CN116583907A - Predictive systems and methods for active intervention in chemical processes - Google Patents

Abstract

Various embodiments of the present disclosure relate to active dosing optimization chemical feed units from which an output solution (e.g., an oxidizing biocide) is produced. An in-line sensor (142) generates a signal corresponding to a variable measured directly for the corresponding process kit part. Information is selectively retrieved from models that relate combinations of input variables to respective industrial process states, where various current process states may be indirectly determined based on directly measured variables for respective system components. Based on the indirectly determined process state, an output feedback signal corresponding to the detected intervention event is automatically generated. The controller may receive the signals and implement, for example, adjustments to the oxidizing biocide feed to optimize the end product and/or performance metrics.

Description

Predictive systems and methods for active intervention in chemical processes

Technical Field

The present invention relates generally to predictive systems and methods for use in chemical processes (e.g., oxidative biocide generation). More particularly, embodiments of the invention as disclosed herein relate to systems and methods for predictively alerting a user via data analysis or implementing automated intervention in a chemical process.

Background

Chemical feed slips (skids) are routinely implemented by those skilled in the art to reliably and consistently dose chemicals into various types of industrial process applications. An illustrative but non-limiting example of an oxidizing biocide solution generated by a chemical feed chute is Monochloramine (MCA). As an alternative to conventional chlorination processes, MCA is implemented for controlling microbial growth across a variety of industries including, for example, pulp and paper industry, food and beverage processing, water treatment and industrial waste treatment facilities, and the like. Monochloramine is much more stable than free chlorine, does not dissipate so rapidly, and has a much lower propensity to convert organic materials to chlorohydrocarbons (such as chloroform and carbon tetrachloride). Monochloramine does not evaporate into the environment but remains in solution when dissolved in aqueous solution and does not ionize to form a weak acid. This property is responsible, at least in part, for the biocidal effect of monochloramine over a wide pH range.

One example of a conventional process for producing monochloramine includes mixing an oxidizing agent (e.g., sodium hypochlorite solution), an amine source (e.g., a mixture of ammonia-containing substances), and water in a particular ratio using a monochloramine generator. The chlorine source used to generate monochloramine may preferably be a solution of industrial bleach (sodium hypochlorite) at a concentration ranging typically between 10 and 16% (calculated as chlorine).

Currently, chemical feed slips of this nature are prone to mineral scale formation inside the system infrastructure, including pipes, mixers, valves, instrumentation, and other devices exposed to the highly alkaline conditions of the chemical components. To combat this phenomenon, many applications use softeners upstream of the system to remove water hardness, or the ability to feed scale inhibitor (anti-scale) products on-line to reduce scale adhesion. However, over time, most units must be acid washed (acid clean) to remove the scale that builds up inside the system. For systems using softeners, the softener must be maintained manually, for example by refilling with salt.

Fouling of the system can typically result in reduced efficiency, reduced capacity, erroneous readings of instrumentation, reduced mixing capability, and/or failure of mechanical components. The overall impact can range from ineffective biological treatment of the final process (end process) to system failure.

Conventional systems lack an on-line method of determining pickling demand using directly and automatically measured system parameters that would then trigger an automated cleaning procedure. The current method of detecting and removing soil from these systems is primarily based on manual visual inspection of the system followed by a cleaning-in-place (CIP) procedure performed manually.

Conventional practice is associated with a number of undesirable limitations. For example, manual detection of the device may be performed too frequently, which results in unnecessary detection and corresponding wastage of time (i.e., field visits, mechanical disassembly, etc.) and effort. Alternatively, if the detection is performed too infrequently, this may result in undetected fouling. Manual detection of upstream conditions (e.g., softeners) also typically results in unnecessary detection, wasting time (on-site visits) and effort, or detection is too infrequent, resulting in reduced softener performance and increased scaling rate (scaling rate) of the equipment.

On-line sensors for direct measurement of pollution (e.g. dirt) are potentially available, but are particularly expensive, difficult to maintain, and unreliable in terms of results.

Conventional methods of implementing sensors to measure softener function have many of the same limitations: they are expensive and/or require manual testing (not on-line).

It is therefore desirable to provide a series of on-line sensors capable of measuring one or more process variables directly from which the contamination status of the chemical feed slide can be indirectly determined, thereby optimizing unit performance by cleaning at appropriate intervals (but frequently, this costs money and not too frequently, which can reduce efficiency and cause operational problems).

It is further desirable to provide a feedback signal based on the determined contamination status of the chemical feed slide (e.g., based on a contamination event as an automatic trigger) to perform an automated cleaning function or adjustment of upstream features, thereby reducing the likelihood of contamination (e.g., scaling) events occurring due to optimizing upstream content (e.g., softener) operation.

Those skilled in the art will further appreciate that the formation of an oxidizing biocide (e.g., monochloramine) depends on precise control of the stoichiometric ratio of two or more chemical precursors, at least one of which contains halogenated materials that are susceptible to decomposition over time, temperature, exposure to sunlight, exposure to contaminants, or other aspects. Proper maintenance of the stoichiometric ratio of the two precursors is critical to the production of the desired oxidizing biocide solution.

Conventional devices that directly measure the concentration of halogenated materials are expensive, inaccurate, unreliable, prone to drift, prone to fouling, and/or require routine maintenance and service. Thus, end users routinely manually take samples of halogenated precursors to measure their concentration in order to manually adjust the volumetric ratio of the precursor pumping rates. This manual solution is undesirable because it results in unnecessary service accesses (service services), time, and effort. It can also result in too slow a response to deterioration (degradation) of the halogenated material concentration or other meaningful changes, where the efficiency of generating the oxidizing biocide solution is reduced such that its effectiveness as a microbial control agent is compromised. Disruption of the stoichiometric ratio of the active ingredient contained within the precursor can result in a significant reduction in the expected biocide generation efficiency, as well as the generation of unsafe/hazardous materials.

In general, it is also desirable to provide systems and methods that include or implement models/algorithms based on measurement of one or more control parameters. Such systems and methods may directly address the problems in the art, including: lack of quantifiability of the incoming water quality variability and its effect on the measurement of the one or more control parameters; measurements of upstream process conditions that lack a response affecting one or more measured control parameters, such as a measurement of the quality of pre-dilution of one or more chemical precursors, ambient conditions, the presence of upstream equipment that accounts for changing the quality of the feed water/precursor, etc.; and lack of scalability to variability in quality, purity, etc. for different precursor suppliers/manufacturers.

