WO2025088389A1

WO2025088389A1 - Automatic central tuning on a compressed air system

Info

Publication number: WO2025088389A1
Application number: PCT/IB2024/058543
Authority: WO
Inventors: Christophe VERBIST; Peter Verhoeven; Wim Van Roy
Original assignee: Atlas Copco Airpower NV
Current assignee: Atlas Copco Airpower NV
Priority date: 2023-10-23
Filing date: 2024-09-03
Publication date: 2025-05-01
Anticipated expiration: 2026-04-23
Also published as: BE1032082A1; TW202526530A; CN119878511A; BE1032082B1

Abstract

The invention relates to an apparatus and a method for the centralized control of compressed air systems. In particular, the present invention concerns a method and a device for a compressed air system with a centralized controller system providing automatic tuning procedures to implement the most optimal tuning parameters for the specific site. Controlling a compressor system (100) comprises at least one compressor (101-103) configured to provide compressed air or gas. The automatic tuning method comprising the steps of, gathering initial data by monitoring said compressor system, estimating a new set of data through an estimation unit until the estimated data configured to be different from the initial data, compiling a new set of parameters based on the estimated data, verifying the new set of parameters derived from the estimated data, repeating all of the steps until the verification of the new parameters is achieved, apply the new tuning based on the verified parameters, instructing at least one of the elements in a compressor room (113) to perform the actions in accordance with the tuning procedure.

Description

Title of Invention: AUTOMATIC CENTRAL

TUNING ON A COMPRESSED AIR SYSTEM

Technical Field of the Present Invention

[0001] The present invention relates to an apparatus for the centralized control of compressed air systems. In particular, the present invention concerns a method and a device for a compressed air system with a centralized controller system providing automatic tuning procedures to implement the most optimal tuning parameters for the specific site.

Background of the Present Invention

[0002] Air compressors are used in various industries, such as automotive, manufacturing, construction, and more. They are essential for powering tools and equipment, making them an integral part of any industrial process. However, air compressors may have limitations that can negatively impact their efficiency, energy consumption, and productivity. Fortunately, with the rise of smart technologies, air compressors are being revolutionized, offering improved performance, sustainability, and costeffectiveness.

[0003] As a part of smart technologies, artificial intelligence already helps air compressors achieve optimal performance by using algorithms to monitor and notify the users of these systems. The Al algorithms can identify patterns in the system's operation, predicting and preventing potential failures.

[0004] Traditionally, air compressor maintenance is performed on a fixed schedule, which can result in unnecessary downtime and wasted resources. With smart technologies, maintenance can be performed based on the compressor's actual condition. The Al technology can analyze the compressor's data to identify potential issues before they become major problems, allowing maintenance to be performed at the right time.

[0005] Predictive maintenance can save time and money by reducing the need for frequent maintenance checks. It can also prevent unexpected equipment failures, which can lead to costly repairs and downtime. Furthermore, predictive maintenance can increase the lifespan of the air compressor, reducing the need for replacements and lowering costs.

[0006] Al-controlled air compressors can also improve safety in industrial settings. For example, if a tool is not operating correctly, Al algorithms on the compressor can detect this and inform the system to shut down, preventing accidents and injuries. Additionally, Al technology can help with identifying sources of noise with the goal of noise reduction. This is crucial in workplaces where noise levels can lead to hearing damage. Also, noise reduction will prolong the compressor's lifespan by reducing the overall wear due to, for example, vibrations.

[0007] The use of Al in air compressors is still in its early stages, but the potential for growth is enormous. As Al technology continues to evolve, even more advancements in air compressor technology should be expected.

[0008] Air compressors are known for their high energy consumption. The Al algorithms can optimize the compressor's operation, reducing energy usage and costs. Additionally, Al technology can identify when the compressor is not in use and shut it down, reducing energy waste. Energy efficiency is crucial in today's world, where energy costs are constantly rising, and environmental concerns are paramount. By using Al-controlled air compressors, industries can reduce their energy consumption and carbon footprint, making them more sustainable and environmentally friendly. However, these Al algorithms are not always safe to use. They do not provide any guarantees on the system's performance or on the safety of the system. As a result, these systems cannot be used in practice due to safety concerns.

