Secure Access Required
Welcome back!
Your password has been reset. Please set a new password to continue.
Verify reactor ownership to reset your password
Step 1: Read Verification
Enter the current SP value for
Step 2: Write Verification
Change the SP value for to a different value (>10% change), then enter the new value below.
Password Reset Successful
A temporary password has been sent to the email address on file for this account.
You must change this password on your next login.
Select devices to register
Your bioreactor needs to be connected to the AI cloud service. Download and run the setup wizard on the computer running D2MS.
Download Setup Wizard (54 MB)D:\D2MS Native\OAJ\D2MS_Gateway\TunnelSetup.exe and select "Run as administrator"Requirements: Windows 10/11 (64-bit), Administrator privileges
您的生物反应器需要连接到AI云服务。请在运行D2MS的电脑上下载并运行安装向导。
下载安装向导 (54 MB)D:\D2MS Native\OAJ\D2MS_Gateway\TunnelSetup.exe 选择"以管理员身份运行"系统要求： Windows 10/11（64位），管理员权限
Need help? Email: support@tjxbio.com 需要帮助？邮箱：support@tjxbio.com
Register your bioreactor
Ownership Verification - Step 1 of 2
Please check the current Set Point (SP) value for:
--
Find this value in your D2MS Software and enter it below.
Ownership Verification - Step 2 of 2
Now, please change the Set Point (SP) for:
--
to a different value (must differ by more than 10%), then enter the new value below.
After changing the value in D2MS Software, enter it here.
Don't need write access to the reactor?
Verification Successful!
Create a password and agree to the terms to complete registration.
End User License Agreement
Key Points:
| Parameter | Measured | Target | Target set by | Controller | Profile |
|---|
Usual and safe ranges, how a value is controlled and what to watch. Mia uses this when it explains or suggests changes.
Fill in what you know; everything except the value is optional.
The documents Mia can look things up in. Only you: in your library. Everyone: published for every user. Removing one takes it out of your library only.
User Requirement Specifications (URS) you uploaded. They belong to you, not to one reactor. Ask Mia about one and it answers each section from your product data and vessel designs.
When Mia holds a reading at a target (e.g. DO at 30 %), it moves one setpoint step by step. These settings say how hard and how fast. A loop without settings here uses Mia's cautious defaults.
Mia uses these settings whenever it holds this reading by moving this setpoint.
The readings and setpoints of this reactor, with their values now. Choose the ones that drive what you want to predict.
Usually a reading you cannot measure online often enough, or one you want to see coming.
How large the model is, how much history it looks at and how it learns.
Based on current configuration, you need:
Copy this reactor's config to another reactor, or load config from another reactor.
Predictor not configured
Please configure the predictor first by selecting input features and output parameter.
Choose a run and time range to use for model training.
No predictions available. Configure and train the model first.
Each training creates a new model version. You can load a previous version to use for predictions.
Loading model history...
Compare model predictions against actual historical data to evaluate accuracy.
The control loops Mia ran on this reactor. Run again… repeats one with the same settings (it asks first); Stop halts a running one.
Measurements Mia logged at high frequency on this reactor (e.g. kLa tests). Open the graphs or the analysis, or download the data.
The runs marked good or bad in Run Data. Tick the ones Mia should learn from; good runs define what a normal batch looks like.
The columns of the reference runs, under the names Run Data shows. Columns that never changed are not useful and start unticked.
Select reference runs in the Reference Runs tab, then open this tab to see available features.
Mia compares these with the reference while a batch runs, and corrects when they drift.
Available after loading features above.
Lab measurements of the reference runs, compared with the reference at each sample time.
How large the prediction model is and how long it learns.
Review and clean training data. Trim end-of-run artifacts, remove outliers, then save as a reusable dataset.
Mia learns what a normal batch looks like from the reference runs. Check the answers below, then start.
Score a historical run against the golden reference to see how well it matches.
Rules used by the L6 Supervisor to automatically correct parameter deviations.
Typical and safe operating ranges for each parameter.
| Parameter | Typical Range | Safe Range | Control Method | MV |
|---|
PID tuning for controllable manipulated variables.
| Parameter (PV) | MV | Kp | Deadband | Max Change/Cycle |
|---|
L6 supervisor threshold checks that raise deviation alerts.
| Name | Parameter | Check Type | Threshold | Severity | Phase |
|---|
Choose the organism model Mia uses to estimate what the cells are doing inside (metabolic fluxes) from this reactor's live data. It applies to this session.
Every value Mia knows, under its name and the other names it answers to. If Mia doesn't recognise a name you use, add it here as another name.
The name everyone sees for this reactor in Mia, and where it stands.
How Mia reaches this reactor's controller to read values and write setpoints.
Where Mia reads this reactor's past runs, profiles and setpoint sources.