Yet another current problem in many industries and applications is the inability to optimize the oxidizing biocide dosing regimen. Existing techniques for measuring the amount of oxidizing biocide in a process in real time, including but not limited to reagent-based and amperometric halogen sensors, and oxidation/reduction potential (ORP) sensors, are often unreliable, require constant cleaning and calibration, are prone to scaling, and are not ion specific enough to work in difficult water matrices. Thus, current "closed loop" techniques that use sensors to measure related process variables and control the feed of oxidizing biocides to a set point often result in under-feed or over-feed of chemicals.

Insufficient feeding of chemicals can lead to process anomalies (process set), including reduced equipment (heat exchangers, cooling towers, coolers, etc.) efficiency, negatively affecting quality parameters of the final product produced (tissues, papers, liners, liquid packaging, etc.), and to outbreaks of pathogenic microbiological agents, which can cause illness in potentially exposed organisms.

Overfeeding of chemical components can lead to damage to process equipment and components via long-term corrosion, which can lead to system breakage and overall unreliability/efficiency degradation. Too much biocide addition can also lead to worker exposure caused by vapors released by the process, resulting in tear or other health problems. Overfeeding can also result in higher costs to the end user and greater potential impact on wastewater treatment and/or emission limits.

Disclosure of Invention

In view of some or all of the foregoing problems and objectives, it is desirable to supplement conventional control methods with advanced analytical techniques and on-line and off-line key performance metrics for processes/applications to develop application-specific models, such as adjusting the dosage of oxidizing biocides and identifying actionable recommendations (including but not limited to overfeed and underfeed conditions) to alert end users. In addition, predictive models according to the present disclosure may be deployed to automatically adjust or enable adjustment of the feed equipment such that, for example, the dosage of biocide is optimized to minimize microbial contamination without overfeeding.

The system as disclosed herein may preferably implement accessible visual graphics, alerts, notifications, etc. via an on-board user interface, mobile computing device, web-based interface, etc. to supplement any automation capabilities with executable suggestions related to the associated process.

Exemplary techniques for predictive model development may include supervised and unsupervised learning, hard and soft clustering, classification, prediction, and so forth.

Various sensors may be implemented in accordance with the present disclosure for real-time knowledge of the amount of oxidizing biocide contained within an application or process. Exemplary such sensors may include, but are not limited to, sensors configured to generate output signals corresponding to pH, oxidation-reduction potential (ORP), chlorine, monochloramine, and the like. Various sensors may be further or alternatively implemented in accordance with the present disclosure for real-time knowledge of parameters that may affect the ability of microbial contaminants to reproduce at high rates. Exemplary such sensors may include, but are not limited to, sensors configured to generate output signals corresponding to values of temperature, flow, conductivity, pH, ORP, etc.

Manually generated or otherwise off-line measurements may be further incorporated within the scope of the present disclosure for determining the amount of oxidizing biocide contained within the application or process and/or system parameters that may affect the ability of microbial contaminants to reproduce at a high rate and/or the amount of microbial contaminants contained within the application or process.

One exemplary objective of the systems, methods, and associated algorithms as disclosed herein may be to implement a novel combination of sensor and online data to learn key performance metrics of an industrial process in real time, which, by optimizing the data, may result in improved operability and performance.

Another exemplary objective of the systems, methods and associated algorithms as disclosed herein may be to provide advanced analytical techniques to develop scalable and reliable control algorithms for specific applications or processes to regulate and control the oxidizing biocide feed as described above, thereby preventing over-and under-feed of the oxidizing biocide.

Another exemplary objective of the systems, methods, and associated algorithms as disclosed herein may be to deploy customized algorithms to an edge device that is capable of controlling and adjusting an oxidizing biocide feed in substantially real-time to meet the objectives as described above.

Another exemplary objective of the systems, methods, and associated algorithms as disclosed herein may be to deploy an updated system configuration from a remote location to an edge device that is capable of controlling and adjusting the oxidizing biocide feed in substantially real-time to meet the objectives as described above.

Disclosed herein are embodiments of a method for dosing optimization of a chemical feed unit that receives at least one input water source and produces at least one output solution. The plurality of in-line sensors generate signals corresponding to the directly measured variables for the respective process components, and data corresponding to the directly measured variables for each of the respective components is transmitted to a remote server. Information is selectively retrieved from a model that associates a combination of input variables with a respective process state in at least one of the chemical feed unit, the output solution, and the at least one input water source. In response to the real-time data corresponding to the directly measured variable for the respective system component, the method further includes indirectly determining a process state in at least one of the chemical feed unit, the output solution, and the at least one input water source, and automatically generating an output feedback signal corresponding to the detected intervention event based on the indirectly determined process state.

In one exemplary aspect of the above-mentioned embodiment, the output solution from the chemical feed unit comprises an oxidizing biocide, such as monochloramine.

In another exemplary aspect of the above-mentioned embodiments, the indirectly-determined process state may comprise a final product quality and/or performance metric corresponding to the amount of oxidizing biocide present, and an output feedback signal may be generated to adjust at least one dosage rate of the oxidizing biocide.

In another exemplary aspect of the above-mentioned embodiment, the variable directly measured by the on-line sensor for the respective chemical feed unit assembly may comprise a measured variable corresponding to one or more of: the pH of the dilute hypochlorite; the pH of the monochloramine mixture; an oxidation-reduction potential; and the conductivity of the inlet water.

In another exemplary aspect of the above-mentioned embodiment, at least one of the models correlates a combination of input variables with a predicted concentration of contaminants in at least a portion of the chemical feed unit. The indirectly determined process state comprises an indirectly determined contamination state of at least a portion of the chemical feed unit, and the output feedback signal corresponds to a detected contamination event based on the indirectly determined contamination state. For example, a contamination event may be detected based on a threshold violation (threshold violation) regarding an indirectly determined contamination state.

In another exemplary aspect of the above-mentioned embodiment, the contamination status of at least a portion of the chemical feed unit is determined indirectly further based on derived variables of the respective chemical feed unit components, the derived variables based on one or more of the variables measured directly by the on-line sensor, the derived variables corresponding to one or more of: hypochlorite dilution rate; hypochlorite volume through the chemical feed unit over time; and the volume of water passing through the chemical feed unit over time.

In another exemplary aspect of the above-mentioned embodiment, at least one of the models correlates the combination of input variables with a predicted true ratio between two or more chemical precursors used to generate the output solution, the indirectly determined process state comprises an indirectly determined active ingredient state for at least one of the two or more chemical precursors, and the output feedback signal corresponds to a detected intervention event based on the indirectly determined active ingredient state. The intervention event may be detected, for example, based on a threshold violation regarding an active ingredient state indirectly determined for at least one of the two or more chemical precursors.

