A method of monitoring a chemical plant

Abstract

Monitoring a chemical plant (100, fig 1) which is implementing a chemical process uses 310 sensors to obtain a sensor data of the chemical process, storing 320 the sensor data on a remote server, anal

Description

The present invention relates to the monitoring of a chemical plant by remote computational means. Particularly, but not exclusively, the methods that follow have particular application to processes with unpredictable inputs in the form of feedstock flow rates and compositions or sources of power. Moreover, the methods that follow have particular application to processes that use catalysts, the performance of which can vary overtime based on both predictable and unpredictable factors. A preferred example in which the disclosed methods may be employed is the process of producing methanol and in particular e-methanol. The process of producing methanol conventionally includes the generation of syngas (synthesis gas) comprising hydrogen and carbon oxides from fossil fuels by partial oxidation or steam reforming, followed by adjustment of the syngas stoichiometry and reaction over a methanol synthesis catalyst to produce a crude methanol with water as a by-product. The synthesis gas generation, stoichiometry adjustment and catalytic conversion are subject to change overtime, especially as the catalysts used in the process age overtime. The beginning of life (BOL) performance of a catalyst can be significantly different to the end of life (EOL) performance and adjustments of the process feeds and operating conditions are generally required to maintain optimal production between catalyst changeout and to maximise the useful lifetime of the catalysts. The process of producing e-methanol includes reacting hydrogen produced by electrolysis of water with carbon dioxide obtained from renewable sources or captured from air. The process is typically powered, at least in part, by a renewable source of energy (e.g., wind power). As such, the supply of feedstock and electricity is significantly variable and unpredictable, since it depends on the composition of the feedstock and availability of power. Although this problem is present in numerous types of chemical plants, it is particularly exacerbated in e-methanol production, which involves many non-linear processes because of, for example, the presence of recycle loops. Owing to this complexity of the chemical process, in scenarios such as this, the processing of the feedstock in a chemical plant provides a more difficult monitoring problem than previously considered. In particular, the variation in the process makes the performance of a catalyst, which typically cannot be directly measured using a sensor, harder to monitor. It is thus very difficult to optimise a chemical plant implementing such a chemical process based on the state of a catalyst, or to predict when such catalysts should be replaced. There is therefore a need for a better approach to the problem of monitoring chemical processes by computational means. According to the invention, there is provided a method of monitoring a chemical plant, comprising: operating a first chemical plant to implement a chemical process; repeatedly using sensors of the first chemical plant to obtain a set of parameters of the chemical process to produce first sensor data; 07 05 25 storing the obtained first sensor data on a remote server; analysing the first sensor data stored on the remote server using a remote controller; and communicating the result of the analysis to a first local controller, wherein analysing the first sensor data stored on the remote server using the remote controller comprises using the first sensor data with a model of at least one component of the first 5 chemical plant to calculate a first parameter of the chemical process and further comprises: using the model to generate a first trajectory of the first parameter over a first time period; using the model to generate a second trajectory of the first parameter over a second time period; and identify a difference between the calculated first and second trajectories. 10 By analysing the first sensor data using a remote controller, the model can be more accurate and up-to-date. For example, the model may be based on data stored on a remote server that is obtained from multiple chemical plants, and possibly be updated periodically or intermittently. The method is applicable to any chemical process, but is preferably used for any heterogeneously-15 catalysed processes using electrolytic hydrogen or electrolytic oxygen, i.e. hydrogen or oxygen produced by electrolysis, such as: methanol synthesis; ammonia synthesis; synthesis of formaldehyde; autothermal reforming; partial oxidation; methanation of carbon dioxide to form methane; Fischer-Tropsch reactions; water gas shift processes; or reverse water-gas shift processes. 20 For a better understanding of the invention, and to show how the same may be put into effect, reference will now be made, by way of example only, to the accompanying drawings in which: Figure 1 shows a schematic representation of a chemical plant for producing e-methanol; Figure 2A shows a schematic representation of a monitoring system for the chemical plant of Figure 25 1; Figure 2B shows a schematic representation of a monitoring system having multiple chemical plants; Figure 3 shows a flow chart of a method carried out by the system of Figure 2; Figure 4A shows a flow chart depicting a first embodiment of a method of analysing data stored on the remote server using a remote controller; 30 Figure 4B shows a flow chart depicting a second embodiment of a method of analysing data stored on the remote server using a remote controller; Figure 4C shows a flow chart depicting a third embodiment of a method of analysing data stored on the remote server using a remote controller; and Figure 5 shows a flow chart of a method carried out by the system of Figure 2. 