[0009] Another approach is to combine the Al Algorithms with existing and proven technologies. According to the present invention, Al algorithms learn from the compressor's data and adjust its operation to achieve even better efficiency and performance. The performance of multiple compressors, working together in a compressed air system, can be optimized using the central tuning technique which includes Al support in compressor systems. It considers the system instead of looking at the performance of each individual machine. By using a combination of existing and predictable control schemes together with Al technology, a safe and performance system can be designed. Proven technology can still take care of the safety requirements while being tuned to achieve the best possible efficiency by Al algorithms. Furthermore, Al technology can be used to predict future energy demands, adding valuable information for more advanced control approaches to adjust the operation point of the air compressors, accordingly, further reducing energy consumption.

[0010] Configuring every installation in the most efficient way is not a straightforward task. There is a large variety of these systems that make improving the optimization of the systems difficult for humans. Multiple compressors may be connected in a compressed air system to supply compressed air to the same distribution network. These compressors could be of several types, sizes, and ages, which could cause an unequal load and system inefficiencies. There is a large variety of pneumatic networks that can be divided into various categories according to their structure, size, layout, and capacity. Customers may have several requirements for the system in terms of energy efficiency and demand requirements. [0011] To handle the large variety of systems in which a compressor can be configured, there is a plurality of parameters available. These parameters may be technical, such as power, pressure ratio and/or start-up time. Other parameters may include economic parameters, such as investment and/or maintenance costs, or even environmental requirements, such as energy-efficiency and/or noise requirements during operation. Taking all these parameters into account and correcting for the different system layouts is difficult in practice. Service engineers quickly lose the overview of all the possibilities to configure a system.

[0012] To obtain a balanced system and maximum efficiency, central tuning entails, among others, modifying the controls of each compressor considering the interplay between the compressors. Measurements and analysis of the system's performance, including elements such as pressure, flow rate, and energy consumption, is part of the central tuning process. In addition to this, the system properties should be considered that are important to the system such as wearing, humidity, noise, and temperature effects. This is often done as a onetime effort by technicians when installing these systems. However, after time, when machines de gradate or when the system changes, this process needs to be repeated.

[0013] Al assisted central tuning can result in significant energy savings, improved system reliability and reduced maintenance costs. Although machine learning and Al technologies are developing quickly, skilled technicians or engineers with specific expertise and experience in compressed air systems are still generally responsible for doing the central tuning of compressor systems. This is due to a few factors. First, compressor systems can be complicated and require a knowledge of how various parts of a system work together. Second, the central tuning procedure comprises modifying the compressor controls, which necessitates a comprehensive understanding of the unique compressor models and their control schemes. Finally, central tuning involves optimizing the performance of the entire compressed air system, which requires a holistic view of the system and its requirements.

[0014] In the present invention, machine learning and Al technologies are enabled to assist and even take over the complicated tuning process. The central tuning process involves analyzing system data, learning the working principle of the system, verifying the necessary parameters, providing recommendations for adjustments to the compressor controls and executing it.

[0015] By analyzing data from sensors and other sources, Al algorithms can adjust control parameters in real-time to achieve the best possible performance, considering factors such as temperature, pressure, and flow rate. They can identify changes in the system that intrinsically lead to a retuning of the system. This can be done in real-time without the need for external interactions or stopping the system from performing benchmark tests. The real-time tuning efforts can guarantee that the system will always operate at the best achievable efficiency. As a result, the control of a compressor system will be able to work at peak efficiency continuously.

[0016] By building in a verification step, the new tuning parameters are guaranteed to be safe for the system. This avoids downtime or degradation of the system. It allows that the produced compressed air always satisfies the minimum required air quality requested by the customer.

