving privacy and allowing you to model the Al response to various specific contaminants (heavy metals, hydrocarbons, nutrients).ii. Automated validation pipeline: Each newly generated model goes through automatic validation modules (k-fold cross-validation, contaminant class confusion matrix) and produces detailed reports on accuracy / latency trade-offs, making it easy to deploy only the optimal weights. iii. Proprietary Dataset APIs: C / C++ and Python APIs allow you to easily integrate new data streams (indirect sensor measurements such as flow and pressure), extending the training database with multimodal features to improve robustness under variable operating conditions.

7. Secure OTA upgradeability i. Firmware, model, and quantization tables are versioned and signed: Each component (bootloader, RTOS firmware, Al weights, quantization parameters) is digitally signed with an ECC-secp256rl signature. Upon OTA arrival, the bootloader verifies the chain of trust before writing to flash, preventing unauthorized installations. ii. Dual-bank flash and transparent rollback: The controller maintains two independent partitions: one running and one standby for updates. If the new firmware or model fails integrity tests (diagnostic self-test or inference test on a small dataset) upon reboot, the system automatically rolls back to the working release. iii. Encrypted and authenticated communication: Updates travel via TLS mutual-TLS, with packets encrypted in AES-256 GCM and HMAC SHA-256 to ensure confidentiality and integrity even in non-isolated industrial networks.

8. Hardware modularity i. Plug-and-play interfaces (l2C, SPI, UART) for direct and indirect sensors:Hot-swappable connectors provide digital buses, power lines, and GPIOs. This makes it easy to add or replace a fiber-optic surface plasmon resonance (SPR) module or a MEMS-based flow / pressure sensor without redoing the main PCB layout. ii. Automatic recognition and configuration: At boot the firmware queries each module via EEPROM or l2C ID descriptor, dynamically loading the appropriate drivers and setting pre-recorded calibration parameters. iii. Coordinated infrastructure: Multiple units can operate in clusters: a master unit coordinates flow diversions across multiple electromagnetic valves, synchronizing readings and Al decisions. This enables the creation of fluid control networks for large systems (smart irrigation, wastewater management), where each node optimizes locally but contributes to a centralized network logic.

9. Advanced energy management i. Dedicated power manager with configurable states: Sensory and inference activity profiling automatically defines "Deep-Sleep" (low- frequency SPR sampling), "Active" (full inference on contaminants) and"Burst" (maximum reaction speed in case of alarm) profiles. ii. DVFS and clock-gating: The Al coprocessor scales voltage and frequencies based on the load: under steady-state flow conditions the system reduces power by up to 50%, instantly raising it again in the presence of composition variations reported by the SPRs. iii. Energy harvesting for remote nodes: For applications in sites without an electricity grid (monitoring stations upstream of reservoirs, pipelines in natural areas), a PMIC with MPPT for solar panels or micro-hydraulic turbines is integrated, charging supercapacitors or Li-Ion batteries with smart algorithms that prioritize uptime and critical response. iv. Infrastructure consumption policy: in the mode In the "infrastructure" operating mode, where the functions of multiple fluid-control devices are integrated in a combined and coordinated manner, each node communicates its charge status and workload to the others according to specific interoperability policies and protocols. The network adapts the collective sampling frequency, balancing autonomy and monitoring coverage, and delegating inference tasks to the nodes with greater residual resources .

Images