35 Figure 1 shows a chemical plant 100. An exemplary chemical plant 100 comprises: chemical processing apparatus 101; a source of power 105; a source of feedstock 110; a catalyst 115; one or more sensors 120; one or more controllable components 125; and a local controller 130. 40 The chemical processing apparatus 101 is configured for processing the feedstock provided by the source of feedstock 110 to produce a product 135. Preferably, the chemical process is the production of methanol using electrolytic hydrogen. The precise apparatus shown in the figure is merely exemplary. The source of power 105 may be a source of electricity. In particular embodiments, the source of power 105 is a renewable source. For example, the source of power 105 may include at least one of: a solar panel; a wind turbine; and / or a hydroelectric generator. The power may be used for some or all of the components of the chemical plant 100. The source of feedstock 110 may be a supply of raw materials, or may be apparatus that processes raw materials to provide a refined feedstock 110. In the example shown in Figure 1, the source of feedstock 110 may be an electrolysis system, powered by the source of power 105 (in this case renewable). Carbon dioxide (CO2) may be passed to the feedstock 110 from any source, such as direct air capture, CO2 recovery from combustion gases, or from sequestered sources of CO2. For example, the electrolysis system may use electricity supplied by a renewable source 105 to produce hydrogen as a feedstock 110 for the chemical plant 100. It is also possible to consider the electrolysis system 110 to be a component of the chemical processing apparatus 101, in which case, the source of fluid for electrolysis would be considered the source of feedstock 110. In either case, the intermittent nature of the power from the power source 105 can produce a significant, and unpredictable, variation in the production of hydrogen by the electrolysis system 110. The catalyst 115 can be of any type, but for a methanol plant may be a component of, or a coating of, pellets forming at least part of a reactor bed. The catalyst 115 may be of a type that fouls or degrades over time. The performance of a catalyst 115 can vary based on temperature, pressure and the presence of contaminants in the feedstock. It is typically not possible to sense the state of the catalyst directly. The degradation of the catalyst is a function of the conditions in the chemical plant 100, which are a function of the feedstock 110. So, if the conditions are not revised in response to changing feedstock then the catalyst 115 degradation can be unnecessarily accelerated. For example, the catalyst 115 can be sensitive to heat and steam, both of which are produced in the synthesis reaction, so the conditions influence the rate of degradation. Accordingly, the state of the catalyst 115 must be inferred from the set of parameters obtained from the sensor data. The sensors 120 (there may be more than those depicted) may be conventional sensors for sensing parameters of the chemical process. The parameters of the chemical process may include one or more of: the parameters of the feedstock, the chemical plant, and / or the product produced. For example, the parameters of the chemical process may include a state of a catalyst used in the chemical process and forming part of the chemical plant 100. The sensors 120 may be located in, mounted on, or integrated within components of the chemical processing apparatus 101. The sensors 120 communicate the sensor data to the local controller 130. Collectively, the sensors 120 provide a set of parameters representing a state of the chemical plant 100 and chemical process. In what follows, the phrase “set of parameters” refers to a plurality of values representing the quantity measured by each sensor. For example, these may be presented as a vector. For example, a set of parameters may be a measurement of each of temperature, input pressure, output pressure, and flow rate. The sensors 120 can simultaneously provide sensor data or provide sensor data at different times. However, the local controller 130 can form a set of parameters from the sensor data representing a common time. This may be done, for example, by grouping sensed data from similar times, or by interpolating the time series of sensed values obtained from each sensor 120 to obtain a set of parameters for a particular time. Irrespective of the particular method, the local controller 130 is arranged to use the sensors 120 to repeatedly obtain a set of parameters of the chemical process. The sensors 120 may include one or more of: a temperature sensor; a flow rate sensor; a pH sensor; a pressure sensor; a concentration sensor; a power sensor; a differential pressure sensor; a speed sensor; a direction sensor; a level sensor; a displacement sensor; a valve sensor for sensing the opening state of a valve; etc. The controllable components 125 (there may be more than those depicted) may include one or more of: an actuator; a valve; a pump; a heater; a fan; a compressor; a turbine; a burner; a cooler; and / or an agitator. The controllable components 125 may be controlled by the local controller 130 to influence the chemical process. The local controller 130 communicates with the sensors 120 and controllable components 125 to monitor and influence the chemical process. For example, the local controller 130 may include a computer. The local controller 130 may include storage, such as a local historian 136, and / or store data on a remote server 140. The local controller 130 may include