In another exemplary aspect of the above-mentioned embodiment, the active ingredient status of at least one of the two or more chemical precursors is determined indirectly from the derivative variable further based on one or more of the variables directly measured by the on-line sensor.

In another exemplary aspect of the above-mentioned embodiment, an output feedback signal is provided to dynamically adjust the composition of the at least one chemical precursor in response to the indirectly determined active ingredient state. The intervention event may be predicted, for example, based on a non-threshold violation (non-threshold violation) with respect to the indirectly determined active ingredient state.

It will be appreciated that each of the above-mentioned aspects may be provided separately or in other combinations with respect to the above-mentioned embodiments.

In another embodiment, disclosed herein is a system for dosing optimization in a chemical feed unit that receives at least one input water source and produces at least one output solution. The system may include a plurality of online sensors, one or more communication devices functionally linked to the plurality of online sensors and configured to generate a message to a remote server via a communication network, wherein the generated message contains data corresponding to directly measured variables for each of the respective components, the remote server contains or is functionally linked to a data store and may additionally be further configured to direct execution of steps according to any one or more of the above-mentioned method embodiments and the above-mentioned aspects, the data store further containing a model that associates combinations of input variables with respective process states in at least one of the chemical feed unit, the output solution, and the at least one input water source.

Many objects, features, and advantages of the embodiments set forth herein will be apparent to those of skill in the art upon reading the following disclosure in conjunction with the accompanying drawings.

Drawings

FIG. 1 is a block diagram representing an exemplary embodiment of a system as disclosed herein.

FIG. 2 is a flow chart representing an exemplary embodiment of a method as disclosed herein.

Fig. 3 is a flow chart representing another exemplary embodiment of a method as disclosed herein.

FIG. 4 is a flow chart representing an exemplary embodiment of a method as disclosed herein.

Fig. 5 is a flow chart representing another exemplary embodiment of a method as disclosed herein.

Best mode for carrying out the invention

Referring generally to fig. 1-5, various exemplary embodiments of the present invention will now be described in detail. Where various figures may describe embodiments sharing various common elements and features with other embodiments, similar elements and features are given the same reference numerals, and redundant descriptions thereof may be omitted below.

Briefly, the systems and methods as disclosed herein may be implemented to predictively alert a user via data analysis or to implement automatic dosing optimization in a chemical process.

In one particular embodiment, described in more detail below, a system and method may be provided for determining whether acid boiling is required to remove mineral scale and/or biofouling of a chemical feed slide, where the chemical feed slide is a chemical feed slide that generates an oxidizing biocide solution from a plurality of precursors at an alkaline pH (such as monochloramine in the specific examples mentioned throughout this disclosure, but without limiting the scope of the invention). Algorithms to determine if boiling is required are established using data directly from the feed chute and including, but not limited to, water conductivity, temperature, flow rate, pH, run time, etc.

In another embodiment as described in more detail below, a system and method (which may be independent, or an additional part of the same system, and further supplementing the same method as previously discussed) may be provided to predictably model the true stoichiometric ratio of two chemical precursors for the real-time generation of an oxidizing biocide, wherein at least one chemical precursor having a concentration of an active ingredient that varies over time is indirectly monitored and/or remotely determined such that, for example, volumetric flow adjustments may be made to optimize the efficiency and performance of the oxidizing biocide.

In another embodiment as described in more detail below, the system and method (which may also be stand alone, or be an additional part of the same system, and further supplement the same methods as previously discussed) may involve controlling the amount of oxidizing biocide fed to a commercial or industrial process to regulate the amount of microbial contamination within the process. Such methods may include capturing both on-line and off-line operation of the process and quality data to develop and deploy application-specific control logic such that microbial contamination is minimized while optimizing key process performance metrics and oxidizing biocide dosing efficiency. Both the flow data and manually entered data (streaming and manually entered data) may be sent, for example, to a remote server where application specific algorithms are developed and pushed back down the edge device to adjust biocide feed along one or more points of the process.

Referring first to fig. 1, an embodiment of a host system 100 as disclosed herein may be provided in association with, or even in some cases including, various stages in an industrial plant, including an input stage 110 that provides one or more content streams to a chemical feed stage 120 that further provides an output solution, such as monochloramine (hereinafter "MCA"). In one embodiment, the input stage may include a first precursor comprising a bleaching solution and a second precursor comprising an amine solution, each of which is fed to a defined area (defined area) to form a mixture (e.g., a reaction mixture) from which MCA products are produced. The MCA product may be used, for example, to treat an aqueous final solution, such as water, pulp, aqueous streams, and the like, and in certain alternative embodiments, the supplied oxidizing agent and amine reactants used to prepare the MCA product may be combined directly in the final solution to produce the treated product in situ, or the reactants may be combined in situ and prior to the final solution. The defined area in which the reactants are shown combined may comprise a vessel or line, such as a tank, pipe, conduit, reactor, bath, stream, or vessel, or the like. Additional supply reactants not shown in this illustration may be used depending on the reaction chemistry involved.

As used herein, the term "industrial plant" may generally connote a facility that produces goods independently or as part of a group of such facilities, and may for example, but not be limited to, those related to industrial processes and chemical business, manufacturing, catering, agricultural, swimming pool, home automation, leather processing, paper making, and the like.

The system "host" as referred to herein may generally be independent of a given industrial plant, but this aspect is not required within the scope of the present disclosure. The system host may be directly associated with an embodiment of the cloud-based server system 100 and is capable of performing predictive analysis and preventative maintenance operations as disclosed herein, either directly or indirectly, for each of a group of industrial plants.

The data collection stage 140 may, for example, include a plurality of sensors 142 positioned in-line with various respective components of the chemical feed stage 120 and/or the input stage 110 and/or the output solution 130. Some or all of the sensors 142 may preferably be configured to continuously generate signals corresponding to real-time values of conditions and/or states of the respective components. The sensors may be configured to calibrate or otherwise convert the raw measurement signals to output data in a form or protocol to be processed by the downstream computing device, or in various embodiments, one or more intermediate computing devices or controllers (not shown) may be implemented to receive the raw signals from some or all of the sensors and provide any necessary calibration or conversion to a desired output data format.

The term "sensor" may include, but is not limited to, a physical level sensor (physical level sensor), a relay, and equivalent monitoring devices that may be provided to directly measure a value or variable of an associated process component or element, or to measure an appropriately derived value (derivative value) from which the process component or element can be measured or calculated.