[0017] Further in the present invention, Al algorithms can identify the root cause of a fault in advance. By analyzing data from sensors and other sources, the Al algorithms enable technicians to take corrective action and avoid downtime. Correctly diagnosed faults in compressor systems allows them to be repaired quickly and efficiently. The condition of the compressor is monitored in real-time, and any signs of impending failure are detected. By analyzing data from sensors and other sources, Al algorithms can identify patterns that indicate the need for maintenance, allowing maintenance to be performed proactively rather than retroactively. In a further embodiment, a service technician is notified about the impending failure based on the analysis.

[0018] Additionally, Al can be used to analyze the performance of compressor systems over time, identifying trends and patterns that can help operators optimize their operation. By analyzing data from sensors and other sources, Al algorithms can provide insights into areas where performance can be improved, such as energy efficiency or reliability. It can make suggestions on where additional measures can be taken such as the installation of additional sensors, changes to the system or replacement of the current compressors.

[0019] Overall, Al can be a valuable tool for optimizing the performance of compressor systems, reducing energy consumption, and improving reliability, ultimately leading to cost savings and improved operational efficiency.

[0020] To achieve this behavior, the Al algorithm can optimize and control a large variety of parameters in the system. The tuning parameters for the central tuning system can vary depending on the specific compressor system being used.

[0021] Pressure setpoints parameter determines the desired operating pressure of the compressed air system. The central tuning system adjusts the compressor’s output pressure to match the setpoint and maintain consistent pressure throughout the system. In an embodiment of the invention, the pressure in inlet and outlet points can be monitored by pressure sensors.

[0022] Load/unload setpoints parameters determine when the compressor should start and stop running based on demand for compressed air. The central tuning system monitors the system’s air demand and adjusts the parameters which influence the compressor’s output accordingly to ensure efficient operation. [0023] Cycle time parameter controls how often the compressor runs and rests. The central tuning system adjusts the cycle time to minimize energy consumption while still meeting the system’s air demand while avoiding issues with the compressors such as overheating or condensation.

[0024] The central tuning system can optimize the system's efficiency by adjusting the compressor's output to match the demand for compressed air. This helps to reduce energy consumption and prolong the life of the compressor system.

[0025] By optimizing these kinds of parameters, the central tuning system can ensure that the compressor system operates at its maximum efficiency while still meeting the demand for compressed air.

[0026] The attempts made in the state of the art to alleviate the problems associated with optimized performance of multiple compressors working together in a compressed air system are described in the following patents.

[0027] Chinese Appl. No. CN113741174 discloses a self-adaptive pressure control algorithm of the reciprocating natural gas compressor, when a gas lift compressor works, gathering, transportation and pressurization of natural gas can be achieved by adopting a PID control algorithm and logical judgment on the exhaust pressure of the gas lift compressor. A set of PID controller parameters matched with the compressor need to be set when each well works, and the parameters capable of automatically adjusting the PID controller according to the change of the working condition characteristics of the well sites are added on the basis of the conventional PID controller parameters. The parameters are in an intelligent control algorithm in an optimal state. Online parameter self-correction of the PID controller is achieved through the fuzzy control technology.

[0028] W02023001903A1 discloses a method for providing at least one design configuration of a compressed-air system comprising at least two parallel-connected compressors, wherein the method comprises the following steps. A computer receives component data, wherein the component data comprise a compressor list containing multiple compressors of different types. The computer generates a branched data structure. A computer generates compressed-air system configuration data specifying compressed-air system configurations in which two of the compressors from the compressor list are connected in parallel. The computer calculates at least one quality value for at least one of the compressed-air system configurations based on the compressed-air system configuration and at least one technical parameter of the compressors of the compressed-air system configuration, wherein the at least one quality value specifies the quality of the compressed-air system configuration in relation to a quality criterion that is preferably predefined by a user. The computer provides at least one compressed-air system configuration having in each case at least one associated quality value.