or implement one or more control systems arranged to maintain certain parts of the chemical process at set points. In this context a set point is a desired parameter of the chemical process (whether directly measurable by the sensors 120 or not). For example, one set point may define a particular flow rate along a conduit of the chemical processing apparatus 101, while another set point may define a particular temperature within a chamber or vessel of the chemical processing apparatus 101. The controllable components 125 may be more or less direct in their influence on the set points. For example, a heater jacket may be a controllable component 125 that directly heats a vessel with the temperature of the vessel being desired to equal a set point. Alternatively, the influence may be indirect, such as a downstream flow rate of the product being the result of a chemical reaction in the vessel, influenced by that heater jacket, with the flow rate being desired to equal a set point. Moreover, the set points might not be directly measurable, but might be inferred. The local controller 130 is arranged to control the controllable components 125 to influence the chemical process. For example, one controllable component 125 may be a heater and the local controller 130 may increase the heat output of the heater. As another example, one controllable component 125 may be a continuously variable valve and the local controller 130 may open or close the valve to modify a flow rate. Figure 2A shows a schematic representation of a monitoring system 102 for the chemical plant 100 of Figure 1. The monitoring system 102 comprises: the local controller 130; a remote server 140; a remote monitor 150; and one or more display devices 160, 165. The local controller 130 is as described above with respect to Figure 1. The local controller 130 may comprise a local processor 132. The local controller 130 may comprise a local historian 136. The local historian 136 may be arranged to store sensor data from the chemical plant 100. The local historian 136 may be on a local server. In some embodiments, the local historian 136 may for example be cloud-based, but it will be a ‘local historian’ in the sense that it is configured to store data for the chemical plant 100. Data obtained from the sensors 120 may be sent to the remote server 140, either directly, or (more preferably) from the local historian 136. The remote server 140 may be a cloud server. The remote server 140 may be in communication with one or more further chemical plants, and may be configured in the same way as shown with respect to the chemical plant 100 of Figure 1. Preferably, the chemical plants 100 carry out a common process, such as producing e-methanol, and / or comprise a common component (such as catalyst 115), such that the process and / or component can be modelled using a common model using the remote monitor 150. The remote monitor 150 may comprise or access the remote server 140. The remote monitor 150 uses the method set out below in relation to Figure 3 to analyse the sensor data stored on the remote server 140. The remote monitor 150 communicates a result of the analysis to the local controller 130. The one or more display devices 160, 165 may include a remote display device 160 or a local display device 165, which may form part of the local controller 130, local processor 132, and / or local historian 135. The local display device 165 may communicate directly with the local processor 132 and / or local historian 135. The remote display device 160 may communicate directly with the remote server 140. The one or more display devices 160, 165 may display a dashboard, visually indicating for a user the parameters of the chemical process. Figure 2B shows a schematic representation of a monitoring system 102 having multiple chemical plants. As depicted, each local controller 130X, 130Y, 130Z for respective chemical plants (which may be as shown in Figure 1) may provide sensor data to the same remote server 140. That sensor data may be used to tune or update models of the chemical plant 100. Figure 3 shows a flow chart of a method 300 carried out by the system of Figure 2. The method comprises: using sensors 120 to produce first sensor data in step 310; storing the first sensor data on a remote server 140 in step 320; analysing the first sensor data using a remote controller 150 in step 330; and communicating the result to the first local controller 130 in step 340. Step 310 involves repeatedly using sensors 120 of the first chemical plant 100 to obtain a set of parameters of a chemical process to produce first sensor data. The sensors 120 may capture the data regularly (e.g., at a fixed frequency) or be event-driven. Preferably, the timing of the use of the sensors 120 is controlled by the local controller 130. Not all sensors 120 need to be used simultaneously, and some may be used more frequently than others. The first sensor data may be used by the controller 130 for the control of the first chemical plant 100. The step 320 of storing the obtained first sensor data on a remote server 140 may involve firstly storing data from the sensors 120 on a local historian 136, and then communicating the data to a remote server 140. In such cases, the sensors 120 may preferably be used at a fixed frequency (whether for each sensor 120 or for the set of sensors 120). The communication from local historian 136 to remote server 140 can occur in batches at a different frequency from the use of the sensors 120. That is, the communication of batches of sensor data from local historian 136 to remote server 140 may occur at a batch transmission frequency. This frequency may be controllable, as explained below. Alternatively, step 320 may involve directly communicating the data