As used herein, the term "on-line" may generally refer to the use of devices, sensors, or corresponding elements located near a container, machine, or associated process element, and generating output signals corresponding to the desired process element in substantially real-time, as distinguished from manual or automatic sample collection and "off-line" analysis in a laboratory or visual observation by one or more operators.

Each sensor 142 may be individually mounted and configured, or the system 100 may provide a modular housing including, for example, a plurality of sensors or sensing elements 142. The sensors or sensor elements may be permanently or portably mounted in specific locations corresponding to the chemical feed stage 120, or may be dynamically adjustable in location to collect data from multiple locations (e.g., further including the input stage 110 and/or the output solution 130 from the chemical feed stage) during operation.

The in-line sensor 142 as disclosed herein may provide substantially continuous measurements with respect to various process components and elements, and is substantially real-time. As used herein, the terms "continuous" and "real-time" with respect to at least the disclosed sensor outputs do not require an explicit degree of continuity, but rather may generally describe a series of measurements corresponding to the physical and technical capabilities of the sensor, the physical and technical capabilities of the transmission medium, the physical and technical capabilities of any intermediate local controller, communication device and/or interface configured to receive the sensor output signals, and the like. For example, based on the relevant hardware components or based on the communication network configuration, the measurements may be made and provided periodically and at a slower rate than the maximum possible rate, which smoothes the input values over time, and still be considered "continuous".

While sensors may be used to directly measure control parameters, such as contamination levels in particular stages or components of an industrial process, or concentration of halogenated materials in chemical precursors, such sensors may be extremely expensive or unreliable, as previously described herein. Accordingly, various embodiments of the system 100 as disclosed herein implement a sensor 142 in the data collection phase 140 that directly senses the value, level, status, etc. of variables other than the specified control parameters (e.g., contaminants) in question, and which is more reliable and readily available for implementation, wherein process conditions (e.g., pollution conditions and/or active ingredient conditions) are indirectly determined or predicted during the predictive maintenance (cloud-based calculation) phase of the system.

The data collection phase 140 may further include a Graphical User Interface (GUI) 144 in which a user, such as an operator, administrator, or the like, may provide periodic input regarding conditions or status of additional components relevant to downstream algorithms, as discussed further herein. The GUI 144 may also be in functional communication with a host server 152 and/or a local process control unit (not shown) to receive and present process-related information or to provide other forms of feedback regarding, for example, a cleaning or replenishment process, as discussed further herein. As used herein, unless otherwise indicated, the term "user interface" may include any input-output module with respect to a host data server, including, but not limited to: a fixed operation panel with key data input, a touch screen, buttons, a dashboard and the like; portal sites such as individual web pages or those that collectively define a hosted web site; mobile device applications, and the like. Thus, one instance of the user interface may be generated remotely, such as on the user computing device 120, and communicatively linked to the remote server 110.

Alternatively, it is within the scope of the present disclosure that one instance of the GUI 144 may be generated on a fixed display unit in an operation control panel (not shown) associated with a production phase of an industrial plant.

The data collection stage 140 may further include one or more communication devices 146 configured to receive the output signals from the online sensors 142 and transmit corresponding output data to a host server 152 via, for example, a communication network. The communication device may be stand alone or alternatively contain a local controller configured, for example, to direct the collection and transmission of data from the industrial plant to a cloud server and further direct output signals from the server to other process controllers at the plant level or more directly to the process executors in the form of control signals to implement automated interventions. In some implementations, the communication device or local controller may be omitted, where, for example, the data collection tool is distributed for directly transmitting the data stream via a communication network, and the user computing device that also exposes and implements GUI 144 is implemented to receive output signals from a server or the like. In some implementations, the communication device or local controller may include at least a portion of a resident control system of the industrial plant.

In one embodiment (not shown), the conversion stage may be increased in order to convert the raw signals from one or more online sensors 142 into signals compatible with the communication network and/or cloud server-based storage and application data transmission or data processing protocols. The conversion phase may involve not only input requirements, but may further provide data security between one or more sensors and the cloud-based server 152, or between the local communication device 146 (e.g., a local controller) and the server.

As used herein, the term "communication network" with respect to data communication between two or more system components or, in addition, between communication network interfaces associated with two or more system components may refer to any one, or a combination of any two or more, of the following: telecommunication networks (wired, wireless, cellular, etc.), global networks such as the internet, local networks, network links, internet Service Providers (ISPs), and intermediate communication interfaces. Any one or more conventionally identified interface standards may be used in its implementation including, but not limited to, bluetooth, RF, ethernet, and the like.

The preventative maintenance phase 150 as represented in fig. 1 may be provided with a host server 152 or a network of host servers linked to the communication device 146 as discussed above. The host server 152 may be associated with a third party of the industrial plant or alternatively may be a server associated with the industrial plant or an administrator thereof, which may further include or be linked to a data store or network 154, the data store or network 154 including models and/or algorithms related to process states and/or intervention events from the input stage 110, chemical feed stage 120, and/or solution 130 of the industrial plant. Thus, the cloud-based server 152 implements a preventative maintenance model that may be configured to process data provided from an industrial plant in view of iterative development residing in the data storage network 154, and generate feedback to corresponding components in the industrial plant regarding, for example, automatic upstream adjustments 156 at the input stage and/or automated cleaning procedures 158 at the chemical feed stage.

The system 100 mentioned above may be implemented in an embodiment of the method 200 as discussed further below with reference to fig. 2, or in an embodiment of the method 300 as discussed further below with reference to fig. 3. Unless otherwise indicated, alternative embodiments of the system may be implemented for any of the methods 200, 300 within the scope of the present disclosure. Depending on the implementation, certain acts, events, or functions of any of the algorithms described herein can be performed in a different order, may be added, combined, or omitted entirely (e.g., not all of the described acts or events are necessary for the practice of the algorithm). Further, in some implementations, actions or events may be performed concurrently, e.g., through multi-threaded processing, interrupt processing, or multiple processors or processor cores, or on other parallel architectures, rather than sequentially.

Those skilled in the art will appreciate that many of the steps in the process of generating the desired output solution 130 via the input stage 110 and the chemical feed stage 120 are conventionally known and generally depend on the type of solution being generated, and that a detailed discussion of such steps or processes may be omitted herein as they are generally outside the scope of the invention as disclosed herein.

Referring more particularly to fig. 1 and 2, for a given process, such as the generation of an MCA mixture, the method 200 includes on-line data collection (step 210) regarding a plurality of process components in one or more of the input stage 110, the chemical feed stage 120, and the output solution 130 itself.