[0029] US20160245273 relates to an electronic control device for a component of compressed-air generation, compressed-air processing, compressed-air storage, and/or compressed-air distribution, wherein the electronic control device falls back upon one or more models, which, as component-related models, contain information relevant to the structure, or the behavior of the component, in order to determine, simulate, or evaluate operation-relevant data and performs, as an evaluation purpose, either — openloop control, closed-loop control, diagnosis, and/or monitoring of the component or — a determination, provision, prediction, or optimization of operating data, operating states, operating modes, operating behaviors, and/or operating effects on the basis of the models in a concrete evaluation routine, and wherein current or historical structure information operating data, operating states, and/or measurements/sensor values of the component at least partially available in the electronic control device are used as initial values.

[0030] The present invention enables tuning process on a separate device or a server and by calculating the optimized parameters without any simulation. A combination of measurement data together with a priori models is used to improve the parameters of the system. In prior art, PI settings are manipulated for monitoring purposes in optimization or iterative adapted model simulations are used.

[0031] The commissioning of compressed air systems is a critical process that ensures their efficient operation and performance. During commissioning, various parameters related to the system's components need to be tuned to achieve the desired output. In prior art, this process has been manual and time-consuming, requiring skilled technicians to make iterative adjustments and validate the results. The present invention addresses this challenge by introducing an automated tuning system that significantly reduces the commissioning time and continuously improves the overall performance of compressed air systems.

Objects of the Present Invention

[0032] One of the objectives of this invention is to provide a way to manage a compressor system optimally and efficiently.

[0033] The object of the present invention is to provide an automatic tuning process for a compressed gas system to determine most optimal parameters contributing to improved performance, lower maintenance cost and more stable control of the system.

[0034] Another object of the present invention is to reduce or eliminate the amount of time a technician needs to spend for the tuning process.

[0035] Another object of the present invention is to provide a central tuning which determines the most efficient parameters regardless of the adjustments in the system and specific conditions in the site, keeping the system in the most optimal working conditions at all times regardless of changes in the system.

Summary of the Present Invention

[0036] The present invention shows an autotuning algorithm that can tune various parameters of the system related to the operation of the system. To streamline the tuning process at commissioning, central tuning process reduces the amount of time a technician needs to spend at a commissioning. Central tuning proposes to auto tune these types of parameters based on algorithms reading the required input from the running control unit. Central auto tuning procedures determines the best tuning parameters for the specific site.

[0037] The present invention relates to an automated tuning system for compressed air systems, specifically designed to streamline the tuning process at commissioning. The system utilizes algorithms and data gathered from a centralized controller to predict and implement optimal tuning parameters for specific sites, thereby reducing the time and effort required by technicians during commissioning. By continuously learning and applying tuning parameters, the system ensures the most efficient operation of the compressed air system, leading to increased energy savings and improved performance.

Figures of the Present Invention

[0038] Accompanying drawings are given solely for the purpose of exemplifying an apparatus for hemostasis testing, whose advantages over prior art were outlined above and will be explained in brief hereinafter.

[0039] The drawings are not meant to delimit the scope of protection as identified in the claims nor should they be referred to alone in an effort to interpret the scope identified in said claims without recourse to the technical disclosure in the description of the present invention.

[0040] Figure 1 is a general perspective view of a compressor system in accordance with this invention.

[0041] Figure 2 is a diagram displaying operational states of a compressor system according to the present invention.

[0042] Figure 3 is a flow diagram showing an algorithm cycle in accordance with this invention.

[0043] Referenced Parts List

[0044] 100 Compressor system

[0045] 101, 102, 103 Compressors

[0046] 104, 105, 106 Local controllers

[0047] 107 Airnet volume [0048] 110 Flow sensor

[0049] 111 Consumer

[0050] 112 Central controller

[0051] 113 Compressor room

Detailed Description of the Present Invention

[0052] Figure 1 illustrates a compressor system 100 according to an illustrative embodiment of the invention. The compressor system 100 comprises a compressor room 113 and a control unit. In an embodiment of the present invention, said control unit comprises a central controller 112 or at least one local controller 104, 105, 106 for operation of the elements in the compressor system 100. Said elements comprise compressors 101, 102, 103 and further, in an embodiment of the invention, the compressor room 113 comprises other elements such as dryers, coolers, sensors, filters, heaters and valves. Said compressor room 113 also comprises an aimet volume 107 for compressed gas as a space within existing piping or a vessel.