to the remote server 140 (e.g., without a local historian 136 being present). Step 330 involves analysing the first sensor data stored on the remote server 140 using a remote controller 150. The remote server 140 and remote controller 150 may be collocated or physically separated. The analysis may be in accordance with the method depicted in Figure 4. Step 340 involves communicating the result of the analysis to the first local controller 130. For example, the result of the analysis may be stored in the local historian 136 for access by the first local processor 132 or the local display device 165. Figure 4A shows a flow chart depicting a first embodiment of a method 400A of analysing data stored on the remote server 140 using a remote controller 150. The method 400A may be used in combination with system 102. The data stored on the remote server 140 includes the data stored thereon in step 320. Optionally, the data stored on the remote server 140 may additionally include data obtained from further chemical plants 130X, 130Y, 130Z other than chemical plant 100. Those further chemical plants are preferably configured in the same way as the chemical plant 100. Preferably, the further chemical plants 130X, 130Y, 130Z carry out the same process as chemical plant 100 and / or comprise a common component (such as catalyst 115), such that the process and / or component can be modelled using a common model. The method 400A comprises: providing a model of at least one component of the first chemical plant 100 in step 410; applying the stored data to the model to calculate a first parameter in step 420; generating an expected trajectory of the first parameter over time in step 430; and identifying a difference between the data and the expected trajectory in step 440. Step 410 involves providing a model of at least one component of the first chemical plant 100. Preferably, a plurality of interconnected components of the chemical plant 100 are modelled. Most preferably, the at least one component comprises a catalyst 115. In some embodiments, the model may be a mechanistic model. Such a model may comprise systems of deterministic equations describing underlying physical processes, with the coefficients of those equations being the model coefficients. In other embodiments, the model may be a statistical model. Such a model may comprise a representation of the training data used to train the model, with the model coefficients defining the representation. Optionally, the model may comprise both a mechanistic model and a statistical model. The model may be arranged to simulate the chemical process implemented by one or more of the chemical plants 100, 100X, 100Y, 100Z. The model can preferably enable the estimation or prediction of parameters of the chemical process that are not measured, or that are not measurable. The model can also preferably enable the more accurate estimate of parameters that have unreliable measurement accuracies. The model can preferably enable a user to estimate parameters of the chemical process between times at which sensor data was captured or extrapolate to predict the future trajectory of parameters of the chemical process. Parameters that may be estimated by the model may include one or more of: product yield (e.g. methanol yield); energy efficiency of the chemical plant 100; catalyst performance; remaining catalyst lifespan; economic cost; feedstock use; a kinetic activity factor for the catalyst 115, normalised pressure drop across the catalyst 115, product selectivity, approach to equilibrium, remaining unused capacity of the catalyst 115, consumed poison capacity of the catalyst 115; etc. The model may include a set of model coefficients. The model coefficients may be derived from data captured during the use of a chemical plant. In other words, the model coefficients may be derived by fitting the model to the data obtained through the use of a chemical plant 100, 100X,100Y, 100Z. For example, the data may be obtained from historic experiments carried out using a chemical plant comprising sensors to produce the data. For example, the chemical plant 100, or another chemical plant or physical model thereof (e.g., a laboratory experiment orsmaller-scale process), may be used and data recorded. If data is desired for a parameter that cannot be directly sensed in the normal use of a chemical plant, this can be obtained by intervening in the process in a way that would not occur during eventual usage of the chemical plant. For example, if a parameter were required to represent clogging of pores in a catalyst, and if this cannot be easily sensed directly, the data can be obtained by destruction of the catalyst. This is not possible during normal use, but may be done in experimental use to produce a model. As is known in the art, for a mechanistic model, this may involve a process of optimising the model coefficients to fit data obtained from the use of a chemical plant to provide the lowest error metric (for example the sum-of-squared errors). For a statistical model, the training process will depend upon the model chosen. One example of a suitable statistical model is a Gaussian process, which may be derived from training data by Gaussian process regression. An example of how Gaussian process regression can be used for modelling chemical processes may be found in “A Bayesian data modelling framework for chemical processes using adaptive sequential design with Gaussian process regression" by L. Fleming et al, Applied Stochastic Models in Business and Industry, vol. 38, no. 5, pp. 787-805, 2022. As noted above, training data may be obtained from the chemical plant 100, or another chemical plant, or an experimental set-up representative of the chemical process. Using the system shown in Figure 