What the reactor is used for, and the vessel geometry Mia uses to estimate mixing and oxygen transfer.
What is fitted, so Mia only offers pumps, gases and probes the reactor has.
Kept with the reactor's record for reports and audits.
The reactor's maximums: Mia never plans or suggests beyond them.
Who can see or work on this reactor. Their role decides whether they can change anything.
Loading users...
New users are created with observer role (read-only access). You can promote them later.
No configuration selected.
Loading users...
How Mia shows each value of this reactor, and the range it keeps every write inside. Click a row to change it.
Changes apply to this reactor only.
Mia warns (yellow) or raises an alarm (red) when a reading leaves its range.
Leave a box empty for no limit on that side.
Mia raises an alarm when an actuator's reading stays too far from its setpoint — a stuck pump, a blocked line.
The alarm fires when the reading differs from the setpoint by more than the allowed deviation.
Calculations Mia can run on your run data (kLa, growth rate…). Test one on a run before relying on it.
The steps behind each calculation, written out so a person can check them.
A soft sensor turns readings into a value you cannot measure directly (oxygen uptake, biomass, growth rate). On = Mia calculates it for this reactor's runs and live data. Shared = other users can find and use your sensor.
In your own words — Mia uses the description to write the calculation for you.
Choose a run — the Reactor column says where it was recorded — then tick the columns the calculation needs, under the names Run Data shows.
Let Mia write the calculation from your description and the chosen columns, then check it. You can edit the code if you know Python.
After saving you can try it on a run straight away.
Preview calculates the values and shows them — nothing is saved. Save into the run writes them into that run's data, where charts, Golden Batch and Model Fit can use them.
What the formula is called and what it is for.
Define expected ranges to validate results
The Titer Predictor uses a recurrent neural network (GRU/LSTM) to forecast future values based on historical process data. The model learns patterns from your bioreactor's time-series data and predicts values at configurable time horizons.
The minimum raw data points needed for training is calculated as:
Example: With Sequence=60 steps, Max Horizon=480 steps, Min Sequences=100, Avg Window=30s:
Downsampled = 60 + 480 + 100 = 640 points
Raw = 640 × 6 = 3,840 points (~5.3 hours at 5-second intervals)
| Input Features | Process variables (PVs) and setpoints (SPs) used as inputs to the model. Select parameters that influence your target output. |
| Output Parameter | The target variable to predict (e.g., titer, biomass). This is what the model learns to forecast. |
| Output as Input (AR) | Autoregressive mode - uses past output values as additional input features. Lag steps controls how many past values to include. |
| Model Type | GRU - Faster training, fewer parameters. LSTM - Better for very long sequences, more memory. |
| Avg Window (sec) | Downsamples training data by averaging points. Higher = less data, faster training. 30s (6x reduction) is a good default. |
| Min Sequences | Minimum training sequences required. Lower = less data needed but potentially less accurate. 50-200 typical. Directly affects data requirements. |
| Auto-Retrain (min) | Automatically retrain the model at this interval. Set to 0 to disable. Uses "Quick Train Epochs" for faster training. |
| Sequence (steps) | How much history the model sees, in steps of the averaging window (30 s by default: 60 steps = 30 min). Longer = more context but more data needed. Prediction horizons are counted in the same steps. |
| Hidden Size | Number of neurons in each recurrent layer. Larger = more capacity but slower. 64-256 typical. |
| Layers | Number of stacked recurrent layers. 1-2 usually sufficient. More layers = more complex patterns but harder to train. |
| Dropout | Regularization to prevent overfitting. 0.1-0.3 typical. Higher if you see overfitting. |
| Learning Rate | How fast the model learns. 0.001 is a good default. Lower if training is unstable. |
| Batch Size | Samples processed before updating weights. 16-64 typical. Larger = faster but uses more memory. |
| Max Epochs | Maximum training iterations for manual "Train Now". Early stopping may end training sooner. |
| Quick Train Epochs | Epochs used for auto-retrain (fewer for faster background training). |
| Early Stop Patience | Stop training if loss doesn't improve for this many epochs. Saves time and prevents overfitting. |
| Horizons (min) | Comma-separated list of future times to predict. E.g., "60, 120, 240" predicts 1h, 2h, 4h ahead. Max horizon affects data requirements! |
| Log Predictions | Which horizons to log to the data history for later analysis and charting. |
Short tips Mia shows users as the tip of the day. A tip that is not shown stays here and can be shown again.
Loading tips...
Changes proposed by your devices, apps and agents that use your Chat API key. Nothing changes on a reactor until you approve.
Generate an API token to connect an MCP client or third-party device to this bioreactor.
Copy now - it won't be shown again.
Paste this into your MCP client's settings.
| Reactor | Type | User | Token | |
|---|---|---|---|---|
| Loading... | ||||