[0053] The compressor room 113 comprises local controllers 104, 105, 106. These local controllers 104, 105, 106 are configured to manage the compressors 101, 102, 103. In an embodiment of the invention, a local controller 104, 105, 106 is limited to one compressor and some other elements in the compressor system 100. In some embodiments of the present invention, said local controllers 104, 105, 106 can also control other compressors 101, 102, 103 who are in series connected to the main compressor that they are connected. Managing means that operating parameters or states can be inspected, controlled, monitored, set up and/or verified, or perform any other manipulation to influence the compressor operation. In an embodiment of the invention, the tasks that executed by the central controller 112 can be executed in a remote fashion instead, such as a cloud server. Moreover, each local controller 104, 105, 106 can be internally embedded or external to each compressor 101, 102, 103.

[0054] In an embodiment of the invention, a flow sensor 110 is sending signal to the control unit. Said flow sensor 110 monitors the aimet volume 107.

[0055] Figure 2 demonstrates operational states of the compressor system 100 according to the present invention. The plant where compressed air is used gets signal from the control unit or central control algorithm on a cloud server as auto tuning algorithms. The automated tuning procedure is initiated either manually or based on a predefined time schedule when the system is online. Once triggered, the system proceeds through a series of steps to estimate, verify, and apply the best tuning parameters for the compressed air system.

[0056] The system starts by collecting real-time data from various sensors, as well as static compressor parameters from a P&I Diagram. The data collection phase is essential to detect the current operating conditions of the compressed air system and serves as input for the subsequent tuning steps. Most of the compressors have outlet sensors and some defined properties. These properties can include but not limited to, speed limitations of the motor, valve positions, pressure differences over elements in the compressor. All of this information can contribute to the data which should be gathered for automatic tunning process.

[0057] For each compressor 101, 102, 103, there is a model with parameters which can be tuned in themselves. These parameters include different timings of the state machine, settings related to filtering of the gathered data, tuning of the local controller settings such as parameters of a PI-loop, calibration settings and regulation limits on the pressure, flow, temperature, and humidity measurements.

[0058] A similar set of settings exists for other machines related to the compressor system 100 such as, but not limited to dryers, chillers, coolers, and energy recovery systems.

[0059] There is a certain time needed to transition between states of the compressors 101, 102, 103. The states that participate in the estimation of the transition time are:

[0060] - The main-states: "Stopped (off)", "Unloaded (on but not producing)" and "Loaded

(on and producing)"

[0061] - The substates: The main states can further be divided into two other states a preparation-substate and a ready-substate. In the preparation-substate, the compressor is already in its main state but there are some auxiliary processes going on preventing an immediate change to another state.

[0062] - The transition states between any of the main-states, these are: starting, stopping, loading, unloading, loading from stop, stopping from load.

[0063] In addition to these states, there is a shutdown state for which no estimations are made, including the transition state from any state to shut down. The transition from shutdown to stop can be included in the transition estimation if needed.

[0064] Using the gathered data and historical information from previously installed sites, the system employs sophisticated algorithms to estimate the best set of tuning parameters for the specific site and current operating conditions. Several methods can be used, including neural networks, optimization algorithms, heuristic approaches designed by engineers, or system identification techniques. In an embodiment, the optimization technique is a moving horizon estimator (MHE).

[0065] In an embodiment of the invention, neural networks are trained with unsupervised learning and/or deep reinforcement learning.

[0066] Unsupervised learning techniques such as clustering, anomaly detection, and principal component analysis (PCA) can be used to detect faults in compressor systems. These methods can be applied to sensor data from the compressor system to identify patterns that deviate from normal operating conditions, indicating a potential fault. [0067] Also, deep reinforcement learning can be used to optimize the control of compressor systems, improving efficiency and reducing energy consumption. In an embodiment of the invention, the compressor system is modeled as an environment, and the DRL agent learns to take actions (e.g., adjusting the speed of the compressor) to maximize a reward signal (e.g., minimizing energy consumption). By learning from experience and exploring different control strategies, the DRL agent can learn to achieve better performance than traditional control methods.