2B, sensor data from any one or more of chemical plants 130X, 130Y, and 130Z may be stored on the remote server and used for the tuning, training, and / or updating of the models stored on the remote server 140. In step 420, the stored first sensor data is applied to the model to calculate a first parameter. The first sensor data stored on the remote server 140 is analysed using the remote controller 150 by using the first sensor data with a model of at least one component of the first chemical plant 100 to calculate a first parameter of the chemical process. The first parameters of the chemical process may include one or more of: the parameters of the feedstock, the chemical plant 100, and / or the product produced. For example, the parameters of the chemical process may include a state of a catalyst used in the chemical process and forming part of the chemical plant 100, such as the parameters listed above. Preferably, in step 420, the first sensor data is applied to the model to calculate a plurality of parameters. Applying the stored first sensor data to the model to calculate a first parameter may be part of a sequential process, with the model defining a state of the chemical process and the addition of new first sensor data enabling the state to be updated using the model. Step 330 described above may involve collecting a time-series of the first sensor data over a time period. Step 420 may use the first sensor data over the time period with the model to calculate the first parameter of the chemical process. Preferably, the time period is at least one week, preferably at least four weeks, more preferably at least eight weeks. This can enable a more accurate calculation of the parameter. In step 430, an expected trajectory of the first parameter over time is generated. When a plurality of parameters are generated, expected trajectories may be generated for one or more of the parameters. Preferably, a plurality of expected trajectories are generated, one for each of the plurality of parameters. The expected trajectory of the first parameter may be a function known to represent an expected change in the first parameter, that has been fitted to the parameter. As an example, the expected trajectory may be a linear function, which may be fit to the calculated first parameter values by minimising sum of squared error. Other functions or fitting methods may be used. The expected trajectory can enable interpolation or extrapolation. For example, a straight line fit to a parameter can provide an expected value of that parameter at any selected point along the line. As an example of extrapolation, the expected trajectory may be based on an older subset of the data, and so indicate the historic change in the first parameter. That is, if a time series of six data points is available, the expected trajectory may be generated on the first five data points. This can enable comparison of the expected trajectory with the sixth data point. As an example of interpolation, the expected trajectory may be fitted to the latest available data, and so indicate the difference between the first parameter and its general trend over the recent period. That is, if a time series of six data points is available, the expected trajectory may be fitted to all six data points, with the sixth data point compared with the trajectory. The expected trajectory need not be a linear function. For example, the remote server may include a model of expected performance of a component of the chemical plant 100 as a function of time dependent upon historic usage data (sensor data representing how the component has been used). The expected trajectory may therefore represent the expected decline in the performance of the component of the chemical plant 100 due to a set of one or more expected factors which may be represented by the first sensor data. The component is preferably a catalyst and the expected factors preferably include: fouling; mechanical degradation; thermal sintering, poisoning, physical deposition. In step 440, a difference between the first sensor data and the expected trajectory is identified. The first sensor data may be used to calculate the first parameter and this compared with the expected trajectory to provide the difference. Preferably, the expected trajectory may be calculated for a first period in time by using the first parameter (derived from the first sensor data as described for step 420). The expected trajectory may be extrapolated to a second period of time. First sensor data corresponding to the second period of time may be compared with the extrapolation by using the first parameter (derived from the first sensor data as described for step 420). The comparison may provide a difference. The difference can indicate a divergence of the first parameter (derived from first sensor data for the second period of time) from the expected trajectory. In some embodiments, data representative of the identified difference may then be communicated to the local controller 130. Optionally, one or both of steps 450A and 450B may be provided. In step 450A, the difference is displayed on one or both of the local and remote displays 160, 165. In step 450B, the difference is used to generate an alert. For example, the alert may be generated by comparing the magnitude of the identified difference with a threshold. The alert may be generated when the magnitude exceeds the threshold. Alternatively, in order to avoid false alarms due to noisy data, in some embodiments, the alert may be generated when the magnitude exceeds the threshold for a period of time exceeding a time threshold. The determination of whether an alert is to be generated may be carried out by the remote controller 150 or the local controller 130. Figure 4B shows