The output from the data collection stage 140 is transmitted via a communication network to a remote (e.g., cloud-based) server network 152 (step 220).

The server 152 may further transmit the output from the data collection stage 140 of the industrial plant to a separate server and/or data storage network 154 for iterative development and updating of predictive models associated with the present disclosure (step 230). As just one illustrative example, a predictive model may be constructed to account for changes in feed, level, equipment, etc., where a "digital twinning" virtual representation in a cloud-based network continuously compares actual performance to expected performance to enable or otherwise facilitate predicting future trends and proactive intervention. Such a virtual representation may include a pairing of digital and physical data and further be combined with a learning system (e.g., an artificial neural network). Real-time data may be provided throughout the process or lifecycle of the respective assets to generate a virtual representation for estimating a given parameter or performance metric, wherein subsequent comparisons of such predicted or estimated parameter or metric with corresponding measured or determined parameters or metrics may preferably be implemented as feedback for machine learning algorithms performed at the server level.

The initial model may be constructed, for example, based on data collected and optionally aggregated from multiple chemical feed skids distributed across any number of industrial locations.

In particular exemplary embodiments involving automated cleaning (e.g., acid boiling) or upstream softener replenishment, data may be collected according to each of the following components:

on-line measurement of the pH of dilute hypochlorite may be one of the driving factors for the model and does not require human intervention from anyone in the field. Data may be collected, for example, every sixty seconds, and higher system pH typically correlates to higher fouling rates.

For example, online measurement of the pH of the MCA mixture collected every sixty seconds may also be one of the driving factors of the model and does not require human intervention from anyone in the field. A low pH indicates an excess of chlorine, which causes an unintended reaction that lowers the pH. The standard pH range may be between 10.5 and 11.5, with higher system pH typically associated with higher fouling rates.

For example, the conductivity (related to hardness) of the incoming water collected every sixty seconds may also be one of the driving factors for the model. Higher water conductivity typically correlates to higher fouling rates. Since the softener replaces calcium with sodium ions, the conductivity difference between the functional softener and the non-functional softener cannot be detected via conductivity alone.

The dilution rate of hypochlorite in water can be measured or derived every sixty seconds to adjust or correlate the pH of the dilute hypochlorite. Higher dilutions may result in higher diluted hypochlorite pH readings.

Hypochlorite/mcap volume ratios can be measured or derived every sixty seconds to adjust or correlate MCA pH values, where higher dilutions can result in higher MCA pH readings.

One or more associated pulp flow rates may be measured every sixty seconds to enable calculation of the hypochlorite dilution rate and the hypo/mcap volume ratio, as well as calculation of the total volume of each flow stream.

The cell status (e.g., dosing, flushing, idle) may be measured as each process changes, e.g., to filter out flushing/idle data.

The hypochlorite volume over time can be measured or derived every sixty seconds to determine how much hypochlorite has passed through the unit in a given amount of time. For example, a higher ratio of hypochlorite volume to pipe diameter may be associated with a faster fouling rate.

The MCA volume over time can be measured or derived every sixty seconds to determine how much MCA passes through the unit in a given amount of time. For example, a higher ratio of MCA volume to conduit diameter may be associated with a faster fouling rate.

The volume of water over time may be measured or derived every sixty seconds to determine the total flow through the system. For example, for a given pipe diameter, a higher total volume may be associated with a faster fouling rate.

The system inner tube diameter, in-line mixer size, and/or application line size may be a single measurement provided to the system, as a smaller inner diameter may require more frequent cleaning, for example, when all other variables are kept constant.

The online scale inhibitor flow may be measured or otherwise derived every sixty seconds.

Whether or not to utilize the softener may be a single measurement provided to the system. Further, measurements may be provided regarding process variations in refilling the softener or cleaning with salt.

Visual inspection may be provided daily with respect to a given unit to provide feedback as to whether pickling is required, thereby, for example, further developing or otherwise providing confirmation of model parameters.

As previously described, once a sufficient data set is established, a model can be developed that correlates combinations of input variables with predicted accumulation of contaminants in at least a portion of the chemical feed units, for example, to predict when some or all of the chemical feed units need to be pickled, or to supplement brine in a water softener. Various embodiments of a model for predicting pollution events (e.g., one or more events identified as requiring pickling) may be constructed for corresponding system implementations, such as: a system using a softening agent; a system using an online scale inhibitor; a system that uses neither softener nor scale inhibitor; a system using both a softener and an anti-fouling agent, and the like.

In various exemplary embodiments, a contamination event may be identified via threshold-based analysis of indirectly determined contamination states. Alternatively or in addition, non-threshold based analysis may be used to predict the timing of a contamination event, for example, based on indirectly determined contamination states. In the context of a pickling process, for example, for a chemical feed stage, the system may typically automatically implement such a process upon determining the presence of a contamination event, or may schedule such a process at a defined time in the future based on a predicted contamination event. In the context of brine replenishment in, for example, a water softener, the system can implement non-threshold based analysis to adjust brine replenishment based on the determined contamination status and with the aim of at least delaying a contamination event (predicted or otherwise) in the chemical feed stage.

The various models may require only automatic streaming or only manual data collection once (e.g., no "routine" manual data collection is required).

Various embodiments of these models may be deployed in the cloud to provide alerts to users to prompt them to pickle their systems or replenish their softeners. The user may then be automatically prompted to provide feedback regarding the accuracy of the model, which will preferably be used to fine tune the model. In one embodiment, when the system predicts that pickling is required, a message may be generated to a user interface associated with an operator, administrator, representative (representational), or the like to confirm or approve initiation of the automated cleaning process. Such approval may be received, for example, via user actuation of a dedicated button or other interface tool. Alternatively, and as additionally noted in this disclosure, an automated cleaning procedure may be dynamically implemented upon determining a contamination event without human intervention.

With further reference to the flow chart in FIG. 2, in view of the model residing in the data storage network 154, implementing data from the data collection stage 140 of the industrial plant, the contamination status of one or more components of the monitored client systems and processes may be indirectly predicted and/or determined (step 240).

If one or more of the predicted and/or determined contamination states corresponds to a determined contamination event (i.e., yes in response to the query represented in step 250), the method 200 continues by providing feedback to the industrial plant to trigger an automated cleaning process (step 260).