[0068] Further, unsupervised learning techniques such as PCA and clustering can be used to analyze large amounts of sensor data from compressor systems to identify patterns and trends in performance. By analyzing the data in an unsupervised manner, it is possible to identify patterns that may not be obvious using traditional analysis methods. This information can be used to improve the performance of the compressor system, optimize maintenance schedules, and identify opportunities for energy savings.

[0069] Overall, both unsupervised learning and DRL are valuable tools for the present invention in the analysis and control of compressor systems, providing insights into performance and opportunities for optimization.

[0070] The algorithm also has as a reward function a combination of KPIs and can control the tuning parameters of the central controller 112 and local controllers 104, 105, 106. The KPI parameters can consist of the metrics below:

[0071] - Metrics for energy consumption (energy consumed, unloaded hours, total blow-off flow, average network pressure drop and so on)

[0072] - Metrics for wear (total running time, total time at low rpm (since this can be dangerous for some compressors with integrated dryers), total starts/stops, total unload cycles, average acceleration, vibrations, exposure to humidity, and so on)

[0073] - Metrics for maintenance cost (total operation time, maximal duration of standstill per unit, exposure to heat, and so on)

[0074] - Metrics for overall performance of the airnet volume 107 (pressure deviation from setpoint or from the band, number of aimet shutdowns, violations on airnet humidity and so on)

[0075] The estimated tuning parameters are subjected to a verification process to ensure their effectiveness and stability. There are multiple criteria for verifying the parameters, including:

[0076] - Checking whether the parameters remain consistent for a new set of data, indicating stability.

[0077] - Simulating the system's performance with the new parameters and comparing the results to ascertain energy savings.

[0078] - In an embodiment of the invention, involving the expertise of a service technician who may provide approval or feedback on the proposed parameters. [0079] - Passing through a safety filter based on learned models from laboratory and field sites.

[0080] If the verification process confirms the suitability of the calculated tuning parameters, they are either presented to the user for approval or directly applied to the compressed air system without user intervention. In cases where remote monitoring and control are possible, the application can be performed remotely, eliminating the need for on-site visits.

[0081] For a new set of parameters, a small simulation can be started. If the new set of parameters provide better results in the simulation compared to the old parameters, new set of parameters are switched with the old parameters. The uncertainty error of the simulation is applied as a tolerance measure. This depends on the site and the simplifications of the simulator.

[0082] The auto tuning procedure is designed to be iterative, continuously monitoring and learning from the system's performance. If the applied tuning parameters are not optimal or the system's operating conditions change significantly, the auto tuning process repeats, ensuring the compressed air system consistently operates at its peak efficiency.

[0083] The automated tuning system can adjust a wide range of parameters to optimize the compressed air system's performance.

[0084] The system can tune the compressor system to maintain the correct pressure required for the specific application. This involves setting optimal operating pressures, considering safety margins and implementing automatic backup plans in the event of any issues within the control chain.

[0085] Each compressor 101, 102, 103 has various tunable parameters, including state machine timings, data filtering settings, and tuning of local controller parameters like PI-loop parameters. Additionally, calibration settings and operation limits for pressure, flow, temperature, and humidity measurements can be fine-tuned.

[0086] The control unit plays an essential role in deciding which machines run at which operation point. The auto tuning system can adjust parameters that influence the control unit’s actions, such as the set of running machines, switching timings, and control action reaction speed.

[0087] Controlling valve positions within the compressed air delivery system significantly impacts efficiency. The automated tuning system optimizes valve positions to enhance the system's overall performance.