a flow chart depicting a second embodiment of a method 400B of analysing data stored on the remote server 140 using a remote controller 150. The method 400B may be used in combination with system 102. The method 400B comprises: steps 410 and 420 from method 400A; generating a first trajectory of the first parameter over a first time period in step 434; generating a second trajectory of the first parameter over a second time period in step 438; and identifying a difference between the first and second trajectories in step 442. Steps 450A and 450B may then be implemented based on the difference determined in step 442. Whereas in the method 400A of Figure 4A, an expected trajectory of the first parameter was compared with a single instance of the first parameter to establish whether it was closely following the expected trajectory, in the method 400B of Figure 4B, a first trajectory of the first parameter over a first time period is compared with a second trajectory of the first parameter over a second time period, the first time period being greater than the second time period. Specifically, the first trajectory may provide an indication of a long-term trend, while the second trajectory provides the most recent short-term trend. The comparison can provide an indication of divergence of the present short-term trend from the long-term trend. This can indicate an unwanted effect to be investigated. In step 434, a first trajectory of the first parameter over a first time period is generated. The first time period may be a time period ending at the time the most recent first sensor data was received by the remote server 140. The length of the first time period may be at least one week, preferably at least four weeks, more preferably at least eight weeks. As with method 400A, when a plurality of parameters are generated, first trajectories may be generated for one or more of the parameters. Preferably, a plurality of first trajectories are generated, one for each of the plurality of parameters. The first trajectory of the first parameter may be a function known to represent an expected change in the first parameter, that has been fitted to the parameter. As an example, the first trajectory may be a linear function, which may be fitted to the calculated first parameter values by minimising sum of squared error. Other functions may be used. For example, the remote server may include a model of expected performance of a component of the chemical plant 100 as a function of time dependent upon historic usage data (sensor data representing how the component has been used). The first trajectory may therefore represent the expected decline in the performance of the component of the chemical plant 100 due to a set of one or more expected factors which may be represented by the first sensor data. The component is preferably a catalyst and the expected factors preferably include: fouling; mechanical degradation; thermal sintering, poisoning, physical deposition. In step 438, a second trajectory of the first parameter over a second time period is generated. The second time period may be a time period ending at the time the most recent first sensor data was received by the remote server 140. The second time period is different from the first time period. The length of the second time period is preferably less than the length of the first time period. For example, the second time period may be less than or equal to 20%, or 10%, of the first time period. Whereas the first and second time periods have been described as ending at a common time (the time the most recent first sensor data was received by the remote server 140), this is not essential. They may even be non-overlapping periods, For example, they may be sequential time periods. The second trajectory may or may not be represented by the same function as the first trajectory. In step 442, a difference between the first and second trajectories is identified. In one implementation, the first and second trajectories may be linear functions. The first trajectory may be a best fit linear function of the first parameter over the first time period. The second trajectory may be a best fit linear function of the first parameter over the second time period. The difference may be calculated as the difference in gradient or magnitude between the first and second trajectories. Some other metric could be used. In another implementation, the first and second trajectories may be different functions. The first trajectory may be a more complicated function of the first parameter (a higher order function than linear) fitted to the first sensor data obtained over the first period. Since the second period is preferably shorter than the first period, a simpler model may be used for the second trajectory. For example, the second trajectory may be a best fit linear function of the first parameter over the second time period. The difference may be calculated as the difference in gradient or magnitude between the first and second trajectories at a point where they overlap. Some other metric could be used. Preferably, in either of methods 400 or 400B, the local controller 130 may modify the timing of the use of the sensors 120 of the chemical plant 100 to produce the first sensor data. For example, the local controller may increase the frequency with which the sensors 120 obtain the set of parameters based on an increase in the identified difference. This can enable the system to more quickly identify when the chemical plant 100 has a problem that causes a parameter to diverge from its expected behaviour. In some implementations of method 400A or400B, the local controller 130 modifies the timing of the use of the sensors 120 of the chemical plant 100 to produce the first sensor data. For example, the