With limited or no human interaction, an exemplary automated cleaning procedure triggered via the model may be performed on the chemical feed skid, and may include some or all of the following operations. First, the method may initiate shutdown or disabling of normal unit operations (e.g., MCA production and dosing), followed by a water-only system flush to remove any precursor of MCA from the system. The system pH can be checked to ensure that all precursors and MCA are removed from the system, followed by dosing the acid into the system via a pump connected to the pickling port. Once filled with acid, the soaking system may be set/configured according to the user, wherein the dosing/soaking cycle may optionally be repeated according to the user's configuration. MCAP/water flushing may be performed to remove all acid from the system and return the pH of the system to normal levels, as well as flushing with water only and checking the pH to ensure that the system is completely cleaned and flushed. Finally, an automatic restart may be implemented to return the system to normal dosing/operating conditions, or in one embodiment, a notification may be generated to the user for approval prior to the restart.

If none of the predicted and/or determined pollution states corresponds to a determined pollution event (i.e., a "no" in response to the query represented in step 250), or alternatively after or concurrently with the automated cleaning process 260, the method 200 continues by providing feedback to the industrial plant to adjust upstream conditions associated with the potential pollution event (step 270). For example, the method may include the system determining (when in use) whether the water supplying the softener to the chemical feed stage requires additional capacity, e.g., to reduce the need for acid washing at the chemical feed stage.

In various embodiments, the event determined based on the indirectly determined process state may be an intervention prompt other than an automatic corrective action, such as a purge or system adjustment, including, for example, a prompt for service or maintenance of one or more system components, or an automatic scheduling of such service or maintenance to prevent future system failures. Examples of system components that may be monitored to determine the need for service or maintenance may include pump failure, valve failure, sensor failure, etc., which may generally complement the aforementioned automated cleaning or regulation/control.

Certain embodiments of the method 200 as disclosed herein may be fully automated without requiring or prompting human intervention via, for example, a graphical user interface. The method may additionally optionally implement one or more intermediate steps in which an operator or other authorized person may approve or modify the automated cleaning procedure and/or control adjustments. For example, the host server and/or local controller may be configured to determine the amount and direction of the recommended amount of brine replenishment or other adjustments to control the valve position in the input phase and further generate notifications thereof to a designated user interface (e.g., an operator dashboard, a mobile application on a phone, etc.). Thus, the authorized personnel may be prompted to manually formulate the suggested intervention, or provide feedback via, for example, approval or editing of recommended adjustments, wherein the server/controller resumes automatic control of one or more relevant system components based thereon.

Referring now to fig. 1 and 3, another embodiment of the method 300 may be described with respect to the same process, such as the generation of an MCA mixture, which still includes on-line data collection with respect to a plurality of process components in one or more of the input stage 110, the chemical feed stage 120, and the output solution 130 itself (step 310). The output from the data collection stage 140 is transmitted via a communication network to a remote (e.g., cloud-based) server network 152 (step 320). The server 152 may further transmit the output from the data collection stage 140 of the industrial plant to a separate server and/or data storage network 154 for iterative development and updating of predictive models associated with the present disclosure (step 330). As just one illustrative example, a predictive model may be constructed to account for changes in feed, grade, equipment, etc., where a "digital twin" virtual representation in a cloud-based network continuously compares actual performance to expected performance to enable or otherwise facilitate predicting future trends and proactive interventions. Such a virtual representation may include a pairing of digital and physical data and further be combined with a learning system (e.g., an artificial neural network). Real-time data may be provided throughout the process or lifecycle of the respective asset to generate a virtual representation for estimating a given parameter or performance metric, wherein subsequent comparisons of such predicted or estimated parameter or metric with corresponding measured or determined parameters or metrics may preferably be implemented as feedback for machine learning algorithms performed at the server level.

The initial model may be constructed, for example, based on manual/batch data and measurable flow data that is reliably collected and optionally aggregated from multiple chemical feed slips distributed across any number of industrial locations.

In a specific exemplary embodiment involving modeling of the true stoichiometric ratio of two chemical precursors for real-time generation of an oxidizing biocide, where at least one chemical precursor has a concentration of active ingredient that varies over time, data may be collected according to the same several components as the embodiment discussed above with respect to fig. 2. For example, online measurement of the pH of a diluted precursor solution (e.g., hypochlorite) and online measurement of the pH of an oxidizing biocide solution (e.g., monochloramine mixture) may be driving factors for modeling molar ratios. Other measurements that are also included in the embodiments discussed above may include the dilution rate of the precursor (e.g., hypochlorite) in water, the hypochlorite/mcap volume ratio, one or more associated pulp flows (e.g., water and chemical precursors), unit status (e.g., dosing, flushing, idle), online antiscalant flow (e.g., to determine its concentration in the final solution and its effects), and whether or not the softener is being utilized, at substantially the same data collection rate and for substantially the same reasons.

Additional measurements related to the embodiment represented in fig. 3 may include the following components:

for example, the on-line bleach concentration collected every sixty seconds (if available) may optionally be obtained using a hypochlorite sensor to verify model accuracy.

Manual measurements of hypochlorite concentration can be made daily to build and train the model, but are typically not used during actual operation of the process. A large difference in hypochlorite concentration between the new/incoming hypochlorite and the old/remaining hypochlorite will drive the decomposition more rapidly.

Manual measurement of hypochlorite alkalinity may be performed each time a new hypochlorite is delivered, or at any time a change is suspected, as well as modeling and training, but typically not during actual operation of the process. In many cases, the alkalinity may be globally constant for all hypochlorites, and any change in this may be tracked as potentially affecting pH readings as all other variables remain constant.

Bulk hypochlorite temperature and/or ambient temperature may optionally be collected, for example, every sixty seconds, as temperature is one of the driving factors for hypochlorite degradation.

The feed water conductivity may be measured, for example, every sixty seconds to adjust or correlate to changes in feed water conductivity and/or dissolved solids.

The temperature of the dilute hypochlorite may be measured, for example, every sixty seconds to determine the baseline temperature prior to the reaction.

The temperature of the MCA mixture may be measured, for example, every sixty seconds to determine exothermic changes based on chemical reaction activity.

A one-time data input regarding the type of scale inhibitor may be provided to determine which particular chemical composition to use and whether to utilize the softener.

The bleach manufacturer may optionally be associated with data as a one-time input, unless the vendor changes, of course, for example, to determine and attribute differences between hypochlorite manufacturers.

Once a sufficient data set is established, an initial model can be developed that correlates the combination of input variables with the predicted true stoichiometric ratio of the active ingredient in the precursor or precursors in question. It may be desirable that the model relies primarily on the flow data, but manual data may also be augmented over time to improve model accuracy.