[0088] The automated tuning system controls the compressor room 113 that houses at least one compressor 101, 102, 103. Each compressor is equipped internally or externally with the local controller 104, 105, 106 that communicates with the control unit. [0089] Figure 3 demonstrates that the auto tuning logic follows a structured and iterative approach to optimize the compressed air system continuously. It involves data gathering, estimation, verification, and application of tuning parameters, ensuring that the system adapts to changing conditions.

[0090] The automated tuning system can adjust a comprehensive set of tuning parameters, including local compressor parameters (e.g., Pl-tuning of speed and valve positions), physical parameters (e.g., vessel volume and airnet modeling), and a wide range of operating parameters. The physical parameters include the properties of the installation, such as the piping properties, volumes, delays within the system, reduction of safety levels. The control parameters include local machine parameters as well as central control parameters. The local machine parameters include Pl-tuning of speed and valve positions, delay times, maximal ramping or drop velocities, properties regarding the filtering of sensor data and other control parameters that can be found on a local controller. It can also include model updates of the local machine, such as surge and choke limits, pressure drops, and flow and power models. The central control parameters are important to operate the system as efficiently as possible. These parameters often include delay times, parameters regarding the robustness of the system (e.g. safety margins), penalties to prevent unwanted behavior (e.g. shutdowns or pressure drops), and parameters to improve the overall cost. Also, here, model updates of the global system are possible.

[0091] The verification to check if a new parameter can be applied consists of different rules depending on the installation. In an embodiment of the invention, the rules can include: [0092] • The new parameters should fall within a specific range or safe set.

[0093] • The relationship between the new parameters should be acceptable. For example, this can be a ratio that should fall within a specific range, a sum that should be large or small enough.

[0094] • A heuristic algorithm should approve the parameters.

[0095] • The new parameters should be close enough to the original parameters.

[0096] • The new parameters should be stable for at least a minimum time (or within a predefined range for a minimum time in the past).

[0097] In brief, the automated tuning system for compressed air systems described in this application provides an automated approach to commissioning and optimizing the performance of such systems. By utilizing advanced algorithms and historical data, the system ensures that each installation operates at its peak efficiency, leading to substantial energy savings and reduced technician involvement during commissioning. The flexible nature of the system allows it to continuously adapt and improve its tuning parameters, making it an invaluable tool for achieving optimal performance in various compressed air applications.

Claims

[Claim 1] - A method for improving the control performance of a compressor system (100) comprising a control unit and at least one compressor (101-103) configured to provide compressed air or gas, said method comprising the steps of:

- gathering initial data by monitoring said compressor system (100),

- estimating a new set of parameters through an estimation unit until the estimated parameters configured to be different from the initial parameters of the system,

- verifying the new set of parameters derived from the initial data,

- repeating all of the steps until the new set of parameters are verified,

- instructing at least one of the elements in a compressor room (113) to apply the new parameters and accordingly perform the required actions.

[Claim 2] - The method for controlling a compressor system (100) of claim

1, wherein the control unit is a central controller (112) which is configured to manage at least one element in the compressor room (113).

[Claim 3] - The method for controlling a compressor system (100) of claim 1, wherein the control unit is at least one local controller (104-06) which configured to manage at least one element in the compressor room (113).

[Claim 4] - The method for controlling a compressor system (100) of claim

3, wherein said local controllers (104-106) are configured to work together in a distributed manner.

[Claim 5] - The method for controlling a compressor system (100) of claim

1, further including the step of gathering more data until the sufficient data volume is achieved in said gathering initial data step.

[Claim 6] - The method for controlling a compressor system (100) of claim 1, further including the step of simulating the new set of parameters derived from the estimated data to compare the results with the initial parameters.

[Claim 7] - The method for controlling a compressor system (100) of claim 1, further including the step of generating the new set of parameters derived from the estimated data according to a set of predefined rules to compare the results with the initial data parameters.

[Claim 8] - The method for controlling a compressor system (100) of claim 1, further including the step of generating the new set of parameters derived from the estimated data according to a supervisory Al system to compare the results with the initial data parameters.

[Claim 9] - The method for controlling a compressor system (100) of claim

1, wherein the initial data comprises static compressor parameters.