local controller may increase the frequency with which the sensors 120 obtain the set of parameters based on an increase in the identified difference. Instead of (or as well as) modifying the use of the sensors, in some implementations of method 400A or400B, the local controller 130 modifies the timing of the communication of batches of sensor data from local historian 136 to remote server 140. For example, the batch transmission frequency may be increased based on an increase in the identified difference. Thus, the step of storing the first sensor data on the remote server may comprise, storing the first sensor data on a local historian; transmitting the first sensor data from the local historian to the remote server, and wherein: the first local controller controls a transmission timing of the step of transmitting the first sensor data; and the first local controller controls the transmission timing based on the result of the analysis. This can enable the system to more quickly identify when the chemical plant 100 has a problem that causes a parameter to diverge from its expected behaviour. Figure 4C shows a flow chart depicting a third embodiment of a method 400C of analysing data stored on the remote server 140 using a remote controller 150. The method 400C may be used in combination with system 102. The method 400C comprises the steps 410 and 420 of the first embodiment, and optionally (not shown) steps 430, 440, 450A and / or 450B. In addition, further steps 445 and 460 are provided. In step 445, the first parameter may be used to identify a change of the mode of operation of the chemical plant 100. The mode of operation of the chemical plant 100 may be defined by a set of control variables. The control variables are coefficients of the model. Identifying a change of the mode of operation of the chemical plant 100 in step 445 may include estimating the set of control variables for the chemical plant that optimises an objective function. For example, the remote controller 150 may use the model to simulate the chemical process and thereby establish the effect of the control variables on the objective function for a given set of parameters. The remote controller 150 may simulate the chemical process using the sensor data and the model to estimate the set of control variables that optimises the objective function. For example, the control variables may be randomly perturbed to see which positively influence the objective function. The objective function may be any suitable function, but could be a function of one or more of: product yield (e.g. methanol yield); energy efficiency of the chemical plant; catalyst performance; catalyst lifespan; economic cost; emissions cost; feedstock use; etc. In some cases, there may be a scheduled time to replace the catalyst 115. In such cases, a preferred objective function may involve optimising, or at least include a term for, the product yield of the chemical plant 100 prior to a scheduled time to replace the catalyst 115. In other cases, the objective function may involve optimising, or at least include a term for, the economic cost of running the chemical plant 100. This can provide different control variables, since it may involve maximising the life of the catalyst 115 or provide a recommendation of a time to replace the catalyst 115. An optimal time to replace the catalyst 115 may be a control variable. In step 460, the first local controller 130 may modify the mode of operation of the first chemical plant using the result of the analysis in step 445. Figure 5 shows a flow chart of a method 500 carried out by the system of Figure 2. As with the earlier embodiments, the method 500 analysing data stored on the remote server 140 uses the remote controller 150. In this embodiment, the result of the analysis is provided to the local controller 130 for display. The method includes steps 310, 320 and 340 from the method 300 of Figure 3, and additionally includes steps 530 and 550. Step 530 is similar to step 330. Step 530 involves analysing the first sensor data stored on the remote server 140 using the remote controller 150 using the first sensor data with a model of at least one component of the first chemical plant 100 to calculate a first parameter of the chemical process. In the method 500 of Figure 5, the first parameter can be a useful parameter for the operator of the chemical plant 100 to visualise to assist their manual control of the operation of the chemical plant 100. For example, the first parameter may be a parameter of the chemical process that is not measured using the plurality of sensors 120. It may even be a parameter that a sensor cannot directly measure. The parameter may be the performance of a catalyst 115 of the chemical plant 100. In some scenarios, the remote server 140 may store a different or more accurate model for calculating the first parameter than the local controller 130. For example, the model may be based on data stored on the remote server 140 obtained from other chemical plants 100. The model may be updated periodically or intermittently. In step 530, the remote controller 150 may optionally further process the calculated first parameter to provide an indicia for graphically representing information to the operator to the chemical plant 100. Alternatively, this processing may be carried out on the local controller 130 in step 550. In step 550, a display device 165 of the local controller 130 may display the result of the processing carried out by the remote controller 150. For example, there may be displayed a dashboard, visually indicating for an operator of the chemical plant 100 one or more of: the first parameter; a graphical representation of the first parameter; other parameters of the chemical process. 07 05 25