The developed model advantageously enables real-time prediction and/or estimation of the true stoichiometric ratio of active ingredients in chemical precursors and takes into account the upstream, downstream and environmental conditions of the oxidizing biocide generating apparatus. There are a number of exemplary results and advantages of such methods, including improved accuracy and reliability, as well as a wider range of applicability of the model, to include situations in which one or more conditions (not monitored or included in conventional systems and methods) have an impact on the model and/or measured controlled parameters. The model may further facilitate a reduction in waste consumption of one or more precursors, resulting in improved efficiency and reduced environmental impact, as well as obvious time and money savings for manual testing of precursor concentrations.

When the modeled stoichiometry is determined to be outside of optimal conditions, the system can be configured to automatically adjust the precursor volume ratio to optimize the stoichiometry of the active ingredient in the one or more precursors. Alternatively, a non-threshold determination may be made when the predictive modeling ratio will need correction. In various embodiments, these models may also be deployed remotely to provide an alert to the user to prompt them to manually adjust the volume ratio of the active ingredients in the two or more precursors. The user may be automatically prompted to provide feedback regarding the accuracy of the model, which will preferably be used to fine tune the model. In one embodiment, when the system predicts that an adjustment of the precursor volume ratio is required to optimize the stoichiometric ratio of the active ingredient in the one or more precursors, a message may be generated to a user interface associated with an operator, administrator, agent, etc. to confirm or approve the initiation of the automatic adjustment. Such approval may be received, for example, via user actuation of a dedicated button or other interface tool.

With further reference to the flow chart in fig. 3, further in view of the models residing in the data storage network 154, implementing data from the data collection stage 140 of the industrial plant, the active ingredient status of one or more precursors of the monitored client systems and processes may be indirectly predicted and/or determined (step 340).

If one or more of the predicted and/or determined active ingredient states corresponds to a determined intervention event (i.e., a "yes" response to the query represented in step 350), the method 300 continues by providing feedback to the industrial plant to adjust an upstream condition related to the composition of the at least one chemical precursor (step 370). For example, the feed rate of the amine solution may be controlled using an associated valve or pump, or the controller may be configured to adjust the feed rate of either or both of the oxidant solution and the amine solution based on a predicted and/or determined measurement of the active oxidant and further in view of the desired molar ratio, according to the specified requirements of the monochloramine production process. The process control operation may be proportional in nature, with the controller identifying a desired corrected orientation aspect in order to obtain (or drive system orientation) an optimal molar ratio, and the process control operation may further include an integral and/or derivative aspect in some embodiments, with the correction step taking into account the rate of change over time to substantially prevent overshoot.

If one or more of the predicted and/or determined active ingredient states have not yet corresponded to a determined intervention event (i.e., "no" in response to the query represented in step 350), then method 300 simply continues with online data collection and repeats the foregoing steps.

Referring now to fig. 1 and 4, another embodiment of the method 400 may be described with respect to substantially the same process, such as the generation of an oxidizing biocide solution, such as a MCA mixture, that still includes on-line data collection with respect to a plurality of process components in one or more of the input stage 110, the chemical feed stage 120, and the output solution 130 itself (step 410). The output from the data collection stage 140 is transmitted via a communication network to a remote (e.g., cloud-based) server network 152 (step 420). The server 152 may further transmit the output from the data collection stage 140 of the industrial plant to a separate server and/or data storage network 154 for iterative development and updating of predictive models associated with the present disclosure (step 430). As just one illustrative example, a predictive model may be constructed to account for changes in feed, grade, equipment, etc., where a "digital twin" virtual representation in a cloud-based network continuously compares actual performance to expected performance to enable or otherwise facilitate predicting future trends and proactive intervention. Such a virtual representation may include a pairing of digital and physical data and further be combined with a learning system (e.g., an artificial neural network). Real-time data may be provided throughout the process or lifecycle of the respective assets to generate a virtual representation for estimating a given parameter or performance metric, wherein subsequent comparisons of such predicted or estimated parameter or metric with corresponding measured or determined parameters or metrics may preferably be implemented as feedback for machine learning algorithms performed at the server level.

The initial model may be constructed, for example, based on manual/batch data and measurable flow data that is reliably collected and optionally aggregated from a plurality of process locations, such as chemical feed slips, distributed across any number of industrial locations. Once a sufficient data set is established, an initial model associated with the combination of input variables may be developed to determine or predict in real time the amount of oxidizing biocide contained within the application or process, the amount of microbial contamination contained within the application or process, the quality of the resulting end product and/or key performance metrics of the customer process, which may result in improved operability and performance by optimizing the determined or predicted data.

Accordingly, and with further reference to the flow chart in FIG. 4, further in view of the model residing in the data storage network 154, implementing data from the data collection stage 140 of the industrial plant, the final product quality and/or key performance metrics associated with the industrial process may be indirectly predicted and/or determined (step 440).

If the one or more predicted and/or determined final product quality and/or key performance metrics correspond to the determined intervention event (i.e., "yes" in response to the query represented in step 450), the method 400 continues by providing feedback to the industrial plant to adjust the feed rate of the oxidizing biocide in at least one point of the process (step 470). The process control operation may be proportional in nature, with the controller identifying a desired corrected directional aspect in order to obtain (or drive system orientation) an optimal feed rate, and the process control operation may further include an integral and/or derivative aspect in some embodiments, with the correction step taking into account the rate of change over time to substantially prevent overshoot.

If one or more of the predicted and/or determined final product quality and/or key performance metrics have not yet corresponded to the determined intervention event (i.e., in response to the query represented in step 450, "NO"), method 400 simply continues with online data collection and repeats the foregoing steps.

The above-mentioned embodiment 400 may preferably include a model and related control scheme that improves over time to optimize the dosage rate of the biocide for a given commercial or industrial process. It will be appreciated by those skilled in the art that the prevention of an overfeed of an oxidizing biocide may result in a reduction in any one or more of the following: corrosion problems and damage to terminal processing equipment; processing program cost; the burden of the wastewater treatment system; impact on emissions restrictions/permissions; etc. Those skilled in the art will further appreciate that the prevention of under-feeding of oxidizing biocides may result in a reduction in any one or more of the following: microbial outbreaks in the final process, which may lead to negative effects on process operability or final product quality; transmission of airborne diseases caused by the growth of microorganisms not examined in commercial and/or industrial processes; etc. For example, overfeeding of biocides can lead to detrimental exposure-based effects, such as salivation or other health problems, due to vapors released from the process.