[Claim 10] - The method for controlling a compressor system (100) of claim

1, wherein the initial data comprises the data collected by various sensors, provided through monitoring of the compressor system (100) by collecting real-time data.

[Claim 11] - The method for controlling a compressor system (100) of claim

10, wherein Al algorithms utilized to identify patterns indicating the need for maintenance, notifying a service technician about the impending failure based on the analysis.

[Claim 12] - The method for controlling a compressor system (100) of claim

1, wherein the estimating procedure is triggered manually by a user or automatically on a predefined time basis.

[Claim 13] - The method for controlling a compressor system (100) of claim

1, wherein the algorithms for estimating the optimal tuning parameters include neural networks, optimization algorithms, and/ or system identification techniques.

[Claim 14] - The method for controlling a compressor system (100) of claim

13, wherein the neural networks are trained with unsupervised learning and/or deep reinforcement learning.

[Claim 15] - The method for controlling a compressor system (100) of claim 1, wherein the tuning system adjusts parameters including pressure settings, individual machine settings, central controller settings, and/or compressed air delivery system settings.

[Claim 16] - The method for controlling a compressor system (100) of claim

15, wherein the pressure settings are adjusted to maintain the correct pressure required for the specific application, considering safety margins and/or implementing automatic backup plans.

[Claim 17] - The method for controlling a compressor system (100) of claim 15, wherein the individual machine settings include tuning state machine timings, data filtering settings, local controller parameters; calibration settings for pressure, flow, temperature, and/or humidity measurements.

[Claim 18] - The method for controlling a compressor system (100) of claim

2 and claim 15, wherein the central controller settings influence the set of running machines, switching timings, and control action reaction speed within the compressed air system.

[Claim 19] - The method for controlling a compressor system (100) of claim

15, wherein the compressed air delivery system settings involve tuning valve positions.

[Claim 20] - A compressor system (100) configured to compress and supply a gas comprising a central controller (112) configured to manage at least one compressor (101-103), each compressor (101-103) equipped with local controllers (104-106) characterized in that;

- the compressor system (100) further comprises a compressor room (113) configured to provide initial data to said central controller (112 )

- an estimation unit configured to estimate a new set of data until the estimated data configured to be different from the initial data,

- said central controller (112) configured to compile a new set of parameters based on the estimated data of estimation unit and simulating the new set of parameters derived from the estimated data to compare the results with the initial data parameters,

- a verification module which configured to trigger repetition of the previous steps until the verification of the new parameters is accepted,

- an application module of the central controller (112) configured to apply verified tuning parameters to the compressor system (100),

- instructing at least one of the elements in a compressor room

(113) to perform the actions in accordance with the verified tuning parameters.

[Claim 21] - The compressor system (100) according to claim 20, wherein the central controller (112) configured to trigger the estimating procedure either manually or on a predefined time basis.

[Claim 22] - The compressor system (100) according to claim 20, wherein the central controller (112) configured to trigger the estimating procedure in a continuous fashion. [Claim 23] - The compressor system (100) according to claim 20, wherein the algorithms for estimating optimal tuning parameters include neural networks, optimization algorithms and/or system identification techniques.

[Claim 24] - The compressor system (100) according to claim 20, wherein the tuning parameters adjustable by the tuning system encompass pressure settings, individual machine settings, central controller settings, and/or compressed air delivery system settings.

[Claim 25] - The compressor system (100) according to claim 24, wherein the pressure settings are adjusted to maintain the correct pressure required for the specific application, considering safety margins.

[Claim 26] - The compressor system (100) according to claim 24, wherein the individual machine settings include tuning state machine timings, data filtering settings, local controller parameters; calibration settings for pressure, flow, temperature, heat flow and/or humidity measurements.

[Claim 27] - The compressor system (100) according to claim 24, wherein the central controller settings influence the set of running machines, switching timings, and/or control action reaction speed within the compressor system (100).

[Claim 28] - The compressor system (100) according to claim 24, wherein the compressed air delivery system settings involve tuning valve positions.