Claims

1. A method of monitoring a chemical plant, comprising:operating a first chemical plant to implement a chemical process;repeatedly using sensors of the first chemical plant to obtain a set of parameters of the chemical process to produce first sensor data;storing the first sensor data on a remote server;analysing the first sensor data stored on the remote server using a remote controller; and communicating the result of the analysis to a first local controller,wherein analysing the first sensor data stored on the remote server using the remote controller comprises using the first sensor data with a model of at least one component of the first chemical plant to calculate a first parameter of the chemical process and further comprises:using the model to generate a first trajectory of the first parameter over a first time period;using the model to generate a second trajectory of the first parameter over a second time period; and identify a difference between the calculated first and second trajectories.

2. The method of claim 1, further comprising controlling the first chemical plant based on the obtained first sensor data using the first local controller.

3. The method of claim 1 or claim 2, wherein the second time period is shorter than first time period.

4. The method of any one of claims 1 to 3, further comprising:comparing the magnitude of the identified difference with a threshold; and(a) generating an alert when the magnitude exceeds a threshold; or(b) generating an alert when the magnitude exceeds a magnitude threshold for a period of time exceeding a time threshold.

5. The method of claim 4, wherein communicating the result of the analysis to the local controller comprises communicating data representative of the identified difference.

6. The method of any one of claims 1 to 5, wherein:the model is a model of at least a component of the chemical plant;the first parameter is a measure of the performance of the component; andthe first trajectory is the expected decline in the performance of the component of the chemical plant overtime due to a set of one or more expected factors, the expected factors preferably including: fouling; thermal sintering; poisoning; and / or physical deposition.

7. The method of any preceding claim, wherein the method further comprisesoperating a second chemical plant to implement a chemical process;07 05 25repeatedly using sensors of the second chemical plant to obtain a set of parameters of a chemical process to produce second sensor data;storing the obtained second sensor data on the remote server, wherein analysing the first sensor data stored on the remote server using the remote controller further comprises tuning, training or updating the model using at least the second sensor data.

8. The method of claim 7, comprising controlling the second chemical plant based on the obtained second sensor data using a second local controller.

9. The method of any preceding claim, further comprising displaying an indicia representative of the result of the analysis using the first local controller.

10. The method of claim 9, wherein the indicia forms part of a dashboard, including other data derived from the first sensor data using the remote server.

11. The method of any preceding claim, further comprising modifying the mode of operation of the first chemical plant using the first local controller based on the communicated result of the analysis.

12. The method of any preceding claim, wherein:the first local controller controls a timing of the step of using the sensors of the first chemical plant to obtain the set of parameters of the chemical process to produce the first sensor data; andthe first local controller controls the timing based on the result of the analysis.

13. The method of claim 12, wherein the first local controller controls the timing based on the result of the analysis by controlling the frequency with which the sensors obtain the set of parameters based on the identified difference.

14. The method of any preceding claim, wherein the step of storing the first sensor data on the remote server comprises, storing the first sensor data on a local server;transmitting the first sensor data from the local server to the remote server, andwherein:the first local controller controls a transmission timing of the step of transmitting the first sensor data; andthe first local controller controls the transmission timing based on the result of the analysis.

15. The method of claim 14, wherein the first local controller controls the transmission timing based on the result of the analysis by controlling the frequency with which the first sensor data is transmitted.07 05 2516. The method of any preceding claim, wherein:analysing the first sensor data stored on the remote server using the remote controller comprises collecting a time-series of first sensor data over a time period; andusing the first sensor data over the time period with the model to calculate the first parameter of the chemical process,wherein the time period is at least one week, preferably at least four weeks, more preferably at least eight weeks.

17. The method of any preceding claim, wherein the chemical process is a heterogeneously-catalysed process, preferably the production of methanol or e-methanol.

18. The method of any preceding claim, wherein the at least one component includes a catalyst; and the at least one parameter is one or more of: a measure of performance of the catalyst; and / or a measure of degradation of the catalyst.

19. The method of claim 18, wherein the parameter is used to provide a recommended time to replace the catalyst.

20. The method of claim 18, wherein:there is scheduled time to replace the catalyst; andthe result of the analysis is a set of control variables for controlling the first chemical plant to maximise an objective function ahead of the scheduled time, wherein the objective function is preferably a function of one or more of: product yield; energy efficiency of the chemical plant; catalyst performance; catalyst lifespan; economic cost; emissions cost; and / or feedstock use.

Images