Referring next to fig. 5, the above-mentioned embodiments 200, 300, 400 for a given commercial and/or industrial process are further illustrated in executable combinations as embodiments of the method 500 disclosed herein. It will be appreciated that in alternative embodiments, any two of the disclosed embodiments may be combined, or that the steps associated with the respective embodiments may be performed in a different order than represented in fig. 5, which is intended to be illustrative only.

While embodiments of the invention as disclosed herein may be described for illustrative purposes in the context of certain commercial applications (e.g., drawings, tissues, packaging) for pulping and papermaking production, it will be understood by those skilled in the art that the systems and methods as disclosed herein are predictably provided for other commercial applications including, but not limited to, water treatment applications (e.g., cooling systems, heating systems, potable water systems, influent systems) and biomass applications (e.g., sugar ethanol, corn ethanol, beet sugar).

Throughout the specification and claims, the following terms take at least the meanings explicitly associated herein, unless the context dictates otherwise. The meaning of the following identification does not necessarily limit the term, but merely provides an illustrative example of the term. The meaning of "a", "an", and "the" may include plural referents, and the meaning of "in …" may include "in …" and "on …". As used herein, the phrase "in one embodiment" does not necessarily refer to the same embodiment, although it may. As used herein, the phrase "one or more" when used with a list of items means that different combinations of one or more items may be used and that only one of each item in the list may be required. For example, "one or more" of items a, B, and C may include, for example, but are not limited to, item a, or item a and item B. The instance may also include item a, item B, and item C, or item B and item C.

The term "coupled" means a direct physical or electrical connection between at least the connected items, or an indirect connection through one or more passive or active intermediary devices.

The various illustrative logical blocks, modules, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. The described functionality may be implemented in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.

The various illustrative logical blocks and modules described in connection with the embodiments disclosed herein may be implemented or performed with a machine, such as a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor may be a microprocessor, but in the alternative, the processor may be a controller, a microcontroller, or a state machine, combinations thereof, or the like. A processor may also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.

The steps of a method, process, or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of computer-readable medium known in the art. An exemplary computer readable medium may be coupled to the processor such the processor can read information from, and write information to, the memory/storage medium. In the alternative, the medium may be integral to the processor. The processor and the medium may reside in an ASIC. The ASIC may reside in a user terminal. In the alternative, the processor and the medium may reside as discrete components in a user terminal.

Conditional language used herein such as, inter alia, "may," "might," "for example," etc., is generally intended to convey that certain embodiments include while other embodiments do not include certain features, elements, and/or states unless specifically stated otherwise or otherwise understood within the context of the use. Thus, such conditional language is not generally intended to imply that one or more embodiments require features, elements and/or states in any way or that one or more embodiments must include logic for deciding, with or without author input or prompting, whether these features, elements and/or states are included in or are to be performed in any particular embodiment.

The foregoing detailed description has been provided for purposes of illustration and description. Thus, while there have been described particular embodiments of the invention, which are new and useful, it is not intended that such references be construed as limitations upon the scope of this invention except as set forth in the following claims.

Claims (11)

1. A method of dosing optimization for a chemical feed unit that receives at least one input water source and produces at least one output solution, the method comprising:

generating signals from a plurality of in-line sensors, the signals corresponding to variables measured directly for respective process components;

selectively retrieving information from a model that associates a combination of input variables with respective process states in at least one of the chemical feed unit, the output solution, and the at least one input water source;

indirectly determining a process state of at least one of the chemical feed unit, the output solution, and the at least one input water source based on data corresponding to variables measured directly for the respective system component; and

based on the indirectly determined process state, an output feedback signal corresponding to the detected intervention event is automatically generated.

2. The method of claim 1, wherein the output solution from the chemical feed unit comprises an oxidizing biocide.

3. The method according to claim 2, wherein:

the indirectly determined process state comprises a final product quality and/or performance measure corresponding to the amount of the oxidizing biocide present; and

the output feedback signal is generated to adjust at least one dosage rate of the oxidizing biocide.

4. The method of one of claims 1 to 3, wherein the variables directly measured by the on-line sensor for the respective chemical feed unit components include measured variables corresponding to one or more of:

the pH of the dilute hypochlorite;

the pH of the output solution;

an oxidation-reduction potential; and

conductivity of the incoming water.

5. The method according to claim 1, wherein:

at least one of the models correlates a combination of input variables with a predicted accumulation of contaminants in at least a portion of the chemical feed unit;

the indirectly determined process state comprises an indirectly determined contamination state of at least a portion of the chemical feed unit; and

The output feedback signal corresponds to a detected contamination event based on the indirectly determined contamination state.

6. The method of claim 5, wherein the contamination status of at least a portion of the chemical feed units is indirectly determined further based on derivative variables of the respective chemical feed unit components, the derivative variables based on one or more of the variables directly measured by the online sensor, the derivative variables corresponding to one or more of:

hypochlorite dilution rate;

a hypochlorite volume passing through the chemical feed unit over time; and

water volume passing through the chemical feed unit over time.

7. The method according to claim 1, wherein:

at least one of the models correlates a combination of input variables with a predicted true ratio between two or more chemical precursors used to generate the output solution;

the indirectly determined process state comprises an indirectly determined active ingredient state for at least one of the two or more chemical precursors; and

the output feedback signal corresponds to a detected intervention event based on the indirectly determined active ingredient state.

8. The method of claim 7, wherein the active ingredient status of at least one of the two or more chemical precursors is determined indirectly from derivative variables further based on one or more of the variables directly measured by the online sensor.

9. The method according to claim 7 or 8, wherein:

providing the output feedback signal to dynamically adjust the composition of at least one chemical precursor in response to the indirectly determined active ingredient state.

10. The method according to one of claims 7 to 9, wherein:

the intervention event is predicted based on a non-threshold violation of the indirectly determined active ingredient state.

11. A system for dosing optimization in a chemical feed unit that receives at least one input water source and produces at least one output solution, the system comprising:

a plurality of in-line sensors, each of the in-line sensors configured to generate a signal corresponding to a variable measured directly for a respective chemical feed unit assembly;

one or more communication devices functionally linked to the plurality of online sensors and configured to generate a message to one or more remote servers via a communication network, wherein the generated message includes data corresponding to variables measured directly for each of the respective components;

One or more remote servers including or functionally linked to a data store, the data store further including a model that associates combinations of input variables with respective process states in at least one of the chemical feed unit, the output solution, and the at least one input water source;

the one or more servers are further configured to automatically direct the execution of any remaining steps in the method according to one of claims 1 to 10.